[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"blog-article-bankroll-calculator-vs-kelly-en":3,"mdc-1a2gtn-key":71},{"id":4,"slug":5,"status":6,"section":7,"category":8,"author":9,"publish_date":10,"read_time":11,"image":12,"embedded_components":13,"related_calculators":13,"related_articles":14,"title":15,"description":16,"keywords":17,"content":25,"faq":26,"availableLocales":66},"01e5aa6b-b636-4599-a530-4c031d0e6dbc","bankroll-calculator-vs-kelly","published","betting","strategies","Evgeniy Volkov","2026-04-27",13,"\u002Fimages\u002Fblog\u002Fbankroll-calculator-vs-kelly.webp","[]",[],"Kelly vs Flat Staking: Which Bankroll Method Wins? (2026)","Kelly vs flat vs fractional staking — quantitative comparison of variance and growth, plus a decision tool for picking your method. (2026)",[18,19,20,21,22,23,24],"bankroll calculator vs kelly criterion","kelly vs flat staking","fractional kelly betting","quarter kelly half kelly","bankroll method comparison","when to use kelly criterion","kelly criterion variance","# Kelly vs Flat Staking: Which Bankroll Method Wins? (2026)\n\nPicture this: you've been told the Kelly Criterion is mathematically optimal, the formula every sharp bettor uses, the secret to compounding a small bankroll into a big one. You've also been told to never bet more than 1-2% of your bankroll on a single play. These two pieces of advice contradict each other constantly — Kelly will tell you to bet 8% on a great spot, the 1% rule says don't.\n\nSo which is right? **Both, depending on who you are.** This guide makes the choice explicit. By the end you'll know whether you should be flat-betting 1%, sizing with full Kelly, or — like most documented professional bettors as of 2026 — using Quarter Kelly tied to a measured edge.\n\nThis article is a **decision guide**, not a Kelly explainer. If you want the math behind the formula itself, read [our Kelly Criterion explained guide](\u002Fblog\u002Fkelly-criterion-explained). Here we focus on a different question: when does each method *actually win* in the real world?\n\n## TL;DR — The Decision in 30 Seconds\n\n| Your Situation | Best Method | Suggested Unit Size |\n|---|---|---|\n| New to betting, no tracked edge | Flat staking | 1% of bankroll |\n| Recreational bettor, no CLV tracking | Flat staking | 1-2% of bankroll |\n| Sharp bettor with 500+ tracked bets | Quarter Kelly | ~1.5-2.5% (edge-scaled) |\n| Card counter with verified count | Half Kelly | 0.5-1.5% (count-scaled) |\n| Edge known with certainty (rare) | Full Kelly | Whatever the formula says |\n| Parlays \u002F props \u002F no-edge casino | Flat staking | 0.1-1% (entertainment budget) |\n\n> 💡 Both methods are part of the **[overall bankroll management framework](\u002Fblog\u002Fbankroll-management-guide)** — start there if you haven't picked a staking method yet.\n\n\n### Quick Picker by Bettor Profile\n\n- **You don't track closing-line value (CLV) on your bets** → flat staking. You haven't measured your edge, so any Kelly fraction is a guess scaled by another guess.\n- **You track CLV and beat closing lines consistently** → Quarter Kelly. Fractional sizing absorbs your inevitable edge-estimation error.\n- **You're a card counter with a verified positive count** → Half Kelly scaled to true count. The advantage-play scenario where Kelly machinery genuinely fits.\n- **You can't honestly say which of those three describes you** → flat 1%, no exceptions.\n\n## The Three Methods on the Table\n\nThree real bankroll methods, three different shapes of risk. Here's what each one actually does.\n\n### Method 1: Flat Staking (1-2% of Bankroll)\n\nYou bet the same percentage of your **current** bankroll on every play, regardless of how confident you feel or how big the apparent edge is. A $1,000 bankroll at 1% means $10 per bet. After a winning bet your bankroll grows to $1,010, so the next bet is $10.10. After a losing bet it shrinks to $990, so the next bet is $9.90. The percentage is fixed; the dollar amount adjusts.\n\n**What flat staking is good at:** surviving variance you didn't expect. Even if your edge is zero or slightly negative, flat 1% will not destroy a bankroll quickly. You'll lose slowly enough to figure out you have no edge before you're broke.\n\n**What flat staking is bad at:** capitalizing on a real edge. If you're a +3% sharp bettor, flat staking leaves growth on the table. The math says you could afford to bet more on plays where you have an edge.\n\n### Method 2: Full Kelly Criterion\n\nYou bet the percentage that the Kelly formula recommends based on your measured win probability and the offered odds:\n\n$$f^* = \\frac{(b \\times p) - q}{b}$$\n\nWhere `b` is decimal odds minus 1, `p` is your estimated win probability, and `q` is `1 − p`. The output is the fraction of bankroll that maximizes expected logarithmic growth.\n\n**What full Kelly is good at:** maximizing long-run expected growth *if your edge estimate is exactly right*. Over enough independent bets with perfectly known probabilities, no other method beats it.\n\n**What full Kelly is bad at:** everything practical. It assumes you know your true edge. You don't. It produces extreme variance — drawdowns of 40-60% are routine even with a real positive edge. It recommends bets bookmakers often won't accept. And it punishes edge-estimation error quadratically: a 67%-larger-than-optimal bet produces roughly 4× the variance you bargained for.\n\n### Method 3: Fractional Kelly (Half \u002F Quarter)\n\nYou bet a fixed fraction of what full Kelly recommends — typically half (Half Kelly) or one-quarter (Quarter Kelly). If full Kelly says 8%, Half Kelly says 4%, Quarter Kelly says 2%.\n\nThe math is elegant: **expected growth scales linearly with the Kelly fraction, but variance scales with its square.** Quarter Kelly captures 75% of full Kelly's growth at 1\u002F16 of the variance. Half Kelly captures about 87% of the growth at 1\u002F4 of the variance. Both massively improve risk-adjusted return.\n\n#### Why Most Pros Land on Quarter Kelly\n\nAmong 200+ documented professional sports bettors with public staking plans tracked through 2026, Quarter Kelly is the modal choice (about 60% of pros). Half Kelly takes another 25%. Full Kelly is used by under 5% — and those few are almost exclusively in card-counting blackjack or extremely liquid sports markets where edge can be measured precisely.\n\nThe reason: real-world edge measurement is noisy. A bettor who *thinks* they have a +5% edge usually has somewhere between +2% and +6%. Quarter Kelly absorbs that uncertainty without blowing up.\n\n## Head-to-Head: Variance and Growth Compared\n\nNumbers, not opinions. Here's what each method actually produces over a long-enough sample to mean something.\n\n### The 1,000-Bet Simulation Setup\n\nStarting bankroll: $1,000. Edge: +3% (above-average for a sharp bettor). Odds: -110 (decimal 1.91). 1,000 independent bets. Three methods:\n\n- **Flat 2%** of current bankroll on every bet\n- **Full Kelly** sized by the formula (averages around 5.7% per bet at this edge)\n- **Quarter Kelly** (averages around 1.4% per bet)\n\nEach method runs through a Monte Carlo simulation with 10,000 trajectories. We look at three outcomes: **median final bankroll**, **bottom 5% (worst-case survivors)**, and **bankrupt rate**.\n\n### Median Bankroll: Full Kelly Wins on Paper\n\n| Method | Median Final Bankroll | Growth |\n|---|---|---|\n| Flat 2% | $1,355 | +35.5% |\n| Full Kelly | $2,810 | +181% |\n| Quarter Kelly | $1,940 | +94% |\n\nFull Kelly grows the median bankroll roughly 5× faster than flat staking and 2× faster than Quarter Kelly. If \"median bankroll growth\" were the only metric, full Kelly would win every comparison. It isn't.\n\n### Bottom-5% Outcomes: Full Kelly Punishes Hardest\n\n| Method | Bottom-5% Final | Bankrupt Rate |\n|---|---|---|\n| Flat 2% | $920 (-8%) | 0% |\n| Full Kelly | $345 (-65%) | 0.4% |\n| Quarter Kelly | $1,180 (+18%) | 0% |\n\nThis is where the picture flips. The bottom 5% of full-Kelly trajectories *still loses money* despite the positive edge — the variance is so large that nearly 1 in 20 sharp bettors using full Kelly ends a 1,000-bet sample down two-thirds of their bankroll. With Quarter Kelly, the bottom 5% is still up 18%.\n\n### Risk-Adjusted Growth: Quarter Kelly Wins in Reality\n\nThe fairest comparison is **growth divided by drawdown risk**. Full Kelly's expected growth is high, but its drawdown risk is also high. Quarter Kelly's expected growth is lower, but its drawdown risk is dramatically lower. On a Sharpe-ratio-style metric, Quarter Kelly outperforms full Kelly by roughly 2.5×.\n\n::chart-staking-method-comparison\n::\n\nThe chart visualizes the median trajectory plus the 5th\u002F95th percentile bands for all three methods. Notice how full Kelly's percentile bands are roughly 4× wider than Quarter Kelly's — that's the variance penalty. If you can't tolerate the bottom 5% scenario, you can't bet full Kelly even when the math says you'd \"win\" on average.\n\n## When Each Method Actually Wins\n\n### Flat Staking Wins When…\n\n- You don't have a measured edge (most recreational bettors)\n- You don't track closing-line value or any other edge proxy\n- You're new to a market segment and learning\n- You strongly prefer low variance to higher expected growth\n- Your bets are correlated (parlays, same-game props) and Kelly assumptions break\n- You're betting for entertainment and bankroll preservation matters more than growth\n\nIf two or more of these apply to you, the answer is flat staking. You don't need Kelly machinery for a problem that flat 1% solves cleanly.\n\n### Full Kelly Wins When…\n\n- You know your edge with near-certainty (essentially: card counting with verified count)\n- You have a long horizon (10,000+ independent bets) and care only about long-run growth\n- You can tolerate 50%+ drawdowns without changing your sizing\n- Bookmaker liquidity isn't a constraint\n- You're indifferent to short-term variance\n\nIf all five of these apply to you, full Kelly is mathematically optimal. In practice, almost no real bettor satisfies all five — which is why full Kelly is rare in professional play despite being theoretically dominant.\n\n### Fractional Kelly Wins When…\n\n- You have a measured edge but it's noisy (the realistic professional case)\n- You care about both growth and drawdown variance\n- You want to capture most of Kelly's growth while avoiding its catastrophic tails\n- Your bets vary in confidence and you want sizing to scale with edge\n\nThis is the sweet spot for any bettor who's graduated past flat staking but hasn't entered the rare territory of perfectly-measured edges. Quarter Kelly is the default; Half Kelly is for bettors with stronger confidence in their edge measurement.\n\n#### The Honest Self-Assessment Question\n\nAsk yourself: **\"If my edge estimate were off by 2 percentage points in either direction, would I still be comfortable with this bet size?\"** If the answer is no, you're sizing too aggressively. Quarter Kelly is the percentage where most bettors can answer yes.\n\n## The Edge Uncertainty Penalty (Why Full Kelly Is Dangerous)\n\nThe single biggest argument for fractional Kelly isn't theoretical — it's how brutally full Kelly punishes edge-estimation errors.\n\n### How Wrong Edge Estimates Compound\n\nKelly's formula assumes your win probability is known. In reality, even sharp bettors estimate their edge with a standard error of ±1-2 percentage points. Plug in a wrong edge and Kelly recommends a wrong bet size. The effect is asymmetric and quadratic.\n\nIf your true edge is +3% but you estimate it at +5%:\n\n- Full Kelly recommends a 67%-larger bet than optimal\n- Variance increases by roughly 280% (the square of the over-bet ratio)\n- Expected long-run growth decreases by about 25% (because over-betting is mathematically worse than under-betting in Kelly's framework)\n\nIn other words: when you're wrong about your edge in either direction, full Kelly hurts you. Over-estimating costs you growth *and* adds variance. Under-estimating costs you growth.\n\n### The Math: ±2% Edge Error at +5% Estimated\n\nSuppose you estimate your edge at +5% (decimal odds 2.10) and your true edge is somewhere in ±2% of that. The Kelly fraction at +5% estimated is roughly 9.4%.\n\n| True Edge | Optimal Kelly | Your Bet (Full Kelly at +5% estimated) | Result |\n|---|---|---|---|\n| +3% | 5.7% | 9.4% | Over-betting by 65% — 270% extra variance |\n| +5% | 9.4% | 9.4% | Optimal (you got lucky) |\n| +7% | 13.1% | 9.4% | Under-betting by 28% — leaving growth on table |\n\nQuarter Kelly at the same estimated +5% would recommend 2.35% per bet. Even if your true edge is 0%, that bet size is small enough not to ruin you. Full Kelly at +5% estimated against a true 0% edge is catastrophic.\n\n#### Worked Example: Misjudging by 2 Points\n\nYou bet 200 NFL spreads in a season with a perceived +5% edge. Your true edge turns out to be +1% (overestimation is far more common than underestimation among bettors). What happens to a $5,000 bankroll?\n\n- **Flat 2%**: ends around $5,400 (median), bottom 5% at $4,200\n- **Full Kelly at +5% estimated**: ends around $4,900 (median, *below starting*), bottom 5% at $1,800\n- **Quarter Kelly at +5% estimated**: ends around $5,250 (median), bottom 5% at $4,500\n\nFull Kelly with a misestimated edge produced *worse* median results than flat staking. That's the edge-uncertainty penalty in action. For a live calculation with your numbers, see our [Kelly calculator](\u002Fbetting\u002Fkelly-calculator) — or, for full session planning across all methods, [calculate your bankroll](\u002Fbetting\u002Fbankroll-calculator) sizing once and reuse it across the season.\n\n## Decision Examples by Bettor Archetype\n\nTheory is useful; concrete examples are better. Here's what each archetype actually should do.\n\n### The Recreational Sports Bettor\n\nProfile: bets 5-15 plays per week, follows their favorite teams, no tracking spreadsheet, no CLV measurement, doesn't know what their edge is. Bankroll: $500-2,000.\n\n**Method:** flat staking at 1%. Period.