Risk of Ruin

Risk of Ruin (RoR)

A sharp bettor with a real +3% ROI edge, a $5,000 bankroll, betting flat $200 per match (4% of bankroll). Considers himself a pro. Over 3 months with a 3% edge, he expects to make around $1,000. In practice, 22% of the time he goes broke before that positive EV ever accumulates. That's not bad luck. That's variance math. Without understanding risk of ruin, any betting or poker strategy is a lottery: long-term upside, short-term chance of getting wiped out.

What It Actually Is

Risk of Ruin (RoR) is the probability that your bankroll hits zero before you get the chance to realize your long-term edge. The key word is "probability" — a statistical figure between 0% and 100%. An RoR of 1% means: in one out of every hundred career scenarios with these parameters, you go broke.

RoR depends on three core variables:

• Bankroll: the amount of money you're willing to lose before you stop playing.
• Edge: your average profit per unit of play (ROI for betting, BB/100 for poker).
• Variance: the standard deviation of your results.

A player with a positive edge is not automatically protected from ruin. A thin bankroll combined with high variance and oversized bets can push RoR well above 50%, even with a positive edge. Understanding RoR is the difference between a professional player and a tilt-prone amateur who's broke two months in despite having a "winning" approach.

The concept comes from financial mathematics and random walk theory. If each of your results is an independent draw from a distribution with a positive mean but non-zero sigma, there's a finite probability that a run of consecutive losses wipes out your capital before the long-term plus ever materializes. That probability doesn't depend on luck — it depends on your system's parameters.

Formula: the Classic Silio Model

The classic Risk of Ruin formula using normal approximation:

RoR = ((1 - edge/sigma) / (1 + edge/sigma))^(bankroll/sigma)

Where:

• edge is your average win per bet (in the same units as sigma)
• sigma is the standard deviation of your results
• bankroll is your total bankroll size

A concrete example. A poker player with a +5 BB/100 edge and sigma of 100 BB/100. Bankroll of 200 BB (4 buy-ins for NL cash).

RoR = ((1 - 0.05) / (1 + 0.05))^(200/100) = (0.952)^2 = 90.7%

That's a 90.7% chance of going broke over a career with that bankroll. To bring RoR down to 5%, you'd need to grow the bankroll to:

bankroll = sigma × ln(RoR) / ln((1-edge/sigma)/(1+edge/sigma))

bankroll = 100 × ln(0.05) / ln(0.952) = 100 × (-2.996) / (-0.049) = 6122 BB

That's 60 buy-ins. 15 times more than 4. That's the real bankroll requirement for a regular NL cash player working with realistic numbers.

Use the risk of ruin calculator to work out your exact RoR with your specific edge, sigma, and bankroll. It's the foundational tool for any serious career plan.

Three Variables: Bankroll, Edge, Variance

Bankroll (money size). Direct relationship: bigger bankroll, lower RoR. Doubling your bankroll produces an exponential drop in RoR. That's why pros build their bankroll aggressively until they reach a comfortable range of hundreds of buy-ins, then ease off on adding funds.

Edge (win rate). Higher edge means lower RoR. Doubling your edge roughly squares the reduction in RoR. But edge has practical limits: for most games, an edge above 5% puts you in the top 1% of players, and getting there without serious study and experience is a tall order.

Variance (sigma). Higher variance means higher RoR. This one gets overlooked: you can improve your edge, but if variance is climbing at the same time, RoR stays flat or even goes up. The classic case is moving from microstakes to higher stakes. Edge drops, sigma holds, and bankroll has to grow just to keep RoR where it was.

The edge / sigma ratio is what finance calls the Sharpe ratio, and it fundamentally drives RoR. A low Sharpe (edge/sigma < 0.05) demands a massive bankroll. A high Sharpe (>0.10) lets you build a sustainable career even on a modest roll.

RoR for Different Betting Strategies

Flat betting (fixed stake size). You bet the same amount on every event, say 2% of your bankroll. The safest long-term approach. RoR is minimal with a properly sized bankroll. The downside: slow growth.

Kelly Criterion (Full Kelly). Stake size = (edge / odds) × bankroll. Mathematically maximizes long-term growth, but produces a meaningful RoR (around 13.5%) even for a perfect player. Full Kelly is rarely used in practice because the variance is genuinely brutal.

Fractional Kelly (1/4 or 1/2 Kelly). The most popular approach among professionals. You bet a quarter or half of what Kelly recommends. This slashes RoR dramatically, down to 1–3% with a decent bankroll, while keeping 50–80% of Full Kelly's long-term growth rate. Best practice for most serious bettors.

Martingale and its variants. Double your stake after every loss. Mathematically catastrophic: RoR approaches 100% over any meaningful sample, because a cold run of 7–10 consecutive losses will wipe out a standard 200 buy-in bankroll. No professional uses this with real money.