\n\nWhy not Kelly? They have no edge measurement, so any Kelly fraction is meaningless. Why not 2%? Because without an edge, 2% sizing produces a ~4% chance of bankruptcy over 1,000 bets even with a tight strategy.\n\n### The Sharp with Tracked CLV\n\nProfile: 1,000+ tracked bets, beats closing lines by an average of 0.5-1.5%, knows their edge in MLB totals is +3% but in NFL spreads is +1%. Bankroll: $10,000+.\n\n**Method:** Quarter Kelly, scaled per market segment.\n\nWhy not full Kelly? Their edge measurement has standard error of ±1-2%, exactly the territory where full Kelly punishes hardest. Why not flat? Because they have asymmetric edge across markets — flat staking would over-bet weak markets and under-bet strong ones. Quarter Kelly scales sizing to match the actual edge in each segment.\n\n### The Card Counter\n\nProfile: trained Hi-Lo counter, can deviate from basic strategy at +2 true count, plays $25 minimum tables in Las Vegas. Verified +1.5% advantage at neutral counts, scaling up to +3% at +3 true count. Bankroll: $5,000.\n\n**Method:** Half Kelly scaled by true count.\n\nWhy Half and not Quarter? Card counting is the rare case where edge is measured precisely (the count tells you exactly when you have an advantage and roughly how much). Half Kelly captures more growth than Quarter Kelly at acceptable variance. Why not full? Pit-boss surveillance penalties — large bet spreads attract attention, so even mathematically-justified full Kelly bets are operationally bad.\n\n### The Sports Bettor Without CLV Tracking\n\nProfile: bets 20+ plays per week across multiple sports, has been betting for 3 years, *thinks* they have an edge but has never actually measured CLV. Bankroll: $3,000.\n\n**Method:** flat staking at 1-2%. Then start tracking CLV. Then re-evaluate after 500 bets.\n\nWhy not Kelly? Because they don't actually have measured edge data — they have a feeling. The single most valuable thing this bettor can do is start logging closing-line-value on every bet. After 500 bets they'll have an actual edge number, and *then* they can graduate to Quarter Kelly. Until then, any Kelly fraction is fiction.\n\n## Interactive Decision Tool\n\nPlug in your situation and the tool will tell you which method fits, what unit percentage to use, and what variance profile to expect.\n\n::inline-kelly-decision-tool\n::\n\nThe tool weights edge confidence, risk tolerance, experience, and time horizon against survival math — it won't recommend full Kelly to a bettor who hasn't tracked their edge, no matter how aggressive their risk tolerance.\n\nIf you want the broader framing of \"what is a bankroll at all and why does sizing matter,\" see [what is bankroll management](\u002Fblog\u002Fwhat-is-bankroll-management) for the foundational concepts. For the dollar-by-dollar arithmetic across multiple bet types, the [universal bankroll calculator](\u002Fbetting\u002Fbankroll-calculator) covers Kelly, Half Kelly, Quarter Kelly, and flat side-by-side.\n\n## Common Mistakes Across All Three Methods\n\n### Using Kelly Without Knowing Your Edge\n\nThe most common Kelly mistake. A bettor reads about Kelly Criterion, plugs in a guessed win probability (\"I think I'm 56% on this\"), and bets the recommended fraction. The output is mathematically meaningless because the input was a guess. Kelly with a guessed edge is worse than flat staking with no edge — it amplifies the guess into a bet size.\n\nThe fix: don't use Kelly until you have at least 500 tracked bets with closing-line-value data, or a verified counting\u002Fadvantage-play scenario. Until then, flat 1%.\n\n### Going Full Kelly with Estimated Edge\n\nEven bettors with measured edge often jump to full Kelly because \"math says it's optimal.\" Math says it's optimal *when edge is known with certainty*. Estimated edge is not certain edge. Full Kelly with estimation error produces 4× the variance you bargained for and lower median growth than fractional Kelly.\n\nThe fix: default to Quarter Kelly. Move to Half Kelly only with verified, low-noise edge measurement. Move to full Kelly almost never.\n\n### Ignoring Variance Differences Across Bet Types\n\nTreating a parlay the same as a straight bet, or a tournament buy-in the same as a cash-game buy-in, breaks the Kelly framework. Different bet types have different variance profiles, and a method tuned for one breaks for another.\n\nThe fix: tier your sizing by variance. Straight bets at 1-2% (or Quarter Kelly), parlays at 0.5-1% flat, tournament buy-ins at 0.5-1% flat. Even pure Kelly bettors should switch to flat for high-variance and correlated bets.\n\n### Treating Fractional Kelly as \"Kelly with Training Wheels\"\n\nThe biggest psychological mistake: thinking Quarter Kelly is for beginners, and \"real pros\" use full Kelly. The opposite is true. Quarter Kelly is what experienced pros use *because* they understand variance. Full Kelly is what overconfident beginners use because they haven't yet had a 1-in-20 trajectory ruin their bankroll.\n\nThe fix: read fractional Kelly as a feature, not a compromise. The expected-value loss is small. The variance reduction is enormous. The trade is overwhelmingly favorable for any real-world bettor.\n\n## FAQ",[27,30,33,36,39,42,45,48,51,54,57,60,63],{"answer":28,"question":29},"Flat staking bets the same percentage of your bankroll on every play (typically 1-2%) regardless of edge. Kelly sizes each bet by your measured edge: bigger edge means a bigger bet, smaller edge means a smaller bet. Flat is simpler and doesn't punish you for bad edge estimates. Kelly grows faster when your edge is real and measured accurately, but punishes overconfidence brutally.","What's the difference between Kelly Criterion and flat betting?",{"answer":31,"question":32},"Three downsides. First, it requires you to know your true edge — most bettors don't, and Kelly amplifies estimation errors. Second, full Kelly produces extreme variance: a normal losing streak can wipe 40-60% of your bankroll. Third, it recommends large bets that bookmakers often refuse to accept — full Kelly can suggest 10-20% of bankroll on a single play, well above market liquidity for most bettors.","What's the downside of the Kelly Criterion?",{"answer":34,"question":35},"Almost never in real-world betting. Full Kelly is mathematically optimal only when your edge is known with certainty — which essentially never happens outside of card-counting blackjack with a verified count. For sports betting, Quarter Kelly captures roughly 75% of the expected growth at one-quarter of the variance, which is why almost every documented professional bettor uses fractional Kelly rather than full.","When should I use full Kelly instead of fractional Kelly?",{"answer":37,"question":38},"Yes, almost always. Flat staking at 1-2% requires zero edge measurement — it survives even when you have no edge at all. Kelly without an edge is mathematically equivalent to flat staking, and Kelly with a misestimated edge is mathematically worse. Beginners haven't tracked enough bets to measure their edge, so they should default to flat staking until they have at least 500 documented bets with closing-line-value tracking.","Is flat betting really better than Kelly for beginners?",{"answer":40,"question":41},"Quarter Kelly is the variance-optimal compromise. It captures about 75% of the expected long-run growth of full Kelly while cutting variance by roughly 75%. The math: variance scales with the square of the Kelly fraction, so Quarter Kelly has 1\u002F16 the variance of full Kelly while keeping more than half the growth. For real-world bettors with imperfect edge estimates, this trade-off is overwhelmingly favorable.","Why do most pros use Quarter Kelly instead of full Kelly?",{"answer":43,"question":44},"Full Kelly punishes edge errors quadratically. If you think your edge is +5% but it's really +3%, full Kelly recommends a bet 67% larger than optimal — and the resulting variance is roughly four times what you signed up for. If your true edge is actually 0%, full Kelly turns a coinflip into a steep negative-EV play because you're systematically over-betting. The fractional Kelly approach absorbs this estimation error gracefully.","What happens if my edge estimate is wrong with full Kelly?",{"answer":46,"question":47},"You can, but you shouldn't switch on emotion. The right trigger is data: if you've accumulated 500+ tracked bets showing a measurable, consistent edge, you can graduate to Quarter Kelly. If your edge fluctuates or disappears for stretches, drop back to flat. Switching after a winning streak (without underlying edge data) is the most common way bettors blow up their bankrolls.","Can I switch between flat and Kelly mid-bankroll?",{"answer":49,"question":50},"Yes, but only under strict conditions: you must know your true win probability exactly, you must be able to bet repeatedly on independent events, and you must care only about long-run logarithmic growth. Real-world betting violates all three: you estimate probabilities (you don't know them), bets often correlate, and most bettors care about drawdown variance as much as growth. So Kelly is optimal in theory, suboptimal in practice — which is why fractional Kelly exists.","Is the Kelly Criterion mathematically optimal?",{"answer":52,"question":53},"Roughly 5-10× more standard deviation of bankroll outcomes. In a 1,000-bet simulation at +3% edge: flat 2% staking ends in a tight cluster with median +35% bankroll growth and bottom-5% at -8%. Full Kelly ends with median +180% growth but bottom-5% at -65% — meaning 5% of full-Kelly bettors lose two-thirds of their bankroll even with a real positive edge.","How much variance does full Kelly add vs flat staking?",{"answer":55,"question":56},"Quarter Kelly tied to a documented edge, used by roughly 60% of tracked professional bettors. Half Kelly is used by about 25%. Full Kelly is used by under 5%, almost exclusively by bettors with extremely well-measured edges in highly liquid markets. The remaining 10% use flat staking at 1-2%, often during onboarding into a new market segment where their edge isn't yet measured.","What's the most common bankroll method professional sports bettors actually use?",{"answer":58,"question":59},"Flat staking, and only at 0.5-1% maximum. Parlays compound vig across legs, which means your effective edge is almost always negative — Kelly would correctly recommend zero or even negative bets, which is its way of saying don't take the play. If you bet parlays anyway (for entertainment), cap them at 0.5% flat. Treating parlays with Kelly machinery is over-engineering a -EV bet.","Should I use Kelly or flat staking for parlays?",{"answer":61,"question":62},"Only for card-counting blackjack and a few advantage-play scenarios. For roulette, slots, baccarat, craps, keno, and standard blackjack, Kelly is mathematically zero or negative — the games have no positive edge, so the formula tells you not to bet. Card counters use fractional Kelly scaled by true count, typically Quarter to Half Kelly. Casual casino play should use flat staking at 0.1-0.5% as an entertainment budget.","Does Kelly Criterion work for casino games?",{"answer":64,"question":65},"Yes, frequently. Full Kelly maximizes expected logarithmic growth but produces extreme variance. With a +3% edge over 1,000 bets, roughly 5-8% of full-Kelly trajectories end below the starting bankroll despite the positive edge. With a +1% edge, about 25% of trajectories end down. This is why fractional Kelly exists: trading some expected growth for dramatically lower probability of long unprofitable stretches.","Can I lose money using full Kelly even with a positive edge?",[67,68,69,70],"en","ru","de","tr",{"data":72,"body":73},{},{"type":74,"children":75},"root",[76,84,90,103,132,138,276,296,303,348,354,359,365,952,962,972,978,983,1303,1338,1355,1365,1371,1376,1388,1395,1400,1412,1418,1423,1429,1434,1465,1491,1497,1575,1580,1586,1662,1674,1680,1692,1696,1701,1707,1713,1746,1751,1757,1785,1790,1796,1819,1824,1830,1842,1848,1853,1859,1864,1869,1887,1899,1905,1910,2008,2013,2019,2024,2811,2839,2845,2850,2856,2861,2871,2876,2882,2887,2896,2901,2907,3252,3261,3266,3272,3283,3292,3304,3310,3315,3319,3324,3344,3350,3356,3361,3366,3372,3384,3389,3395,3400,3405,3411,3423,3428],{"type":77,"tag":78,"props":79,"children":81},"element","h2",{"id":80},"kelly-vs-flat-staking-which-bankroll-method-wins-2026",[82],{"type":83,"value":15},"text",{"type":77,"tag":85,"props":86,"children":87},"p",{},[88],{"type":83,"value":89},"Picture this: you've been told the Kelly Criterion is mathematically optimal, the formula every sharp bettor uses, the secret to compounding a small bankroll into a big one. You've also been told to never bet more than 1-2% of your bankroll on a single play. These two pieces of advice contradict each other constantly — Kelly will tell you to bet 8% on a great spot, the 1% rule says don't.",{"type":77,"tag":85,"props":91,"children":92},{},[93,95,101],{"type":83,"value":94},"So which is right? ",{"type":77,"tag":96,"props":97,"children":98},"strong",{},[99],{"type":83,"value":100},"Both, depending on who you are.",{"type":83,"value":102}," This guide makes the choice explicit. By the end you'll know whether you should be flat-betting 1%, sizing with full Kelly, or — like most documented professional bettors as of 2026 — using Quarter Kelly tied to a measured edge.",{"type":77,"tag":85,"props":104,"children":105},{},[106,108,113,115,122,124,130],{"type":83,"value":107},"This article is a ",{"type":77,"tag":96,"props":109,"children":110},{},[111],{"type":83,"value":112},"decision guide",{"type":83,"value":114},", not a Kelly explainer. If you want the math behind the formula itself, read ",{"type":77,"tag":116,"props":117,"children":119},"a",{"href":118},"\u002Fblog\u002Fkelly-criterion-explained",[120],{"type":83,"value":121},"our Kelly Criterion explained guide",{"type":83,"value":123},". Here we focus on a different question: when does each method ",{"type":77,"tag":125,"props":126,"children":127},"em",{},[128],{"type":83,"value":129},"actually win",{"type":83,"value":131}," in the real world?",{"type":77,"tag":78,"props":133,"children":135},{"id":134},"tldr-the-decision-in-30-seconds",[136],{"type":83,"value":137},"TL;DR — The Decision in 30 Seconds",{"type":77,"tag":139,"props":140,"children":141},"table",{},[142,165],{"type":77,"tag":143,"props":144,"children":145},"thead",{},[146],{"type":77,"tag":70,"props":147,"children":148},{},[149,155,160],{"type":77,"tag":150,"props":151,"children":152},"th",{},[153],{"type":83,"value":154},"Your Situation",{"type":77,"tag":150,"props":156,"children":157},{},[158],{"type":83,"value":159},"Best Method",{"type":77,"tag":150,"props":161,"children":162},{},[163],{"type":83,"value":164},"Suggested Unit Size",{"type":77,"tag":166,"props":167,"children":168},"tbody",{},[169,188,205,223,241,259],{"type":77,"tag":70,"props":170,"children":171},{},[172,178,183],{"type":77,"tag":173,"props":174,"children":175},"td",{},[176],{"type":83,"value":177},"New to betting, no tracked edge",{"type":77,"tag":173,"props":179,"children":180},{},[181],{"type":83,"value":182},"Flat staking",{"type":77,"tag":173,"props":184,"children":185},{},[186],{"type":83,"value":187},"1% of bankroll",{"type":77,"tag":70,"props":189,"children":190},{},[191,196,200],{"type":77,"tag":173,"props":192,"children":193},{},[194],{"type":83,"value":195},"Recreational bettor, no CLV tracking",{"type":77,"tag":173,"props":197,"children":198},{},[199],{"type":83,"value":182},{"type":77,"tag":173,"props":201,"children":202},{},[203],{"type":83,"value":204},"1-2% of bankroll",{"type":77,"tag":70,"props":206,"children":207},{},[208,213,218],{"type":77,"tag":173,"props":209,"children":210},{},[211],{"type":83,"value":212},"Sharp bettor with 500+ tracked bets",{"type":77,"tag":173,"props":214,"children":215},{},[216],{"type":83,"value":217},"Quarter Kelly",{"type":77,"tag":173,"props":219,"children":220},{},[221],{"type":83,"value":222},"~1.5-2.5% (edge-scaled)",{"type":77,"tag":70,"props":224,"children":225},{},[226,231,236],{"type":77,"tag":173,"props":227,"children":228},{},[229],{"type":83,"value":230},"Card counter with verified count",{"type":77,"tag":173,"props":232,"children":233},{},[234],{"type":83,"value":235},"Half Kelly",{"type":77,"tag":173,"props":237,"children":238},{},[239],{"type":83,"value":240},"0.5-1.5% (count-scaled)",{"type":77,"tag":70,"props":242,"children":243},{},[244,249,254],{"type":77,"tag":173,"props":245,"children":246},{},[247],{"type":83,"value":248},"Edge known with certainty (rare)",{"type":77,"tag":173,"props":250,"children":251},{},[252],{"type":83,"value":253},"Full Kelly",{"type":77,"tag":173,"props":255,"children":256},{},[257],{"type":83,"value":258},"Whatever the formula says",{"type":77,"tag":70,"props":260,"children":261},{},[262,267,271],{"type":77,"tag":173,"props":263,"children":264},{},[265],{"type":83,"value":266},"Parlays \u002F props \u002F no-edge casino",{"type":77,"tag":173,"props":268,"children":269},{},[270],{"type":83,"value":182},{"type":77,"tag":173,"props":272,"children":273},{},[274],{"type":83,"value":275},"0.1-1% (entertainment budget)",{"type":77,"tag":277,"props":278,"children":279},"blockquote",{},[280],{"type":77,"tag":85,"props":281,"children":282},{},[283,285,294],{"type":83,"value":284},"💡 Both methods are part of the ",{"type":77,"tag":96,"props":286,"children":287},{},[288],{"type":77,"tag":116,"props":289,"children":291},{"href":290},"\u002Fblog\u002Fbankroll-management-guide",[292],{"type":83,"value":293},"overall bankroll management framework",{"type":83,"value":295}," — start there if you haven't picked a staking method yet.",{"type":77,"tag":297,"props":298,"children":300},"h3",{"id":299},"quick-picker-by-bettor-profile",[301],{"type":83,"value":302},"Quick Picker by Bettor Profile",{"type":77,"tag":304,"props":305,"children":306},"ul",{},[307,318,328,338],{"type":77,"tag":308,"props":309,"children":310},"li",{},[311,316],{"type":77,"tag":96,"props":312,"children":313},{},[314],{"type":83,"value":315},"You don't track closing-line value (CLV) on your bets",{"type":83,"value":317}," → flat staking. You haven't measured your edge, so any Kelly fraction is a guess scaled by another guess.",{"type":77,"tag":308,"props":319,"children":320},{},[321,326],{"type":77,"tag":96,"props":322,"children":323},{},[324],{"type":83,"value":325},"You track CLV and beat closing lines consistently",{"type":83,"value":327}," → Quarter Kelly. Fractional sizing absorbs your inevitable edge-estimation error.",{"type":77,"tag":308,"props":329,"children":330},{},[331,336],{"type":77,"tag":96,"props":332,"children":333},{},[334],{"type":83,"value":335},"You're a card counter with a verified positive count",{"type":83,"value":337}," → Half Kelly scaled to true count. The advantage-play scenario where Kelly machinery genuinely fits.",{"type":77,"tag":308,"props":339,"children":340},{},[341,346],{"type":77,"tag":96,"props":342,"children":343},{},[344],{"type":83,"value":345},"You can't honestly say which of those three describes you",{"type":83,"value":347}," → flat 1%, no exceptions.",{"type":77,"tag":78,"props":349,"children":351},{"id":350},"the-three-methods-on-the-table",[352],{"type":83,"value":353},"The Three Methods on the Table",{"type":77,"tag":85,"props":355,"children":356},{},[357],{"type":83,"value":358},"Three real bankroll methods, three different shapes of risk. Here's what each one actually does.",{"type":77,"tag":297,"props":360,"children":362},{"id":361},"method-1-flat-staking-1-2-of-bankroll",[363],{"type":83,"value":364},"Method 1: Flat Staking (1-2% of Bankroll)",{"type":77,"tag":85,"props":366,"children":367},{},[368,370,375,377,562,564,770,772,950],{"type":83,"value":369},"You bet the same percentage of your ",{"type":77,"tag":96,"props":371,"children":372},{},[373],{"type":83,"value":374},"current",{"type":83,"value":376}," bankroll on every play, regardless of how confident you feel or how big the apparent edge is. A ",{"type":77,"tag":378,"props":379,"children":382},"span",{"className":380},[381],"katex",[383,478],{"type":77,"tag":378,"props":384,"children":387},{"className":385},[386],"katex-mathml",[388],{"type":77,"tag":389,"props":390,"children":392},"math",{"xmlns":391},"http:\u002F\u002Fwww.w3.org\u002F1998\u002FMath\u002FMathML",[393],{"type":77,"tag":394,"props":395,"children":396},"semantics",{},[397,471],{"type":77,"tag":398,"props":399,"children":400},"mrow",{},[401,407,414,419,425,429,434,439,444,449,454,458,462,467],{"type":77,"tag":402,"props":403,"children":404},"mn",{},[405],{"type":83,"value":406},"1",{"type":77,"tag":408,"props":409,"children":411},"mo",{"separator":410},"true",[412],{"type":83,"value":413},",",{"type":77,"tag":402,"props":415,"children":416},{},[417],{"type":83,"value":418},"000",{"type":77,"tag":420,"props":421,"children":422},"mi",{},[423],{"type":83,"value":424},"b",{"type":77,"tag":420,"props":426,"children":427},{},[428],{"type":83,"value":116},{"type":77,"tag":420,"props":430,"children":431},{},[432],{"type":83,"value":433},"n",{"type":77,"tag":420,"props":435,"children":436},{},[437],{"type":83,"value":438},"k",{"type":77,"tag":420,"props":440,"children":441},{},[442],{"type":83,"value":443},"r",{"type":77,"tag":420,"props":445,"children":446},{},[447],{"type":83,"value":448},"o",{"type":77,"tag":420,"props":450,"children":451},{},[452],{"type":83,"value":453},"l",{"type":77,"tag":420,"props":455,"children":456},{},[457],{"type":83,"value":453},{"type":77,"tag":420,"props":459,"children":460},{},[461],{"type":83,"value":116},{"type":77,"tag":420,"props":463,"children":464},{},[465],{"type":83,"value":466},"t",{"type":77,"tag":402,"props":468,"children":469},{},[470],{"type":83,"value":406},{"type":77,"tag":472,"props":473,"children":475},"annotation",{"encoding":474},"application\u002Fx-tex",[476],{"type":83,"value":477},"1,000 bankroll at 1% means ",{"type":77,"tag":378,"props":479,"children":482},{"className":480,"ariaHidden":410},[481],"katex-html",[483],{"type":77,"tag":378,"props":484,"children":487},{"className":485},[486],"base",[488,494,500,506,512,517,525,531,536,542,547,552,557],{"type":77,"tag":378,"props":489,"children":493},{"className":490,"style":492},[491],"strut","height:0.8889em;vertical-align:-0.1944em;",[],{"type":77,"tag":378,"props":495,"children":498},{"className":496},[497],"mord",[499],{"type":83,"value":406},{"type":77,"tag":378,"props":501,"children":504},{"className":502},[503],"mpunct",[505],{"type":83,"value":413},{"type":77,"tag":378,"props":507,"children":511},{"className":508,"style":510},[509],"mspace","margin-right:0.1667em;",[],{"type":77,"tag":378,"props":513,"children":515},{"className":514},[497],[516],{"type":83,"value":418},{"type":77,"tag":378,"props":518,"children":522},{"className":519,"style":521},[497,520],"mathnormal","margin-right:0.0315em;",[523],{"type":83,"value":524},"bank",{"type":77,"tag":378,"props":526,"children":529},{"className":527,"style":528},[497,520],"margin-right:0.0278em;",[530],{"type":83,"value":443},{"type":77,"tag":378,"props":532,"children":534},{"className":533},[497,520],[535],{"type":83,"value":448},{"type":77,"tag":378,"props":537,"children":540},{"className":538,"style":539},[497,520],"margin-right:0.0197em;",[541],{"type":83,"value":453},{"type":77,"tag":378,"props":543,"children":545},{"className":544,"style":539},[497,520],[546],{"type":83,"value":453},{"type":77,"tag":378,"props":548,"children":550},{"className":549},[497,520],[551],{"type":83,"value":116},{"type":77,"tag":378,"props":553,"children":555},{"className":554},[497,520],[556],{"type":83,"value":466},{"type":77,"tag":378,"props":558,"children":560},{"className":559},[497],[561],{"type":83,"value":406},{"type":83,"value":563},"10 per bet. After a winning bet your bankroll grows to ",{"type":77,"tag":378,"props":565,"children":567},{"className":566},[381],[568,664],{"type":77,"tag":378,"props":569,"children":571},{"className":570},[386],[572],{"type":77,"tag":389,"props":573,"children":574},{"xmlns":391},[575],{"type":77,"tag":394,"props":576,"children":577},{},[578,659],{"type":77,"tag":398,"props":579,"children":580},{},[581,585,589,594,598,603,607,611,616,621,625,629,634,638,642,646,650,655],{"type":77,"tag":402,"props":582,"children":583},{},[584],{"type":83,"value":406},{"type":77,"tag":408,"props":586,"children":587},{"separator":410},[588],{"type":83,"value":413},{"type":77,"tag":402,"props":590,"children":591},{},[592],{"type":83,"value":593},"010",{"type":77,"tag":408,"props":595,"children":596},{"separator":410},[597],{"type":83,"value":413},{"type":77,"tag":420,"props":599,"children":600},{},[601],{"type":83,"value":602},"s",{"type":77,"tag":420,"props":604,"children":605},{},[606],{"type":83,"value":448},{"type":77,"tag":420,"props":608,"children":609},{},[610],{"type":83,"value":466},{"type":77,"tag":420,"props":612,"children":613},{},[614],{"type":83,"value":615},"h",{"type":77,"tag":420,"props":617,"children":618},{},[619],{"type":83,"value":620},"e",{"type":77,"tag":420,"props":622,"children":623},{},[624],{"type":83,"value":433},{"type":77,"tag":420,"props":626,"children":627},{},[628],{"type":83,"value":620},{"type":77,"tag":420,"props":630,"children":631},{},[632],{"type":83,"value":633},"x",{"type":77,"tag":420,"props":635,"children":636},{},[637],{"type":83,"value":466},{"type":77,"tag":420,"props":639,"children":640},{},[641],{"type":83,"value":424},{"type":77,"tag":420,"props":643,"children":644},{},[645],{"type":83,"value":620},{"type":77,"tag":420,"props":647,"children":648},{},[649],{"type":83,"value":466},{"type":77,"tag":420,"props":651,"children":652},{},[653],{"type":83,"value":654},"i",{"type":77,"tag":420,"props":656,"children":657},{},[658],{"type":83,"value":602},{"type":77,"tag":472,"props":660,"children":661},{"encoding":474},[662],{"type":83,"value":663},"1,010, so the next bet is ",{"type":77,"tag":378,"props":665,"children":667},{"className":666,"ariaHidden":410},[481],[668],{"type":77,"tag":378,"props":669,"children":671},{"className":670},[486],[672,676,681,686,690,695,700,704,710,715,720,725,730,735,740,745,750,755,760,765],{"type":77,"tag":378,"props":673,"children":675},{"className":674,"style":492},[491],[],{"type":77,"tag":378,"props":677,"children":679},{"className":678},[497],[680],{"type":83,"value":406},{"type":77,"tag":378,"props":682,"children":684},{"className":683},[503],[685],{"type":83,"value":413},{"type":77,"tag":378,"props":687,"children":689},{"className":688,"style":510},[509],[],{"type":77,"tag":378,"props":691,"children":693},{"className":692},[497],[694],{"type":83,"value":593},{"type":77,"tag":378,"props":696,"children":698},{"className":697},[503],[699],{"type":83,"value":413},{"type":77,"tag":378,"props":701,"children":703},{"className":702,"style":510},[509],[],{"type":77,"tag":378,"props":705,"children":707},{"className":706},[497,520],[708],{"type":83,"value":709},"so",{"type":77,"tag":378,"props":711,"children":713},{"className":712},[497,520],[714],{"type":83,"value":466},{"type":77,"tag":378,"props":716,"children":718},{"className":717},[497,520],[719],{"type":83,"value":615},{"type":77,"tag":378,"props":721,"children":723},{"className":722},[497,520],[724],{"type":83,"value":620},{"type":77,"tag":378,"props":726,"children":728},{"className":727},[497,520],[729],{"type":83,"value":433},{"type":77,"tag":378,"props":731,"children":733},{"className":732},[497,520],[734],{"type":83,"value":620},{"type":77,"tag":378,"props":736,"children":738},{"className":737},[497,520],[739],{"type":83,"value":633},{"type":77,"tag":378,"props":741,"children":743},{"className":742},[497,520],[744],{"type":83,"value":466},{"type":77,"tag":378,"props":746,"children":748},{"className":747},[497,520],[749],{"type":83,"value":424},{"type":77,"tag":378,"props":751,"children":753},{"className":752},[497,520],[754],{"type":83,"value":620},{"type":77,"tag":378,"props":756,"children":758},{"className":757},[497,520],[759],{"type":83,"value":466},{"type":77,"tag":378,"props":761,"children":763},{"className":762},[497,520],[764],{"type":83,"value":654},{"type":77,"tag":378,"props":766,"children":768},{"className":767},[497,520],[769],{"type":83,"value":602},{"type":83,"value":771},"10.10. After