Variable sizing by edge. Stake size scales with your perceived edge in each situation. Works well when your edge estimates are accurate. Works badly when they're off, which happens more often than you'd think. RoR is moderate.

More detail on Fractional Kelly in the Kelly Criterion article. Use the Kelly calculator to find your optimal stake size for a given edge.

Real Numbers: RoR for Poker and Sports Betting

Concrete RoR estimates for typical player profiles:

Cash poker NL10, edge +5 BB/100, sigma 100, bankroll 50 buy-ins (5,000 BB): RoR ~5.2%. A comfortable buffer, but not fully safe.

Cash poker NL10, same edge, 30 buy-ins: RoR ~22%. Too thin for a long-term grind.

Cash poker NL10, same edge, 100 buy-ins: RoR ~0.3%. Very safe, though possibly overkill for micro-stakes.

MTT, edge +20% ROI, sigma 200%, bankroll 100 buy-ins: RoR ~3%. The floor for MTT play.

MTT, same edge, 50 buy-ins: RoR ~27%. High risk.

High-stakes MTT, edge +10% ROI, sigma 250%, bankroll 100 buy-ins: RoR ~28%. High stakes demand a significantly larger bankroll because of the elevated variance.

Sports betting, edge +3% ROI, sigma 15% per bet, bankroll 100 bets: RoR ~12%. Standard for a disciplined bettor.

Sports betting, same edge, 200 bets: RoR ~1.5%. Much safer.

Sports betting, same edge, 500 bets: RoR <0.1%. Essentially zero risk, provided the edge holds.

Rule of thumb: cash poker (low and mid stakes) needs 50–100 buy-ins, MTTs need 100–200, sports betting needs 200–500 bets. The higher the variance, the more you need. These numbers sound conservative. They're also the ones that save careers.

Kelly Criterion and Its Relationship to RoR

Kelly Criterion is the mathematically optimal bet size for maximizing long-term logarithmic bankroll growth. The formula:

Kelly% = (edge × odds - 1) / (odds - 1)

Where odds are in decimal format (e.g., 2.10 = 110% return on a win).

Concrete example: a 5% edge at 2.0 odds gives a Kelly bet of 5% of your bankroll. On a $5,000 bankroll, that's $250 per bet.

The connection to RoR: Full Kelly maximizes long-term growth but carries meaningful RoR (around 13.5% under ideal parameters). Dropping to 1/4 Kelly delivers roughly 92% of that growth with RoR below 1%. That's exactly why professionals use Fractional Kelly.

Fractional Kelly breakdown:

• Full Kelly: maximum growth, 13.5% RoR
• Half Kelly (1/2): 75% of growth, 2% RoR
• Quarter Kelly (1/4): 50% of growth, RoR <0.1%

Fractional Kelly is the only rational balance between maximizing long-term growth and keeping risk of ruin in check. Full Kelly demands tolerance for extreme swings that most players find psychologically unbearable.

Bankroll Management via RoR Tables

Standard bankroll requirements for common game types:

NL cash poker (BB/100 win-rate, sigma 100):

• 2 BB/100 edge: 80+ buy-ins for RoR <5%
• 3 BB/100 edge: 50 buy-ins
• 5 BB/100 edge: 30 buy-ins
• 8 BB/100 edge: 20 buy-ins (microstakes, where edge is massive)
• 10+ BB/100 edge: 15 buy-ins

MTT poker (ROI%, sigma 200%):

• +10% ROI edge: 150–200 buy-ins
• +20% ROI edge: 100 buy-ins
• +30% ROI edge: 75 buy-ins
• +50% ROI edge: 50 buy-ins (micro MTTs)

Sports betting (ROI per bet, sigma 15–20%):

• +1% ROI edge: 500+ bets (at flat 1% of bankroll staking)
• +2% ROI edge: 250 bets
• +3% ROI edge: 150–200 bets (the professional standard)
• +5% ROI edge: 100 bets (a genuinely rare edge level)

Casino blackjack with card counting:

• +1.5% edge (Hi-Lo average): 200–400 minimum bets
• +2–3% edge (more advanced systems): 100–150 minimum bets

These figures assume disciplined play and a consistent edge across the full sample. If your edge fluctuates with skill variance or you run into extended downswings, build your bankroll requirements up further.

Real Ruin Scenarios

Situations where a disciplined player actually goes broke:

Scenario 1: Moving up too aggressively. A player has 50 buy-ins at NL10 ($500) with a 3% RoR. Decides to move to NL25. Now they have 20 buy-ins at NL25 ($500), RoR jumps to 38%. A normal 15–20 buy-in downswing at that stake is completely routine, and they go broke before ever finding their footing at the new limit.