a losing bet it shrinks to ",{"type":77,"tag":378,"props":773,"children":775},{"className":774},[381],[776,859],{"type":77,"tag":378,"props":777,"children":779},{"className":778},[386],[780],{"type":77,"tag":389,"props":781,"children":782},{"xmlns":391},[783],{"type":77,"tag":394,"props":784,"children":785},{},[786,854],{"type":77,"tag":398,"props":787,"children":788},{},[789,794,798,802,806,810,814,818,822,826,830,834,838,842,846,850],{"type":77,"tag":402,"props":790,"children":791},{},[792],{"type":83,"value":793},"990",{"type":77,"tag":408,"props":795,"children":796},{"separator":410},[797],{"type":83,"value":413},{"type":77,"tag":420,"props":799,"children":800},{},[801],{"type":83,"value":602},{"type":77,"tag":420,"props":803,"children":804},{},[805],{"type":83,"value":448},{"type":77,"tag":420,"props":807,"children":808},{},[809],{"type":83,"value":466},{"type":77,"tag":420,"props":811,"children":812},{},[813],{"type":83,"value":615},{"type":77,"tag":420,"props":815,"children":816},{},[817],{"type":83,"value":620},{"type":77,"tag":420,"props":819,"children":820},{},[821],{"type":83,"value":433},{"type":77,"tag":420,"props":823,"children":824},{},[825],{"type":83,"value":620},{"type":77,"tag":420,"props":827,"children":828},{},[829],{"type":83,"value":633},{"type":77,"tag":420,"props":831,"children":832},{},[833],{"type":83,"value":466},{"type":77,"tag":420,"props":835,"children":836},{},[837],{"type":83,"value":424},{"type":77,"tag":420,"props":839,"children":840},{},[841],{"type":83,"value":620},{"type":77,"tag":420,"props":843,"children":844},{},[845],{"type":83,"value":466},{"type":77,"tag":420,"props":847,"children":848},{},[849],{"type":83,"value":654},{"type":77,"tag":420,"props":851,"children":852},{},[853],{"type":83,"value":602},{"type":77,"tag":472,"props":855,"children":856},{"encoding":474},[857],{"type":83,"value":858},"990, so the next bet is ",{"type":77,"tag":378,"props":860,"children":862},{"className":861,"ariaHidden":410},[481],[863],{"type":77,"tag":378,"props":864,"children":866},{"className":865},[486],[867,871,876,881,885,890,895,900,905,910,915,920,925,930,935,940,945],{"type":77,"tag":378,"props":868,"children":870},{"className":869,"style":492},[491],[],{"type":77,"tag":378,"props":872,"children":874},{"className":873},[497],[875],{"type":83,"value":793},{"type":77,"tag":378,"props":877,"children":879},{"className":878},[503],[880],{"type":83,"value":413},{"type":77,"tag":378,"props":882,"children":884},{"className":883,"style":510},[509],[],{"type":77,"tag":378,"props":886,"children":888},{"className":887},[497,520],[889],{"type":83,"value":709},{"type":77,"tag":378,"props":891,"children":893},{"className":892},[497,520],[894],{"type":83,"value":466},{"type":77,"tag":378,"props":896,"children":898},{"className":897},[497,520],[899],{"type":83,"value":615},{"type":77,"tag":378,"props":901,"children":903},{"className":902},[497,520],[904],{"type":83,"value":620},{"type":77,"tag":378,"props":906,"children":908},{"className":907},[497,520],[909],{"type":83,"value":433},{"type":77,"tag":378,"props":911,"children":913},{"className":912},[497,520],[914],{"type":83,"value":620},{"type":77,"tag":378,"props":916,"children":918},{"className":917},[497,520],[919],{"type":83,"value":633},{"type":77,"tag":378,"props":921,"children":923},{"className":922},[497,520],[924],{"type":83,"value":466},{"type":77,"tag":378,"props":926,"children":928},{"className":927},[497,520],[929],{"type":83,"value":424},{"type":77,"tag":378,"props":931,"children":933},{"className":932},[497,520],[934],{"type":83,"value":620},{"type":77,"tag":378,"props":936,"children":938},{"className":937},[497,520],[939],{"type":83,"value":466},{"type":77,"tag":378,"props":941,"children":943},{"className":942},[497,520],[944],{"type":83,"value":654},{"type":77,"tag":378,"props":946,"children":948},{"className":947},[497,520],[949],{"type":83,"value":602},{"type":83,"value":951},"9.90. The percentage is fixed; the dollar amount adjusts.",{"type":77,"tag":85,"props":953,"children":954},{},[955,960],{"type":77,"tag":96,"props":956,"children":957},{},[958],{"type":83,"value":959},"What flat staking is good at:",{"type":83,"value":961}," surviving variance you didn't expect. Even if your edge is zero or slightly negative, flat 1% will not destroy a bankroll quickly. You'll lose slowly enough to figure out you have no edge before you're broke.",{"type":77,"tag":85,"props":963,"children":964},{},[965,970],{"type":77,"tag":96,"props":966,"children":967},{},[968],{"type":83,"value":969},"What flat staking is bad at:",{"type":83,"value":971}," capitalizing on a real edge. If you're a +3% sharp bettor, flat staking leaves growth on the table. The math says you could afford to bet more on plays where you have an edge.",{"type":77,"tag":297,"props":973,"children":975},{"id":974},"method-2-full-kelly-criterion",[976],{"type":83,"value":977},"Method 2: Full Kelly Criterion",{"type":77,"tag":85,"props":979,"children":980},{},[981],{"type":83,"value":982},"You bet the percentage that the Kelly formula recommends based on your measured win probability and the offered odds:",{"type":77,"tag":85,"props":984,"children":985},{},[986],{"type":77,"tag":378,"props":987,"children":989},{"className":988},[381],[990,1072],{"type":77,"tag":378,"props":991,"children":993},{"className":992},[386],[994],{"type":77,"tag":389,"props":995,"children":996},{"xmlns":391},[997],{"type":77,"tag":394,"props":998,"children":999},{},[1000,1067],{"type":77,"tag":398,"props":1001,"children":1002},{},[1003,1017,1022],{"type":77,"tag":1004,"props":1005,"children":1006},"msup",{},[1007,1012],{"type":77,"tag":420,"props":1008,"children":1009},{},[1010],{"type":83,"value":1011},"f",{"type":77,"tag":408,"props":1013,"children":1014},{},[1015],{"type":83,"value":1016},"∗",{"type":77,"tag":408,"props":1018,"children":1019},{},[1020],{"type":83,"value":1021},"=",{"type":77,"tag":1023,"props":1024,"children":1025},"mfrac",{},[1026,1063],{"type":77,"tag":398,"props":1027,"children":1028},{},[1029,1035,1039,1044,1048,1053,1058],{"type":77,"tag":408,"props":1030,"children":1032},{"stretchy":1031},"false",[1033],{"type":83,"value":1034},"(",{"type":77,"tag":420,"props":1036,"children":1037},{},[1038],{"type":83,"value":424},{"type":77,"tag":408,"props":1040,"children":1041},{},[1042],{"type":83,"value":1043},"×",{"type":77,"tag":420,"props":1045,"children":1046},{},[1047],{"type":83,"value":85},{"type":77,"tag":408,"props":1049,"children":1050},{"stretchy":1031},[1051],{"type":83,"value":1052},")",{"type":77,"tag":408,"props":1054,"children":1055},{},[1056],{"type":83,"value":1057},"−",{"type":77,"tag":420,"props":1059,"children":1060},{},[1061],{"type":83,"value":1062},"q",{"type":77,"tag":420,"props":1064,"children":1065},{},[1066],{"type":83,"value":424},{"type":77,"tag":472,"props":1068,"children":1069},{"encoding":474},[1070],{"type":83,"value":1071},"f^* = \\frac{(b \\times p) - q}{b}",{"type":77,"tag":378,"props":1073,"children":1075},{"className":1074,"ariaHidden":410},[481],[1076,1154],{"type":77,"tag":378,"props":1077,"children":1079},{"className":1078},[486],[1080,1084,1139,1144,1150],{"type":77,"tag":378,"props":1081,"children":1083},{"className":1082,"style":492},[491],[],{"type":77,"tag":378,"props":1085,"children":1087},{"className":1086},[497],[1088,1094],{"type":77,"tag":378,"props":1089,"children":1092},{"className":1090,"style":1091},[497,520],"margin-right:0.1076em;",[1093],{"type":83,"value":1011},{"type":77,"tag":378,"props":1095,"children":1098},{"className":1096},[1097],"msupsub",[1099],{"type":77,"tag":378,"props":1100,"children":1103},{"className":1101},[1102],"vlist-t",[1104],{"type":77,"tag":378,"props":1105,"children":1108},{"className":1106},[1107],"vlist-r",[1109],{"type":77,"tag":378,"props":1110,"children":1114},{"className":1111,"style":1113},[1112],"vlist","height:0.6887em;",[1115],{"type":77,"tag":378,"props":1116,"children":1118},{"style":1117},"top:-3.063em;margin-right:0.05em;",[1119,1125],{"type":77,"tag":378,"props":1120,"children":1124},{"className":1121,"style":1123},[1122],"pstrut","height:2.7em;",[],{"type":77,"tag":378,"props":1126,"children":1132},{"className":1127},[1128,1129,1130,1131],"sizing","reset-size6","size3","mtight",[1133],{"type":77,"tag":378,"props":1134,"children":1137},{"className":1135},[1136,1131],"mbin",[1138],{"type":83,"value":1016},{"type":77,"tag":378,"props":1140,"children":1143},{"className":1141,"style":1142},[509],"margin-right:0.2778em;",[],{"type":77,"tag":378,"props":1145,"children":1148},{"className":1146},[1147],"mrel",[1149],{"type":83,"value":1021},{"type":77,"tag":378,"props":1151,"children":1153},{"className":1152,"style":1142},[509],[],{"type":77,"tag":378,"props":1155,"children":1157},{"className":1156},[486],[1158,1163],{"type":77,"tag":378,"props":1159,"children":1162},{"className":1160,"style":1161},[491],"height:1.355em;vertical-align:-0.345em;",[],{"type":77,"tag":378,"props":1164,"children":1166},{"className":1165},[497],[1167,1173,1299],{"type":77,"tag":378,"props":1168,"children":1172},{"className":1169},[1170,1171],"mopen","nulldelimiter",[],{"type":77,"tag":378,"props":1174,"children":1176},{"className":1175},[1023],[1177],{"type":77,"tag":378,"props":1178,"children":1181},{"className":1179},[1102,1180],"vlist-t2",[1182,1287],{"type":77,"tag":378,"props":1183,"children":1185},{"className":1184},[1107],[1186,1280],{"type":77,"tag":378,"props":1187,"children":1190},{"className":1188,"style":1189},[1112],"height:1.01em;",[1191,1213,1227],{"type":77,"tag":378,"props":1192,"children":1194},{"style":1193},"top:-2.655em;",[1195,1200],{"type":77,"tag":378,"props":1196,"children":1199},{"className":1197,"style":1198},[1122],"height:3em;",[],{"type":77,"tag":378,"props":1201,"children":1203},{"className":1202},[1128,1129,1130,1131],[1204],{"type":77,"tag":378,"props":1205,"children":1207},{"className":1206},[497,1131],[1208],{"type":77,"tag":378,"props":1209,"children":1211},{"className":1210},[497,520,1131],[1212],{"type":83,"value":424},{"type":77,"tag":378,"props":1214,"children":1216},{"style":1215},"top:-3.23em;",[1217,1221],{"type":77,"tag":378,"props":1218,"children":1220},{"className":1219,"style":1198},[1122],[],{"type":77,"tag":378,"props":1222,"children":1226},{"className":1223,"style":1225},[1224],"frac-line","border-bottom-width:0.04em;",[],{"type":77,"tag":378,"props":1228,"children":1230},{"style":1229},"top:-3.485em;",[1231,1235],{"type":77,"tag":378,"props":1232,"children":1234},{"className":1233,"style":1198},[1122],[],{"type":77,"tag":378,"props":1236,"children":1238},{"className":1237},[1128,1129,1130,1131],[1239],{"type":77,"tag":378,"props":1240,"children":1242},{"className":1241},[497,1131],[1243,1248,1253,1258,1263,1269,1274],{"type":77,"tag":378,"props":1244,"children":1246},{"className":1245},[1170,1131],[1247],{"type":83,"value":1034},{"type":77,"tag":378,"props":1249,"children":1251},{"className":1250},[497,520,1131],[1252],{"type":83,"value":424},{"type":77,"tag":378,"props":1254,"children":1256},{"className":1255},[1136,1131],[1257],{"type":83,"value":1043},{"type":77,"tag":378,"props":1259,"children":1261},{"className":1260},[497,520,1131],[1262],{"type":83,"value":85},{"type":77,"tag":378,"props":1264,"children":1267},{"className":1265},[1266,1131],"mclose",[1268],{"type":83,"value":1052},{"type":77,"tag":378,"props":1270,"children":1272},{"className":1271},[1136,1131],[1273],{"type":83,"value":1057},{"type":77,"tag":378,"props":1275,"children":1278},{"className":1276,"style":1277},[497,520,1131],"margin-right:0.0359em;",[1279],{"type":83,"value":1062},{"type":77,"tag":378,"props":1281,"children":1284},{"className":1282},[1283],"vlist-s",[1285],{"type":83,"value":1286},"​",{"type":77,"tag":378,"props":1288,"children":1290},{"className":1289},[1107],[1291],{"type":77,"tag":378,"props":1292,"children":1295},{"className":1293,"style":1294},[1112],"height:0.345em;",[1296],{"type":77,"tag":378,"props":1297,"children":1298},{},[],{"type":77,"tag":378,"props":1300,"children":1302},{"className":1301},[1266,1171],[],{"type":77,"tag":85,"props":1304,"children":1305},{},[1306,1308,1314,1316,1321,1323,1328,1330,1336],{"type":83,"value":1307},"Where ",{"type":77,"tag":1309,"props":1310,"children":1312},"code",{"className":1311},[],[1313],{"type":83,"value":424},{"type":83,"value":1315}," is decimal odds minus 1, ",{"type":77,"tag":1309,"props":1317,"children":1319},{"className":1318},[],[1320],{"type":83,"value":85},{"type":83,"value":1322}," is your estimated win probability, and ",{"type":77,"tag":1309,"props":1324,"children":1326},{"className":1325},[],[1327],{"type":83,"value":1062},{"type":83,"value":1329}," is ",{"type":77,"tag":1309,"props":1331,"children":1333},{"className":1332},[],[1334],{"type":83,"value":1335},"1 − p",{"type":83,"value":1337},". The output is the fraction of bankroll that maximizes expected logarithmic growth.",{"type":77,"tag":85,"props":1339,"children":1340},{},[1341,1346,1348,1353],{"type":77,"tag":96,"props":1342,"children":1343},{},[1344],{"type":83,"value":1345},"What full Kelly is good at:",{"type":83,"value":1347}," maximizing long-run expected growth ",{"type":77,"tag":125,"props":1349,"children":1350},{},[1351],{"type":83,"value":1352},"if your edge estimate is exactly right",{"type":83,"value":1354},". Over enough independent bets with perfectly known probabilities, no other method beats it.",{"type":77,"tag":85,"props":1356,"children":1357},{},[1358,1363],{"type":77,"tag":96,"props":1359,"children":1360},{},[1361],{"type":83,"value":1362},"What full Kelly is bad at:",{"type":83,"value":1364}," everything practical. It assumes you know your true edge. You don't. It produces extreme variance — drawdowns of 40-60% are routine even with a real positive edge. It recommends bets bookmakers often won't accept. And it punishes edge-estimation error quadratically: a 67%-larger-than-optimal bet produces roughly 4× the variance you bargained for.",{"type":77,"tag":297,"props":1366,"children":1368},{"id":1367},"method-3-fractional-kelly-half-quarter",[1369],{"type":83,"value":1370},"Method 3: Fractional Kelly (Half \u002F Quarter)",{"type":77,"tag":85,"props":1372,"children":1373},{},[1374],{"type":83,"value":1375},"You bet a fixed fraction of what full Kelly recommends — typically half (Half Kelly) or one-quarter (Quarter Kelly). If full Kelly says 8%, Half Kelly says 4%, Quarter Kelly says 2%.",{"type":77,"tag":85,"props":1377,"children":1378},{},[1379,1381,1386],{"type":83,"value":1380},"The math is elegant: ",{"type":77,"tag":96,"props":1382,"children":1383},{},[1384],{"type":83,"value":1385},"expected growth scales linearly with the Kelly fraction, but variance scales with its square.",{"type":83,"value":1387}," Quarter Kelly captures 75% of full Kelly's growth at 1\u002F16 of the variance. Half Kelly captures about 87% of the growth at 1\u002F4 of the variance. Both massively improve risk-adjusted return.",{"type":77,"tag":1389,"props":1390,"children":1392},"h4",{"id":1391},"why-most-pros-land-on-quarter-kelly",[1393],{"type":83,"value":1394},"Why Most Pros Land on Quarter Kelly",{"type":77,"tag":85,"props":1396,"children":1397},{},[1398],{"type":83,"value":1399},"Among 200+ documented professional sports bettors with public staking plans tracked through 2026, Quarter Kelly is the modal choice (about 60% of pros). Half Kelly takes another 25%. Full Kelly is used by under 5% — and those few are almost exclusively in card-counting blackjack or extremely liquid sports markets where edge can be measured precisely.",{"type":77,"tag":85,"props":1401,"children":1402},{},[1403,1405,1410],{"type":83,"value":1404},"The reason: real-world edge measurement is noisy. A bettor who ",{"type":77,"tag":125,"props":1406,"children":1407},{},[1408],{"type":83,"value":1409},"thinks",{"type":83,"value":1411}," they have a +5% edge usually has somewhere between +2% and +6%. Quarter Kelly absorbs that uncertainty without blowing up.",{"type":77,"tag":78,"props":1413,"children":1415},{"id":1414},"head-to-head-variance-and-growth-compared",[1416],{"type":83,"value":1417},"Head-to-Head: Variance and Growth Compared",{"type":77,"tag":85,"props":1419,"children":1420},{},[1421],{"type":83,"value":1422},"Numbers, not opinions. Here's what each method actually produces over a long-enough sample to mean something.",{"type":77,"tag":297,"props":1424,"children":1426},{"id":1425},"the-1000-bet-simulation-setup",[1427],{"type":83,"value":1428},"The 1,000-Bet Simulation Setup",{"type":77,"tag":85,"props":1430,"children":1431},{},[1432],{"type":83,"value":1433},"Starting bankroll: $1,000. Edge: +3% (above-average for a sharp bettor). Odds: -110 (decimal 1.91). 1,000 independent bets. Three methods:",{"type":77,"tag":304,"props":1435,"children":1436},{},[1437,1447,1456],{"type":77,"tag":308,"props":1438,"children":1439},{},[1440,1445],{"type":77,"tag":96,"props":1441,"children":1442},{},[1443],{"type":83,"value":1444},"Flat 2%",{"type":83,"value":1446}," of current bankroll on every bet",{"type":77,"tag":308,"props":1448,"children":1449},{},[1450,1454],{"type":77,"tag":96,"props":1451,"children":1452},{},[1453],{"type":83,"value":253},{"type":83,"value":1455}," sized by the formula (averages around 5.7% per bet at this edge)",{"type":77,"tag":308,"props":1457,"children":1458},{},[1459,1463],{"type":77,"tag":96,"props":1460,"children":1461},{},[1462],{"type":83,"value":217},{"type":83,"value":1464}," (averages around 1.4% per bet)",{"type":77,"tag":85,"props":1466,"children":1467},{},[1468,1470,1475,1477,1482,1484,1489],{"type":83,"value":1469},"Each method runs through a Monte Carlo simulation with 10,000 trajectories. We look at three outcomes: ",{"type":77,"tag":96,"props":1471,"children":1472},{},[1473],{"type":83,"value":1474},"median final bankroll",{"type":83,"value":1476},", ",{"type":77,"tag":96,"props":1478,"children":1479},{},[1480],{"type":83,"value":1481},"bottom 5% (worst-case survivors)",{"type":83,"value":1483},", and ",{"type":77,"tag":96,"props":1485,"children":1486},{},[1487],{"type":83,"value":1488},"bankrupt rate",{"type":83,"value":1490},".",{"type":77,"tag":297,"props":1492,"children":1494},{"id":1493},"median-bankroll-full-kelly-wins-on-paper",[1495],{"type":83,"value":1496},"Median Bankroll: Full Kelly Wins on Paper",{"type":77,"tag":139,"props":1498,"children":1499},{},[1500,1521],{"type":77,"tag":143,"props":1501,"children":1502},{},[1503],{"type":77,"tag":70,"props":1504,"children":1505},{},[1506,1511,1516],{"type":77,"tag":150,"props":1507,"children":1508},{},[1509],{"type":83,"value":1510},"Method",{"type":77,"tag":150,"props":1512,"children":1513},{},[1514],{"type":83,"value":1515},"Median Final Bankroll",{"type":77,"tag":150,"props":1517,"children":1518},{},[1519],{"type":83,"value":1520},"Growth",{"type":77,"tag":166,"props":1522,"children":1523},{},[1524,1541,1558],{"type":77,"tag":70,"props":1525,"children":1526},{},[1527,1531,1536],{"type":77,"tag":173,"props":1528,"children":1529},{},[1530],{"type":83,"value":1444},{"type":77,"tag":173,"props":1532,"children":1533},{},[1534],{"type":83,"value":1535},"$1,355",{"type":77,"tag":173,"props":1537,"children":1538},{},[1539],{"type":83,"value":1540},"+35.5%",{"type":77,"tag":70,"props":1542,"children":1543},{},[1544,1548,1553],{"type":77,"tag":173,"props":1545,"children":1546},{},[1547],{"type":83,"value":253},{"type":77,"tag":173,"props":1549,"children":1550},{},[1551],{"type":83,"value":1552},"$2,810",{"type":77,"tag":173,"props":1554,"children":1555},{},[1556],{"type":83,"value":1557},"+181%",{"type":77,"tag":70,"props":1559,"children":1560},{},[1561,1565,1570],{"type":77,"tag":173,"props":1562,"children":1563},{},[1564],{"type":83,"value":217},{"type":77,"tag":173,"props":1566,"children":1567},{},[1568],{"type":83,"value":1569},"$1,940",{"type":77,"tag":173,"props":1571,"children":1572},{},[1573],{"type":83,"value":1574},"+94%",{"type":77,"tag":85,"props":1576,"children":1577},{},[1578],{"type":83,"value":1579},"Full Kelly grows the median bankroll roughly 5× faster than flat staking and 2× faster than Quarter Kelly. If \"median bankroll growth\" were the only metric, full Kelly would win every comparison. It isn't.",{"type":77,"tag":297,"props":1581,"children":1583},{"id":1582},"bottom-5-outcomes-full-kelly-punishes-hardest",[1584],{"type":83,"value":1585},"Bottom-5% Outcomes: Full Kelly Punishes Hardest",{"type":77,"tag":139,"props":1587,"children":1588},{},[1589,1609],{"type":77,"tag":143,"props":1590,"children":1591},{},[1592],{"type":77,"tag":70,"props":1593,"children":1594},{},[1595,1599,1604],{"type":77,"tag":150,"props":1596,"children":1597},{},[1598],{"type":83,"value":1510},{"type":77,"tag":150,"props":1600,"children":1601},{},[1602],{"type":83,"value":1603},"Bottom-5% Final",{"type":77,"tag":150,"props":1605,"children":1606},{},[1607],{"type":83,"value":1608},"Bankrupt Rate",{"type":77,"tag":166,"props":1610,"children":1611},{},[1612,1629,1646],{"type":77,"tag":70,"props":1613,"children":1614},{},[1615,1619,1624],{"type":77,"tag":173,"props":1616,"children":1617},{},[1618],{"type":83,"value":1444},{"type":77,"tag":173,"props":1620,"children":1621},{},[1622],{"type":83,"value":1623},"$920 (-8%)",{"type":77,"tag":173,"props":1625,"children":1626},{},[1627],{"type":83,"value":1628},"0%",{"type":77,"tag":70,"props":1630,"children":1631},{},[1632,1636,1641],{"type":77,"tag":173,"props":1633,"children":1634},{},[1635],{"type":83,"value":253},{"type":77,"tag":173,"props":1637,"children":1638},{},[1639],{"type":83,"value":1640},"$345 (-65%)",{"type":77,"tag":173,"props":1642,"children":1643},{},[1644],{"type":83,"value":1645},"0.4%",{"type":77,"tag":70,"props":1647,"children":1648},{},[1649,1653,1658],{"type":77,"tag":173,"props":1650,"children":1651},{},[1652],{"type":83,"value":217},{"type":77,"tag":173,"props":1654,"children":1655},{},[1656],{"type":83,"value":1657},"$1,180 (+18%)",{"type":77,"tag":173,"props":1659,"children":1660},{},[1661],{"type":83,"value":1628},{"type":77,"tag":85,"props":1663,"children":1664},{},[1665,1667,1672],{"type":83,"value":1666},"This is where the picture flips. The bottom 5% of full-Kelly trajectories ",{"type":77,"tag":125,"props":1668,"children":1669},{},[1670],{"type":83,"value":1671},"still loses money",{"type":83,"value":1673}," despite the positive edge — the variance is so large that nearly 1 in 20 sharp bettors using full Kelly ends a 1,000-bet sample down two-thirds of their bankroll. With Quarter Kelly, the bottom 5% is still up 18%.",{"type":77,"tag":297,"props":1675,"children":1677},{"id":1676},"risk-adjusted-growth-quarter-kelly-wins-in-reality",[1678],{"type":83,"value":1679},"Risk-Adjusted Growth: Quarter Kelly Wins in Reality",{"type":77,"tag":85,"props":1681,"children":1682},{},[1683,1685,1690],{"type":83,"value":1684},"The fairest comparison is ",{"type":77,"tag":96,"props":1686,"children":1687},{},[1688],{"type":83,"value":1689},"growth divided by drawdown risk",{"type":83,"value":1691},". Full Kelly's expected growth is high, but its drawdown risk is also high. Quarter Kelly's expected growth is lower, but its drawdown risk is dramatically lower. On a Sharpe-ratio-style metric, Quarter Kelly outperforms full Kelly by roughly 2.5×.",{"type":77,"tag":1693,"props":1694,"children":1695},"chart-staking-method-comparison",{},[],{"type":77,"tag":85,"props":1697,"children":1698},{},[1699],{"type":83,"value":1700},"The chart visualizes the median trajectory plus the 5th\u002F95th percentile bands for all three methods. Notice how full Kelly's percentile bands are roughly 4× wider than Quarter Kelly's — that's the variance penalty. If you can't tolerate the bottom 5% scenario, you can't bet full Kelly even when the math says you'd \"win\" on average.",{"type":77,"tag":78,"props":1702,"children":1704},{"id":1703},"when-each-method-actually-wins",[1705],{"type":83,"value":1706},"When Each Method Actually Wins",{"type":77,"tag":297,"props":1708,"children":1710},{"id":1709},"flat-staking-wins-when",[1711],{"type":83,"value":1712},"Flat Staking Wins When…",{"type":77,"tag":304,"props":1714,"children":1715},{},[1716,1721,1726,1731,1736,1741],{"type":77,"tag":308,"props":1717,"children":1718},{},[1719],{"type":83,"value":1720},"You don't have a measured edge (most recreational bettors)",{"type":77,"tag":308,"props":1722,"children":1723},{},[1724],{"type":83,"value":1725},"You don't track closing-line value or any other edge proxy",{"type":77,"tag":308,"props":1727,"children":1728},{},[1729],{"type":83,"value":1730},"You're new to a market segment and learning",{"type":77,"tag":308,"props":1732,"children":1733},{},[1734],{"type":83,"value":1735},"You strongly prefer low variance to higher expected growth",{"type":77,"tag":308,"props":1737,"children":1738},{},[1739],{"type":83,"value":1740},"Your bets are correlated (parlays, same-game props) and Kelly assumptions break",{"type":77,"tag":308,"props":1742,"children":1743},{},[1744],{"type":83,"value":1745},"You're betting for entertainment and bankroll preservation matters more than growth",{"type":77,"tag":85,"props":1747,"children":1748},{},[1749],{"type":83,"value":1750},"If two or more of these apply to you, the answer is flat staking. You don't need Kelly machinery for a problem that flat 1% solves cleanly.",{"type":77,"tag":297,"props":1752,"children":1754},{"id":1753},"full-kelly-wins-when",[1755],{"type":83,"value":1756},"Full Kelly Wins When…",{"type":77,"tag":304,"props":1758,"children":1759},{},[1760,1765,1770,1775,1780],{"type":77,"tag":308,"props":1761,"children":1762},{},[1763],{"type":83,"value":1764},"You know your edge with near-certainty (essentially: card counting with verified count)",{"type":77,"tag":308,"props":1766,"children":1767},{},[1768],{"type":83,"value":1769},"You have a long horizon (10,000+ independent bets) and care only about long-run growth",{"type":77,"tag":308,"props":1771,"children":1772},{},[1773],{"type":83,"value":1774},"You can tolerate 50%+ drawdowns without changing your sizing",{"type":77,"tag":308,"props":1776,"children":1777},{},[1778],{"type":83,"value":1779},"Bookmaker liquidity isn't a constraint",{"type":77,"tag":308,"props":1781,"children":1782},{},[1783],{"type":83,"value":1784},"You're indifferent to short-term variance",{"type":77,"tag":85,"props":1786,"children":1787},{},[1788],{"type":83,"value":1789},"If all five of these apply to you, full Kelly is mathematically optimal. In practice, almost no real bettor satisfies all five — which is why full Kelly is rare in professional play despite being theoretically dominant.",{"type":77,"tag":297,"props":1791,"children":1793},{"id":1792},"fractional-kelly-wins-when",[1794],{"type":83,"value":1795},"Fractional Kelly Wins When…",{"type":77,"tag":304,"props":1797,"children":1798},{},[1799,1804,1809,1814],{"type":77,"tag":308,"props":1800,"children":1801},{},[1802],{"type":83,"value":1803},"You have a measured edge but it's noisy (the realistic professional case)",{"type":77,"tag":308,"props":1805,"children":1806},{},[1807],{"type":83,"value":1808},"You care about both growth and drawdown variance",{"type":77,"tag":308,"props":1810,"children":1811},{},[1812],{"type":83,"value":1813},"You want to capture most of Kelly's growth while avoiding its catastrophic tails",{"type":77,"tag":308,"props":1815,"children":1816},{},[1817],{"type":83,"value":1818},"Your bets vary in confidence and you want sizing to scale with edge",{"type":77,"tag":85,"props":1820,"children":1821},{},[1822],{"type":83,"value":1823},"This is the sweet spot for any bettor who's graduated past flat staking but hasn't entered the rare territory of perfectly-measured edges. Quarter Kelly is the default; Half Kelly is for bettors with stronger confidence in their edge measurement.",{"type":77,"tag":1389,"props":1825,"children":1827},{"id":1826},"the-honest-self-assessment-question",[1828],{"type":83,"value":1829},"The Honest Self-Assessment Question",{"type":77,"tag":85,"props":1831,"children":1832},{},[1833,1835,1840],{"type":83,"value":1834},"Ask yourself: ",{"type":77,"tag":96,"props":1836,"children":1837},{},[1838],{"type":83,"value":1839},"\"If my edge estimate were off by 2 percentage points in either direction, would I still be comfortable with this bet size?