Scenario 2: Mixing stakes without separate bankrolls. A player runs NL10, NL25, and NL50 simultaneously with one combined bankroll. The tracker shows overall +EV, but per-stake data tells a different story: -EV at NL50, +EV at NL10. Actual win rate at NL50 is -1 BB/100, making RoR at NL50 close to 100% even when the overall results look positive.

Scenario 3: Betting correlated legs. A player bets on 5 different teams across the same set of matches through a parlay. Portfolio variance spikes sharply because the bets are correlated. Standard RoR calculations assume independent bets and badly underestimate the real RoR here.

Scenario 4: Chasing losses on tilt. Player is down on a session. Convinces himself he'll win it back by doubling up and jacks the stakes to 2x. Repeat this pattern regularly, and RoR scales proportionally to the number of those episodes, even with a positive edge overall.

Scenario 5: Variance early in a career. Player studies poker and makes the jump to real money. Edge isn't there yet: through the first 30–50K hands they can easily be running at -2 BB/100 due to bad early call decisions. RoR is high during this window, and many people quit before their edge ever turns positive.

When RoR Underestimates Your Real Risk

The standard RoR formula rests on assumptions that often don't hold:

1. Stable edge. The formula assumes a constant edge. In practice it erodes: regulars improve, competition gets tougher, and edge thins out at higher stakes. Real ruin probability is often higher than what the model predicts.

2. Normal distribution. The formula assumes a Gaussian distribution. Real results in poker and sports betting have fatter tails. Extreme events happen more often than the model expects. Actual RoR can run 50–100% above the theoretical figure.

3. Independent bets. Each bet or hand is assumed to be independent. In reality, emotional decisions, tilt, and error spirals create correlations between bets. Downswings cluster, pushing real RoR higher.

4. Life doesn't interfere. The formula doesn't account for losing interest, changing careers, divorce, or anything else that ends a gambling career without the bankroll ever hitting zero.

5. No bankroll top-ups. The model assumes a closed system. Plenty of players reload when they run low, whether from savings or outside income. This pulls real RoR below the theoretical number.

Practical rule of thumb: add 50–100% on top of whatever the calculator spits out. If it says 5% RoR, plan for 10%.

Common Mistakes

1. Ignoring sigma when evaluating RoR. A player focuses purely on edge ("I'm a winning player!") without factoring in sigma. At NL50 with a sigma of 120 BB/100 and an edge of +3 BB/100, you still need 50+ buy-ins, same as at lower stakes with the same edge. Edge does not equal safety.

2. Insufficient sample size for estimating edge. A player believes they have a +5 BB/100 edge after 10K hands. The true edge could be +1 or -1. Running RoR calculations with an inflated edge creates a false sense of security.

3. Bumping up stakes after a winning streak. After a big upswing, a player thinks they're running hot and decides to play bigger. Classic gambler's fallacy. Variance has no memory of previous results, and a larger bet size means higher RoR.

4. Underestimating downswings. A player refuses to believe a 20 buy-in downswing at NL10 is realistic. Statistically, it's completely normal, it happens every 3–5 months even for a disciplined player. If your bankroll doesn't account for it, ruin is just a matter of time.

5. Mixing stakes without separate bankrolls. Playing different limits from a single bankroll masks the RoR for each individual stake. Especially dangerous at higher stakes where your edge might actually be negative.

6. Treating high RoR as "normal for poker." A 30% RoR is a disaster for any long-term career. A disciplined player targets RoR below 5%; a professional, below 2%.

Where This Model Falls Short

RoR is a mathematical model built on statistical assumptions. A real playing career depends on factors the formula simply doesn't capture:

Psychological factors: tilt, mental state, lifestyle, motivation, family situation. A player with a mathematically sound RoR can still walk away from poker due to burnout, not bankruptcy.

Edge decay over time: real edge erodes as competition toughens and the game evolves. RoR calculated on today's edge says nothing about your edge six months from now.

Moving up in stakes: takes time and often comes with a temporary loss of edge. A low RoR can effectively mean "stay at current stakes instead of growing your career."

Life events: divorce, illness, family obligations, relocation. Any of these can force a bankroll withdrawal or an extended break. RoR doesn't account for any of it.

Taxes: actual net winnings after taxes are significantly lower than gross in most jurisdictions. Calculate RoR on after-tax figures for realistic long-term planning.

Account restrictions: sites can close accounts, freeze withdrawals, or suspend accounts pending review. That's an external reduction of your bankroll that the standard RoR formula ignores entirely.

RoR is a starting point for bankroll management, not a final answer. Combine RoR calculations with a pragmatic safety buffer (50–100% more than the formula recommends), regular edge reassessments, and consistent bet-sizing discipline.