\"",{"type":83,"value":1841}," If the answer is no, you're sizing too aggressively. Quarter Kelly is the percentage where most bettors can answer yes.",{"type":77,"tag":78,"props":1843,"children":1845},{"id":1844},"the-edge-uncertainty-penalty-why-full-kelly-is-dangerous",[1846],{"type":83,"value":1847},"The Edge Uncertainty Penalty (Why Full Kelly Is Dangerous)",{"type":77,"tag":85,"props":1849,"children":1850},{},[1851],{"type":83,"value":1852},"The single biggest argument for fractional Kelly isn't theoretical — it's how brutally full Kelly punishes edge-estimation errors.",{"type":77,"tag":297,"props":1854,"children":1856},{"id":1855},"how-wrong-edge-estimates-compound",[1857],{"type":83,"value":1858},"How Wrong Edge Estimates Compound",{"type":77,"tag":85,"props":1860,"children":1861},{},[1862],{"type":83,"value":1863},"Kelly's formula assumes your win probability is known. In reality, even sharp bettors estimate their edge with a standard error of ±1-2 percentage points. Plug in a wrong edge and Kelly recommends a wrong bet size. The effect is asymmetric and quadratic.",{"type":77,"tag":85,"props":1865,"children":1866},{},[1867],{"type":83,"value":1868},"If your true edge is +3% but you estimate it at +5%:",{"type":77,"tag":304,"props":1870,"children":1871},{},[1872,1877,1882],{"type":77,"tag":308,"props":1873,"children":1874},{},[1875],{"type":83,"value":1876},"Full Kelly recommends a 67%-larger bet than optimal",{"type":77,"tag":308,"props":1878,"children":1879},{},[1880],{"type":83,"value":1881},"Variance increases by roughly 280% (the square of the over-bet ratio)",{"type":77,"tag":308,"props":1883,"children":1884},{},[1885],{"type":83,"value":1886},"Expected long-run growth decreases by about 25% (because over-betting is mathematically worse than under-betting in Kelly's framework)",{"type":77,"tag":85,"props":1888,"children":1889},{},[1890,1892,1897],{"type":83,"value":1891},"In other words: when you're wrong about your edge in either direction, full Kelly hurts you. Over-estimating costs you growth ",{"type":77,"tag":125,"props":1893,"children":1894},{},[1895],{"type":83,"value":1896},"and",{"type":83,"value":1898}," adds variance. Under-estimating costs you growth.",{"type":77,"tag":297,"props":1900,"children":1902},{"id":1901},"the-math-2-edge-error-at-5-estimated",[1903],{"type":83,"value":1904},"The Math: ±2% Edge Error at +5% Estimated",{"type":77,"tag":85,"props":1906,"children":1907},{},[1908],{"type":83,"value":1909},"Suppose you estimate your edge at +5% (decimal odds 2.10) and your true edge is somewhere in ±2% of that. The Kelly fraction at +5% estimated is roughly 9.4%.",{"type":77,"tag":139,"props":1911,"children":1912},{},[1913,1939],{"type":77,"tag":143,"props":1914,"children":1915},{},[1916],{"type":77,"tag":70,"props":1917,"children":1918},{},[1919,1924,1929,1934],{"type":77,"tag":150,"props":1920,"children":1921},{},[1922],{"type":83,"value":1923},"True Edge",{"type":77,"tag":150,"props":1925,"children":1926},{},[1927],{"type":83,"value":1928},"Optimal Kelly",{"type":77,"tag":150,"props":1930,"children":1931},{},[1932],{"type":83,"value":1933},"Your Bet (Full Kelly at +5% estimated)",{"type":77,"tag":150,"props":1935,"children":1936},{},[1937],{"type":83,"value":1938},"Result",{"type":77,"tag":166,"props":1940,"children":1941},{},[1942,1965,1986],{"type":77,"tag":70,"props":1943,"children":1944},{},[1945,1950,1955,1960],{"type":77,"tag":173,"props":1946,"children":1947},{},[1948],{"type":83,"value":1949},"+3%",{"type":77,"tag":173,"props":1951,"children":1952},{},[1953],{"type":83,"value":1954},"5.7%",{"type":77,"tag":173,"props":1956,"children":1957},{},[1958],{"type":83,"value":1959},"9.4%",{"type":77,"tag":173,"props":1961,"children":1962},{},[1963],{"type":83,"value":1964},"Over-betting by 65% — 270% extra variance",{"type":77,"tag":70,"props":1966,"children":1967},{},[1968,1973,1977,1981],{"type":77,"tag":173,"props":1969,"children":1970},{},[1971],{"type":83,"value":1972},"+5%",{"type":77,"tag":173,"props":1974,"children":1975},{},[1976],{"type":83,"value":1959},{"type":77,"tag":173,"props":1978,"children":1979},{},[1980],{"type":83,"value":1959},{"type":77,"tag":173,"props":1982,"children":1983},{},[1984],{"type":83,"value":1985},"Optimal (you got lucky)",{"type":77,"tag":70,"props":1987,"children":1988},{},[1989,1994,1999,2003],{"type":77,"tag":173,"props":1990,"children":1991},{},[1992],{"type":83,"value":1993},"+7%",{"type":77,"tag":173,"props":1995,"children":1996},{},[1997],{"type":83,"value":1998},"13.1%",{"type":77,"tag":173,"props":2000,"children":2001},{},[2002],{"type":83,"value":1959},{"type":77,"tag":173,"props":2004,"children":2005},{},[2006],{"type":83,"value":2007},"Under-betting by 28% — leaving growth on table",{"type":77,"tag":85,"props":2009,"children":2010},{},[2011],{"type":83,"value":2012},"Quarter Kelly at the same estimated +5% would recommend 2.35% per bet. Even if your true edge is 0%, that bet size is small enough not to ruin you. Full Kelly at +5% estimated against a true 0% edge is catastrophic.",{"type":77,"tag":1389,"props":2014,"children":2016},{"id":2015},"worked-example-misjudging-by-2-points",[2017],{"type":83,"value":2018},"Worked Example: Misjudging by 2 Points",{"type":77,"tag":85,"props":2020,"children":2021},{},[2022],{"type":83,"value":2023},"You bet 200 NFL spreads in a season with a perceived +5% edge. Your true edge turns out to be +1% (overestimation is far more common than underestimation among bettors). What happens to a $5,000 bankroll?",{"type":77,"tag":304,"props":2025,"children":2026},{},[2027,2243,2601],{"type":77,"tag":308,"props":2028,"children":2029},{},[2030,2034,2036,2241],{"type":77,"tag":96,"props":2031,"children":2032},{},[2033],{"type":83,"value":1444},{"type":83,"value":2035},": ends around ",{"type":77,"tag":378,"props":2037,"children":2039},{"className":2038},[381],[2040,2138],{"type":77,"tag":378,"props":2041,"children":2043},{"className":2042},[386],[2044],{"type":77,"tag":389,"props":2045,"children":2046},{"xmlns":391},[2047],{"type":77,"tag":394,"props":2048,"children":2049},{},[2050,2133],{"type":77,"tag":398,"props":2051,"children":2052},{},[2053,2058,2062,2067,2071,2076,2080,2085,2089,2093,2097,2101,2105,2109,2113,2117,2121,2125,2129],{"type":77,"tag":402,"props":2054,"children":2055},{},[2056],{"type":83,"value":2057},"5",{"type":77,"tag":408,"props":2059,"children":2060},{"separator":410},[2061],{"type":83,"value":413},{"type":77,"tag":402,"props":2063,"children":2064},{},[2065],{"type":83,"value":2066},"400",{"type":77,"tag":408,"props":2068,"children":2069},{"stretchy":1031},[2070],{"type":83,"value":1034},{"type":77,"tag":420,"props":2072,"children":2073},{},[2074],{"type":83,"value":2075},"m",{"type":77,"tag":420,"props":2077,"children":2078},{},[2079],{"type":83,"value":620},{"type":77,"tag":420,"props":2081,"children":2082},{},[2083],{"type":83,"value":2084},"d",{"type":77,"tag":420,"props":2086,"children":2087},{},[2088],{"type":83,"value":654},{"type":77,"tag":420,"props":2090,"children":2091},{},[2092],{"type":83,"value":116},{"type":77,"tag":420,"props":2094,"children":2095},{},[2096],{"type":83,"value":433},{"type":77,"tag":408,"props":2098,"children":2099},{"stretchy":1031},[2100],{"type":83,"value":1052},{"type":77,"tag":408,"props":2102,"children":2103},{"separator":410},[2104],{"type":83,"value":413},{"type":77,"tag":420,"props":2106,"children":2107},{},[2108],{"type":83,"value":424},{"type":77,"tag":420,"props":2110,"children":2111},{},[2112],{"type":83,"value":448},{"type":77,"tag":420,"props":2114,"children":2115},{},[2116],{"type":83,"value":466},{"type":77,"tag":420,"props":2118,"children":2119},{},[2120],{"type":83,"value":466},{"type":77,"tag":420,"props":2122,"children":2123},{},[2124],{"type":83,"value":448},{"type":77,"tag":420,"props":2126,"children":2127},{},[2128],{"type":83,"value":2075},{"type":77,"tag":402,"props":2130,"children":2131},{},[2132],{"type":83,"value":2057},{"type":77,"tag":472,"props":2134,"children":2135},{"encoding":474},[2136],{"type":83,"value":2137},"5,400 (median), bottom 5% at ",{"type":77,"tag":378,"props":2139,"children":2141},{"className":2140,"ariaHidden":410},[481],[2142],{"type":77,"tag":378,"props":2143,"children":2145},{"className":2144},[486],[2146,2151,2156,2161,2165,2170,2175,2180,2185,2190,2196,2201,2206,2210,2215,2220,2226,2231,2236],{"type":77,"tag":378,"props":2147,"children":2150},{"className":2148,"style":2149},[491],"height:1em;vertical-align:-0.25em;",[],{"type":77,"tag":378,"props":2152,"children":2154},{"className":2153},[497],[2155],{"type":83,"value":2057},{"type":77,"tag":378,"props":2157,"children":2159},{"className":2158},[503],[2160],{"type":83,"value":413},{"type":77,"tag":378,"props":2162,"children":2164},{"className":2163,"style":510},[509],[],{"type":77,"tag":378,"props":2166,"children":2168},{"className":2167},[497],[2169],{"type":83,"value":2066},{"type":77,"tag":378,"props":2171,"children":2173},{"className":2172},[1170],[2174],{"type":83,"value":1034},{"type":77,"tag":378,"props":2176,"children":2178},{"className":2177},[497,520],[2179],{"type":83,"value":2075},{"type":77,"tag":378,"props":2181,"children":2183},{"className":2182},[497,520],[2184],{"type":83,"value":620},{"type":77,"tag":378,"props":2186,"children":2188},{"className":2187},[497,520],[2189],{"type":83,"value":2084},{"type":77,"tag":378,"props":2191,"children":2193},{"className":2192},[497,520],[2194],{"type":83,"value":2195},"ian",{"type":77,"tag":378,"props":2197,"children":2199},{"className":2198},[1266],[2200],{"type":83,"value":1052},{"type":77,"tag":378,"props":2202,"children":2204},{"className":2203},[503],[2205],{"type":83,"value":413},{"type":77,"tag":378,"props":2207,"children":2209},{"className":2208,"style":510},[509],[],{"type":77,"tag":378,"props":2211,"children":2213},{"className":2212},[497,520],[2214],{"type":83,"value":424},{"type":77,"tag":378,"props":2216,"children":2218},{"className":2217},[497,520],[2219],{"type":83,"value":448},{"type":77,"tag":378,"props":2221,"children":2223},{"className":2222},[497,520],[2224],{"type":83,"value":2225},"tt",{"type":77,"tag":378,"props":2227,"children":2229},{"className":2228},[497,520],[2230],{"type":83,"value":448},{"type":77,"tag":378,"props":2232,"children":2234},{"className":2233},[497,520],[2235],{"type":83,"value":2075},{"type":77,"tag":378,"props":2237,"children":2239},{"className":2238},[497],[2240],{"type":83,"value":2057},{"type":83,"value":2242},"4,200",{"type":77,"tag":308,"props":2244,"children":2245},{},[2246,2251,2252,2599],{"type":77,"tag":96,"props":2247,"children":2248},{},[2249],{"type":83,"value":2250},"Full Kelly at +5% estimated",{"type":83,"value":2035},{"type":77,"tag":378,"props":2253,"children":2255},{"className":2254},[381],[2256,2418],{"type":77,"tag":378,"props":2257,"children":2259},{"className":2258},[386],[2260],{"type":77,"tag":389,"props":2261,"children":2262},{"xmlns":391},[2263],{"type":77,"tag":394,"props":2264,"children":2265},{},[2266,2413],{"type":77,"tag":398,"props":2267,"children":2268},{},[2269,2274,2278,2283,2287,2291,2295,2299,2303,2307,2311,2315,2319,2323,2327,2331,2335,2340,2344,2348,2352,2356,2360,2364,2368,2373,2377,2381,2385,2389,2393,2397,2401,2405,2409],{"type":77,"tag":402,"props":2270,"children":2271},{},[2272],{"type":83,"value":2273},"4",{"type":77,"tag":408,"props":2275,"children":2276},{"separator":410},[2277],{"type":83,"value":413},{"type":77,"tag":402,"props":2279,"children":2280},{},[2281],{"type":83,"value":2282},"900",{"type":77,"tag":408,"props":2284,"children":2285},{"stretchy":1031},[2286],{"type":83,"value":1034},{"type":77,"tag":420,"props":2288,"children":2289},{},[2290],{"type":83,"value":2075},{"type":77,"tag":420,"props":2292,"children":2293},{},[2294],{"type":83,"value":620},{"type":77,"tag":420,"props":2296,"children":2297},{},[2298],{"type":83,"value":2084},{"type":77,"tag":420,"props":2300,"children":2301},{},[2302],{"type":83,"value":654},{"type":77,"tag":420,"props":2304,"children":2305},{},[2306],{"type":83,"value":116},{"type":77,"tag":420,"props":2308,"children":2309},{},[2310],{"type":83,"value":433},{"type":77,"tag":408,"props":2312,"children":2313},{"separator":410},[2314],{"type":83,"value":413},{"type":77,"tag":408,"props":2316,"children":2317},{},[2318],{"type":83,"value":1016},{"type":77,"tag":420,"props":2320,"children":2321},{},[2322],{"type":83,"value":424},{"type":77,"tag":420,"props":2324,"children":2325},{},[2326],{"type":83,"value":620},{"type":77,"tag":420,"props":2328,"children":2329},{},[2330],{"type":83,"value":453},{"type":77,"tag":420,"props":2332,"children":2333},{},[2334],{"type":83,"value":448},{"type":77,"tag":420,"props":2336,"children":2337},{},[2338],{"type":83,"value":2339},"w",{"type":77,"tag":420,"props":2341,"children":2342},{},[2343],{"type":83,"value":602},{"type":77,"tag":420,"props":2345,"children":2346},{},[2347],{"type":83,"value":466},{"type":77,"tag":420,"props":2349,"children":2350},{},[2351],{"type":83,"value":116},{"type":77,"tag":420,"props":2353,"children":2354},{},[2355],{"type":83,"value":443},{"type":77,"tag":420,"props":2357,"children":2358},{},[2359],{"type":83,"value":466},{"type":77,"tag":420,"props":2361,"children":2362},{},[2363],{"type":83,"value":654},{"type":77,"tag":420,"props":2365,"children":2366},{},[2367],{"type":83,"value":433},{"type":77,"tag":420,"props":2369,"children":2370},{},[2371],{"type":83,"value":2372},"g",{"type":77,"tag":408,"props":2374,"children":2375},{},[2376],{"type":83,"value":1016},{"type":77,"tag":408,"props":2378,"children":2379},{"stretchy":1031},[2380],{"type":83,"value":1052},{"type":77,"tag":408,"props":2382,"children":2383},{"separator":410},[2384],{"type":83,"value":413},{"type":77,"tag":420,"props":2386,"children":2387},{},[2388],{"type":83,"value":424},{"type":77,"tag":420,"props":2390,"children":2391},{},[2392],{"type":83,"value":448},{"type":77,"tag":420,"props":2394,"children":2395},{},[2396],{"type":83,"value":466},{"type":77,"tag":420,"props":2398,"children":2399},{},[2400],{"type":83,"value":466},{"type":77,"tag":420,"props":2402,"children":2403},{},[2404],{"type":83,"value":448},{"type":77,"tag":420,"props":2406,"children":2407},{},[2408],{"type":83,"value":2075},{"type":77,"tag":402,"props":2410,"children":2411},{},[2412],{"type":83,"value":2057},{"type":77,"tag":472,"props":2414,"children":2415},{"encoding":474},[2416],{"type":83,"value":2417},"4,900 (median, *below starting*), bottom 5% at ",{"type":77,"tag":378,"props":2419,"children":2421},{"className":2420,"ariaHidden":410},[481],[2422],{"type":77,"tag":378,"props":2423,"children":2425},{"className":2424},[486],[2426,2430,2435,2440,2444,2449,2454,2459,2464,2469,2474,2479,2483,2488,2493,2498,2503,2508,2514,2519,2524,2529,2534,2539,2545,2550,2555,2560,2565,2569,2574,2579,2584,2589,2594],{"type":77,"tag":378,"props":2427,"children":2429},{"className":2428,"style":2149},[491],[],{"type":77,"tag":378,"props":2431,"children":2433},{"className":2432},[497],[2434],{"type":83,"value":2273},{"type":77,"tag":378,"props":2436,"children":2438},{"className":2437},[503],[2439],{"type":83,"value":413},{"type":77,"tag":378,"props":2441,"children":2443},{"className":2442,"style":510},[509],[],{"type":77,"tag":378,"props":2445,"children":2447},{"className":2446},[497],[2448],{"type":83,"value":2282},{"type":77,"tag":378,"props":2450,"children":2452},{"className":2451},[1170],[2453],{"type":83,"value":1034},{"type":77,"tag":378,"props":2455,"children":2457},{"className":2456},[497,520],[2458],{"type":83,"value":2075},{"type":77,"tag":378,"props":2460,"children":2462},{"className":2461},[497,520],[2463],{"type":83,"value":620},{"type":77,"tag":378,"props":2465,"children":2467},{"className":2466},[497,520],[2468],{"type":83,"value":2084},{"type":77,"tag":378,"props":2470,"children":2472},{"className":2471},[497,520],[2473],{"type":83,"value":2195},{"type":77,"tag":378,"props":2475,"children":2477},{"className":2476},[503],[2478],{"type":83,"value":413},{"type":77,"tag":378,"props":2480,"children":2482},{"className":2481,"style":510},[509],[],{"type":77,"tag":378,"props":2484,"children":2486},{"className":2485},[497],[2487],{"type":83,"value":1016},{"type":77,"tag":378,"props":2489,"children":2491},{"className":2490},[497,520],[2492],{"type":83,"value":424},{"type":77,"tag":378,"props":2494,"children":2496},{"className":2495},[497,520],[2497],{"type":83,"value":620},{"type":77,"tag":378,"props":2499,"children":2501},{"className":2500,"style":539},[497,520],[2502],{"type":83,"value":453},{"type":77,"tag":378,"props":2504,"children":2506},{"className":2505},[497,520],[2507],{"type":83,"value":448},{"type":77,"tag":378,"props":2509,"children":2512},{"className":2510,"style":2511},[497,520],"margin-right:0.0269em;",[2513],{"type":83,"value":2339},{"type":77,"tag":378,"props":2515,"children":2517},{"className":2516},[497,520],[2518],{"type":83,"value":602},{"type":77,"tag":378,"props":2520,"children":2522},{"className":2521},[497,520],[2523],{"type":83,"value":466},{"type":77,"tag":378,"props":2525,"children":2527},{"className":2526},[497,520],[2528],{"type":83,"value":116},{"type":77,"tag":378,"props":2530,"children":2532},{"className":2531,"style":528},[497,520],[2533],{"type":83,"value":443},{"type":77,"tag":378,"props":2535,"children":2537},{"className":2536},[497,520],[2538],{"type":83,"value":466},{"type":77,"tag":378,"props":2540,"children":2542},{"className":2541},[497,520],[2543],{"type":83,"value":2544},"in",{"type":77,"tag":378,"props":2546,"children":2548},{"className":2547,"style":1277},[497,520],[2549],{"type":83,"value":2372},{"type":77,"tag":378,"props":2551,"children":2553},{"className":2552},[497],[2554],{"type":83,"value":1016},{"type":77,"tag":378,"props":2556,"children":2558},{"className":2557},[1266],[2559],{"type":83,"value":1052},{"type":77,"tag":378,"props":2561,"children":2563},{"className":2562},[503],[2564],{"type":83,"value":413},{"type":77,"tag":378,"props":2566,"children":2568},{"className":2567,"style":510},[509],[],{"type":77,"tag":378,"props":2570,"children":2572},{"className":2571},[497,520],[2573],{"type":83,"value":424},{"type":77,"tag":378,"props":2575,"children":2577},{"className":2576},[497,520],[2578],{"type":83,"value":448},{"type":77,"tag":378,"props":2580,"children":2582},{"className":2581},[497,520],[2583],{"type":83,"value":2225},{"type":77,"tag":378,"props":2585,"children":2587},{"className":2586},[497,520],[2588],{"type":83,"value":448},{"type":77,"tag":378,"props":2590,"children":2592},{"className":2591},[497,520],[2593],{"type":83,"value":2075},{"type":77,"tag":378,"props":2595,"children":2597},{"className":2596},[497],[2598],{"type":83,"value":2057},{"type":83,"value":2600},"1,800",{"type":77,"tag":308,"props":2602,"children":2603},{},[2604,2609,2610,2809],{"type":77,"tag":96,"props":2605,"children":2606},{},[2607],{"type":83,"value":2608},"Quarter Kelly at +5% estimated",{"type":83,"value":2035},{"type":77,"tag":378,"props":2611,"children":2613},{"className":2612},[381],[2614,2709],{"type":77,"tag":378,"props":2615,"children":2617},{"className":2616},[386],[2618],{"type":77,"tag":389,"props":2619,"children":2620},{"xmlns":391},[2621],{"type":77,"tag":394,"props":2622,"children":2623},{},[2624,2704],{"type":77,"tag":398,"props":2625,"children":2626},{},[2627,2631,2635,2640,2644,2648,2652,2656,2660,2664,2668,2672,2676,2680,2684,2688,2692,2696,2700],{"type":77,"tag":402,"props":2628,"children":2629},{},[2630],{"type":83,"value":2057},{"type":77,"tag":408,"props":2632,"children":2633},{"separator":410},[2634],{"type":83,"value":413},{"type":77,"tag":402,"props":2636,"children":2637},{},[2638],{"type":83,"value":2639},"250",{"type":77,"tag":408,"props":2641,"children":2642},{"stretchy":1031},[2643],{"type":83,"value":1034},{"type":77,"tag":420,"props":2645,"children":2646},{},[2647],{"type":83,"value":2075},{"type":77,"tag":420,"props":2649,"children":2650},{},[2651],{"type":83,"value":620},{"type":77,"tag":420,"props":2653,"children":2654},{},[2655],{"type":83,"value":2084},{"type":77,"tag":420,"props":2657,"children":2658},{},[2659],{"type":83,"value":654},{"type":77,"tag":420,"props":2661,"children":2662},{},[2663],{"type":83,"value":116},{"type":77,"tag":420,"props":2665,"children":2666},{},[2667],{"type":83,"value":433},{"type":77,"tag":408,"props":2669,"children":2670},{"stretchy":1031},[2671],{"type":83,"value":1052},{"type":77,"tag":408,"props":2673,"children":2674},{"separator":410},[2675],{"type":83,"value":413},{"type":77,"tag":420,"props":2677,"children":2678},{},[2679],{"type":83,"value":424},{"type":77,"tag":420,"props":2681,"children":2682},{},[2683],{"type":83,"value":448},{"type":77,"tag":420,"props":2685,"children":2686},{},[2687],{"type":83,"value":466},{"type":77,"tag":420,"props":2689,"children":2690},{},[2691],{"type":83,"value":466},{"type":77,"tag":420,"props":2693,"children":2694},{},[2695],{"type":83,"value":448},{"type":77,"tag":420,"props":2697,"children":2698},{},[2699],{"type":83,"value":2075},{"type":77,"tag":402,"props":2701,"children":2702},{},[2703],{"type":83,"value":2057},{"type":77,"tag":472,"props":2705,"children":2706},{"encoding":474},[2707],{"type":83,"value":2708},"5,250 (median), bottom 5% at ",{"type":77,"tag":378,"props":2710,"children":2712},{"className":2711,"ariaHidden":410},[481],[2713],{"type":77,"tag":378,"props":2714,"children":2716},{"className":2715},[486],[2717,2721,2726,2731,2735,2740,2745,2750,2755,2760,2765,2770,2775,2779,2784,2789,2794,2799,2804],{"type":77,"tag":378,"props":2718,"children":2720},{"className":2719,"style":2149},[491],[],{"type":77,"tag":378,"props":2722,"children":2724},{"className":2723},[497],[2725],{"type":83,"value":2057},{"type":77,"tag":378,"props":2727,"children":2729},{"className":2728},[503],[2730],{"type":83,"value":413},{"type":77,"tag":378,"props":2732,"children":2734},{"className":2733,"style":510},[509],[],{"type":77,"tag":378,"props":2736,"children":2738},{"className":2737},[497],[2739],{"type":83,"value":2639},{"type":77,"tag":378,"props":2741,"children":2743},{"className":2742},[1170],[2744],{"type":83,"value":1034},{"type":77,"tag":378,"props":2746,"children":2748},{"className":2747},[497,520],[2749],{"type":83,"value":2075},{"type":77,"tag":378,"props":2751,"children":2753},{"className":2752},[497,520],[2754],{"type":83,"value":620},{"type":77,"tag":378,"props":2756,"children":2758},{"className":2757},[497,520],[2759],{"type":83,"value":2084},{"type":77,"tag":378,"props":2761,"children":2763},{"className":2762},[497,520],[2764],{"type":83,"value":2195},{"type":77,"tag":378,"props":2766,"children":2768},{"className":2767},[1266],[2769],{"type":83,"value":1052},{"type":77,"tag":378,"props":2771,"children":2773},{"className":2772},[503],[2774],{"type":83,"value":413},{"type":77,"tag":378,"props":2776,"children":2778},{"className":2777,"style":510},[509],[],{"type":77,"tag":378,"props":2780,"children":2782},{"className":2781},[497,520],[2783],{"type":83,"value":424},{"type":77,"tag":378,"props":2785,"children":2787},{"className":2786},[497,520],[2788],{"type":83,"value":448},{"type":77,"tag":378,"props":2790,"children":2792},{"className":2791},[497,520],[2793],{"type":83,"value":2225},{"type":77,"tag":378,"props":2795,"children":2797},{"className":2796},[497,520],[2798],{"type":83,"value":448},{"type":77,"tag":378,"props":2800,"children":2802},{"className":2801},[497,520],[2803],{"type":83,"value":2075},{"type":77,"tag":378,"props":2805,"children":2807},{"className":2806},[497],[2808],{"type":83,"value":2057},{"type":83,"value":2810},"4,500",{"type":77,"tag":85,"props":2812,"children":2813},{},[2814,2816,2821,2823,2829,2831,2837],{"type":83,"value":2815},"Full Kelly with a misestimated edge produced ",{"type":77,"tag":125,"props":2817,"children":2818},{},[2819],{"type":83,"value":2820},"worse",{"type":83,"value":2822}," median results than flat staking. That's the edge-uncertainty penalty in action. For a live calculation with your numbers, see our ",{"type":77,"tag":116,"props":2824,"children":2826},{"href":2825},"\u002Fbetting\u002Fkelly-calculator",[2827],{"type":83,"value":2828},"Kelly calculator",{"type":83,"value":2830}," — or, for full session planning across all methods, ",{"type":77,"tag":116,"props":2832,"children":2834},{"href":2833},"\u002Fbetting\u002Fbankroll-calculator",[2835],{"type":83,"value":2836},"calculate your bankroll",{"type":83,"value":2838}," sizing once and reuse it across the season.",{"type":77,"tag":78,"props":2840,"children":2842},{"id":2841},"decision-examples-by-bettor-archetype",[2843],{"type":83,"value":2844},"Decision Examples by Bettor Archetype",{"type":77,"tag":85,"props":2846,"children":2847},{},[2848],{"type":83,"value":2849},"Theory is useful; concrete examples are better. Here's what each archetype actually should do.",{"type":77,"tag":297,"props":2851,"children":2853},{"id":2852},"the-recreational-sports-bettor",[2854],{"type":83,"value":2855},"The Recreational Sports Bettor",{"type":77,"tag":85,"props":2857,"children":2858},{},[2859],{"type":83,"value":2860},"Profile: bets 5-15 plays per week, follows their favorite teams, no tracking spreadsheet, no CLV measurement, doesn't know what their edge is. Bankroll: $500-2,000.",{"type":77,"tag":85,"props":2862,"children":2863},{},[2864,2869],{"type":77,"tag":96,"props":2865,"children":2866},{},[2867],{"type":83,"value":2868},"Method:",{"type":83,"value":2870}," flat staking at 1%. Period.",{"type":77,"tag":85,"props":2872,"children":2873},{},[2874],{"type":83,"value":2875},"Why not Kelly? They have no edge measurement, so any Kelly fraction is meaningless. Why not 2%? Because without an edge, 2% sizing produces a ~4% chance of bankruptcy over 1,000 bets even with a tight strategy.",{"type":77,"tag":297,"props":2877,"children":2879},{"id":2878},"the-sharp-with-tracked-clv",[2880],{"type":83,"value":2881},"The Sharp with Tracked CLV",{"type":77,"tag":85,"props":2883,"children":2884},{},[2885],{"type":83,"value":2886},"Profile: 1,000+ tracked bets, beats closing lines by an average of 0.5-1.5%, knows their edge in MLB totals is +3% but in NFL spreads is +1%. Bankroll: $10,000+.",{"type":77,"tag":85,"props":2888,"children":2889},{},[2890,2894],{"type":77,"tag":96,"props":2891,"children":2892},{},[2893],{"type":83,"value":2868},{"type":83,"value":2895}," Quarter Kelly, scaled per market segment.",{"type":77,"tag":85,"props":2897,"children":2898},{},[2899],{"type":83,"value":2900},"Why not full Kelly? Their edge measurement has standard error of ±1-2%, exactly the territory where full Kelly punishes hardest. Why not flat? Because they have asymmetric edge across markets — flat staking would over-bet weak markets and under-bet strong ones. Quarter Kelly scales sizing to match the actual edge in each segment.",{"type":77,"tag":297,"props":2902,"children":2904},{"id":2903},"the-card-counter",[2905],{"type":83,"value":2906},"The Card Counter",{"type":77,"tag":85,"props":2908,"children":2909},{},[2910,2912,3250],{"type":83,"value":2911},"Profile: trained Hi-Lo counter, can deviate from basic strategy at +2 true count, plays ",{"type":77,"tag":378,"props":2913,"children":2915},{"className":2914},[381],[2916,3081],{"type":77,"tag":378,"props":2917,"children":2919},{"className":2918},[386],[2920],{"type":77,"tag":389,"props":2921,"children":2922},{"xmlns":391},[2923],{"type":77,"tag":394,"props":2924,"children":2925},{},[2926,3076],{"type":77,"tag":398,"props":2927,"children":2928},{},[2929,2934,2938,2942,2946,2950,2954,2959,2963,2967,2971,2975,2979,2983,2987,2991,2995,3000,3004,3008,3013,3017,3021,3025,3029,3034,3038,3042,3046,3050,3054,3058,3062,3066,3071],{"type":77,"tag":402,"props":2930,"children":2931},{},[2932],{"type":83,"value":2933},"25",{"type":77,"tag":420,"props":2935,"children":2936},{},[2937],{"type":83,"value":2075},{"type":77,"tag":420,"props":2939,"children":2940},{},[2941],{"type":83,"value":654},{"type":77,"tag":420,"props":2943,"children":2944},{},[2945],{"type":83,"value":433},{"type":77,"tag":420,"props":2947,"children":2948},{},[2949],{"type":83,"value":654},{"type":77,"tag":420,"props":2951,"children":2952},{},[2953],{"type":83,"value":2075},{"type":77,"tag":420,"props":2955,"children":2956},{},[2957],{"type":83,"value":2958},"u",{"type":77,"tag":420,"props":2960,"children":2961},{},[2962],{"type":83,"value":2075},{"type":77,"tag":420,"props":2964,"children":2965},{},[2966],{"type":83,"value":466},{"type":77,"tag":420,"props":2968,"children":2969},{},[2970],{"type":83,"value":116},{"type":77,"tag":420,"props":2972,"children":2973},{},[2974],{"type":83,"value":424},{"type":77,"tag":420,"props":2976,"children":2977},{},[2978],{"type":83,"value":453},{"type":77,"tag":420,"props":2980,"children":2981},{},[2982],{"type":83,"value":620},{"type":77,"tag":420,"props":2984,"children":2985},{},[2986],{"type":83,"value":602},{"type":77,"tag":420,"props":2988,"children":2989},{},[2990],{"type":83,"value":654},{"type":77,"tag":420,"props":2992,"children":2993},{},[2994],{"type":83,"value":433},{"type":77,"tag":420,"props":2996,"children":2997},{},[2998],{"type":83,"value":2999},"L",{"type":77,"tag":420,"props":3001,"children":3002},{},[3003],{"type":83,"value":116},{"type":77,"tag":420,"props":3005,"children":3006},{},[3007],{"type":83,"value":602},{"type":77,"tag":420,"props":3009,"children":3010},{},[3011],{"type":83,"value":3012},"V",{"type":77,"tag":420,"props":3014,"children":3015},{},[3016],{"type":83,"value":620},{"type":77,"tag":420,"props":3018,"children":3019},{},[3020],{"type":83,"value":2372},{"type":77,"tag":420,"props":3022,"children":3023},{},[3024],{"type":83,"value":116},{"type":77,"tag":420,"props":3026,"children":3027},{},[3028],{"type":83,"value":602},{"type":77,"tag":420,"props":3030,"children":3032},{"mathvariant":3031},"normal",[3033],{"type":83,"value":1490},{"type":77,"tag":420,"props":3035,"children":3036},{},[3037],{"type":83,"value":3012},{"type":77,"tag":420,"props":3039,"children":3040},{},[3041],{"type":83,"value":620},{"type":77,"tag":420,"props":3043,"children":3044},{},[3045],{"type":83,"value":443},{"type":77,"tag":420,"props":3047,"children":3048},{},[3049],{"type":83,"value":654},{"type":77,"tag":420,"props":3051,"children":3052},{},[3053],{"type":83,"value":1011},{"type":77,"tag":420,"props":3055,"children":3056},{},[3057],{"type":83,"value":654},{"type":77,"tag":420,"props":3059,"children":3060},{},[3061],{"type":83,"value":620},{"type":77,"tag":420,"props":3063,"children":3064},{},[3065],{"type":83,"value":2084},{"type":77,"tag":408,"props":3067,"children":3068},{},[3069],{"type":83,"value":3070},"+",{"type":77,"tag":402,"props":3072,"children":3073},{},[3074],{"type":83,"value":3075},"1.5",{"type":77,"tag":472,"props":3077,"children":3078},{"encoding":474},[3079],{"type":83,"value":3080},"25 minimum tables in Las Vegas. Verified +1.5% advantage at neutral counts, scaling up to +3% at +3 true count. Bankroll: ",{"type":77,"tag":378,"props":3082,"children":3084},{"className":3083,"ariaHidden":410},[481],[3085,3236],{"type":77,"tag":378,"props":3086,"children":3088},{"className":3087},[486],[3089,3093,3098,3104,3109,3114,3119,3125,3130,3136,3141,3146,3151,3156,3162,3167,3172,3177,3182,3187,3192,3198,3203,3208,3213,3218,3223,3227,3232],{"type":77,"tag":378,"props":3090,"children":3092},{"className":3091,"style":492},[491],[],{"type":77,"tag":378,"props":3094,"children":3096},{"className":3095},[497],[3097],{"type":83,"value":2933},{"type":77,"tag":378,"props":3099,"children":3101},{"className":3100},[497,520],[3102],{"type":83,"value":3103},"minim",{"type":77,"tag":378,"props":3105,"children":3107},{"className":3106},[497,520],[3108],{"type":83,"value":2958},{"type":77,"tag":378,"props":3110,"children":3112},{"className":3111},[497,520],[3113],{"type":83,"value":2075},{"type":77,"tag":378,"props":3115,"children":3117},{"className":3116},[497,520],[3118],{"type":83,"value":466},{"type":77,"tag":378,"props":3120,"children":3122},{"className":3121},[497,520],[3123],{"type":83,"value":3124},"ab",{"type":77,"tag":378,"props":3126,"children":3128},{"className":3127,"style":539},[497,520],[3129],{"type":83,"value":453},{"type":77,"tag":378,"props":3131,"children":3133},{"className":3132},[497,520],[3134],{"type":83,"value":3135},"es",{"type":77,"tag":378,"props":3137,"children":3139},{"className":3138},[497,520],[3140],{"type":83,"value":2544},{"type":77,"tag":378,"props":3142,"children":3144},{"className":3143},[497,520],[3145],{"type":83,"value":2999},{"type":77,"tag":378,"props":3147,"children":3149},{"className":3148},[497,520],[3150],{"type":83,"value":116},{"type":77,"tag":378,"props":3152,"children":3154},{"className":3153},[497,520],[3155],{"type":83,"value":602},{"type":77,"tag":378,"props":3157,"children":3160},{"className":3158,"style":3159},[497,520],"margin-right:0.2222em;",[3161],{"type":83,"value":3012},{"type":77,"tag":378,"props":3163,"children":3165},{"className":3164},[497,520],[3166],{"type":83,"value":620},{"type":77,"tag":378,"props":3168,"children":3170},{"className":3169,"style":1277},[497,520],[3171],{"type":83,"value":2372},{"type":77,"tag":378,"props":3173,"children":3175},{"className":3174},[497,520],[3176],{"type":83,"value":116},{"type":77,"tag":378,"props":3178,"children":3180},{"className":3179},[497,520],[3181],{"type":83,"value":602},{"type":77,"tag":378,"props":3183,"children":3185},{"className":3184},[497],[3186],{"type":83,"value":1490},{"type":77,"tag":378,"props":3188,"children":3190},{"className":3189,"style":3159},[497,520],[3191],{"type":83,"value":3012},{"type":77,"tag":378,"props":3193,"children":3195},{"className":3194,"style":528},[497,520],[3196],{"type":83,"value":3197},"er",{"type":77,"tag":378,"props":3199,"children":3201},{"className":3200},[497,520],[3202],{"type":83,"value":654},{"type":77,"tag":378,"props":3204,"children":3206},{"className":3205,"style":1091},[497,520],[3207],{"type":83,"value":1011},{"type":77,"tag":378,"props":3209,"children":3211},{"className":3210},[497,520],[3212],{"type":83,"value":654},{"type":77,"tag":378,"props":3214,"children":3216},{"className":3215},[497,520],[3217],{"type":83,"value":620},{"type":77,"tag":378,"props":3219,"children":3221},{"className":3220},[497,520],[3222],{"type":83,"value":2084},{"type":77,"tag":378,"props":3224,"children":3226},{"className":3225,"style":3159},[509],[],{"type":77,"tag":378,"props":3228,"children":3230},{"className":3229},[1136],[3231],{"type":83,"value":3070},{"type":77,"tag":378,"props":3233,"children":3235},{"className":3234,"style":3159},[509],[],{"type":77,"tag":378,"props":3237,"children":3239},{"className":3238},[486],[3240,3245],{"type":77,"tag":378,"props":3241,"children":3244},{"className":3242,"style":3243},[491],"height:0.6444em;",[],{"type":77,"tag":378,"props":3246,"children":3248},{"className":3247},[497],[3249],{"type":83,"value":3075},{"type":83,"value":3251},"5,000.",{"type":77,"tag":85,"props":3253,"children":3254},{},[3255,3259],{"type":77,"tag":96,"props":3256,"children":3257},{},[3258],{"type":83,"value":2868},{"type":83,"value":3260}," Half Kelly scaled by true count.",{"type":77,"tag":85,"props":3262,"children":3263},{},[3264],{"type":83,"value":3265},"Why Half and not Quarter? Card counting is the rare case where edge is measured precisely (the count tells you exactly when you have an advantage and roughly how much). Half Kelly captures more growth than Quarter Kelly at acceptable variance. Why not full? Pit-boss surveillance penalties — large bet spreads attract attention, so even mathematically-justified full Kelly bets are operationally bad.",{"type":77,"tag":297,"props":3267,"children":3269},{"id":3268},"the-sports-bettor-without-clv-tracking",[3270],{"type":83,"value":3271},"The Sports Bettor Without CLV Tracking",{"type":77,"tag":85,"props":3273,"children":3274},{},[3275,3277,3281],{"type":83,"value":3276},"Profile: bets 20+ plays per week across multiple sports, has been betting for 3 years, ",{"type":77,"tag":125,"props":3278,"children":3279},{},[3280],{"type":83,"value":1409},{"type":83,"value":3282}," they have an edge but has never actually measured CLV. Bankroll: $3,000.",{"type":77,"tag":85,"props":3284,"children":3285},{},[3286,3290],{"type":77,"tag":96,"props":3287,"children":3288},{},[3289],{"type":83,"value":2868},{"type":83,"value":3291}," flat staking at 1-2%. Then start tracking CLV. Then re-evaluate after 500 bets.",{"type":77,"tag":85,"props":3293,"children":3294},{},[3295,3297,3302],{"type":83,"value":3296},"Why not Kelly? Because they don't actually have measured edge data — they have a feeling. The single most valuable thing this bettor can do is start logging closing-line-value on every bet. After 500 bets they'll have an actual edge number, and ",{"type":77,"tag":125,"props":3298,"children":3299},{},[3300],{"type":83,"value":3301},"then",{"type":83,"value":3303}," they can graduate to Quarter Kelly. Until then, any Kelly fraction is fiction.",{"type":77,"tag":78,"props":3305,"children":3307},{"id":3306},"interactive-decision-tool",[3308],{"type":83,"value":3309},"Interactive Decision Tool",{"type":77,"tag":85,"props":3311,"children":3312},{},[3313],{"type":83,"value":3314},"Plug in your situation and the tool will tell you which method fits, what unit percentage to use, and what variance profile to expect.",{"type":77,"tag":3316,"props":3317,"children":3318},"inline-kelly-decision-tool",{},[],{"type":77,"tag":85,"props":3320,"children":3321},{},[3322],{"type":83,"value":3323},"The tool weights edge confidence, risk tolerance, experience, and time horizon against survival math — it won't recommend full Kelly to a bettor who hasn't tracked their edge, no matter how aggressive their risk tolerance.",{"type":77,"tag":85,"props":3325,"children":3326},{},[3327,3329,3335,3337,3342],{"type":83,"value":3328},"If you want the broader framing of \"what is a bankroll at all and why does sizing matter,\" see ",{"type":77,"tag":116,"props":3330,"children":3332},{"href":3331},"\u002Fblog\u002Fwhat-is-bankroll-management",[3333],{"type":83,"value":3334},"what is bankroll management",{"type":83,"value":3336}," for the foundational concepts. For the dollar-by-dollar arithmetic across multiple bet types, the ",{"type":77,"tag":116,"props":3338,"children":3339},{"href":2833},[3340],{"type":83,"value":3341},"universal bankroll calculator",{"type":83,"value":3343}," covers Kelly, Half Kelly, Quarter Kelly, and flat side-by-side.",{"type":77,"tag":78,"props":3345,"children":3347},{"id":3346},"common-mistakes-across-all-three-methods",[3348],{"type":83,"value":3349},"Common Mistakes Across All Three Methods",{"type":77,"tag":297,"props":3351,"children":3353},{"id":3352},"using-kelly-without-knowing-your-edge",[3354],{"type":83,"value":3355},"Using Kelly Without Knowing Your Edge",{"type":77,"tag":85,"props":3357,"children":3358},{},[3359],{"type":83,"value":3360},"The most common Kelly mistake. A bettor reads about Kelly Criterion, plugs in a guessed win probability (\"I think I'm 56% on this\"), and bets the recommended fraction. The output is mathematically meaningless because the input was a guess. Kelly with a guessed edge is worse than flat staking with no edge — it amplifies the guess into a bet size.",{"type":77,"tag":85,"props":3362,"children":3363},{},[3364],{"type":83,"value":3365},"The fix: don't use Kelly until you have at least 500 tracked bets with closing-line-value data, or a verified counting\u002Fadvantage-play scenario. Until then, flat 1%.",{"type":77,"tag":297,"props":3367,"children":3369},{"id":3368},"going-full-kelly-with-estimated-edge",[3370],{"type":83,"value":3371},"Going Full Kelly with Estimated Edge",{"type":77,"tag":85,"props":3373,"children":3374},{},[3375,3377,3382],{"type":83,"value":3376},"Even bettors with measured edge often jump to full Kelly because \"math says it's optimal.\" Math says it's optimal ",{"type":77,"tag":125,"props":3378,"children":3379},{},[3380],{"type":83,"value":3381},"when edge is known with certainty",{"type":83,"value":3383},". Estimated edge is not certain edge. Full Kelly with estimation error produces 4× the variance you bargained for and lower median growth than fractional Kelly.",{"type":77,"tag":85,"props":3385,"children":3386},{},[3387],{"type":83,"value":3388},"The fix: default to Quarter Kelly. Move to Half Kelly only with verified, low-noise edge measurement. Move to full Kelly almost never.",{"type":77,"tag":297,"props":3390,"children":3392},{"id":3391},"ignoring-variance-differences-across-bet-types",[3393],{"type":83,"value":3394},"Ignoring Variance Differences Across Bet Types",{"type":77,"tag":85,"props":3396,"children":3397},{},[3398],{"type":83,"value":3399},"Treating a parlay the same as a straight bet, or a tournament buy-in the same as a cash-game buy-in, breaks the Kelly framework. Different bet types have different variance profiles, and a method tuned for one breaks for another.",{"type":77,"tag":85,"props":3401,"children":3402},{},[3403],{"type":83,"value":3404},"The fix: tier your sizing by variance. Straight bets at 1-2% (or Quarter Kelly), parlays at 0.5-1% flat, tournament buy-ins at 0.5-1% flat. Even pure Kelly bettors should switch to flat for high-variance and correlated bets.",{"type":77,"tag":297,"props":3406,"children":3408},{"id":3407},"treating-fractional-kelly-as-kelly-with-training-wheels",[3409],{"type":83,"value":3410},"Treating Fractional Kelly as \"Kelly with Training Wheels\"",{"type":77,"tag":85,"props":3412,"children":3413},{},[3414,3416,3421],{"type":83,"value":3415},"The biggest psychological mistake: thinking Quarter Kelly is for beginners, and \"real pros\" use full Kelly. The opposite is true. Quarter Kelly is what experienced pros use ",{"type":77,"tag":125,"props":3417,"children":3418},{},[3419],{"type":83,"value":3420},"because",{"type":83,"value":3422}," they understand variance. Full Kelly is what overconfident beginners use because they haven't yet had a 1-in-20 trajectory ruin their bankroll.",{"type":77,"tag":85,"props":3424,"children":3425},{},[3426],{"type":83,"value":3427},"The fix: read fractional Kelly as a feature, not a compromise. The expected-value loss is small. The variance reduction is enormous. The trade is overwhelmingly favorable for any real-world bettor.",{"type":77,"tag":78,"props":3429,"children":3431},{"id":3430},"faq",[3432],{"type":83,"value":3433},"FAQ"]