# How to Use the No-Vig Calculator and Interpret Its Results

> Use the no-vig calculator for complete 2–4 outcome markets. Follow a 3-way example, compare methods, and understand adjusted probabilities and EV.

- Source: https://toolsgambling.com/blog/no-vig-calculator
- Author: Evgeniy Volkov (ToolsGambling)
- Published: 2026-02-23
- Updated: 2026-10-09
- Cite as: "No-Vig Calculator: Remove Margin and Read Fair Odds", ToolsGambling, https://toolsgambling.com/blog/no-vig-calculator

A no-vig calculator removes the margin from a set of bookmaker odds to produce **method-derived probability estimates and no-vig odds**, sometimes called fair odds. These are not true probabilities, new prices a bookmaker must offer, guaranteed profit, or a prediction.

Use a complete market with two, three, or four mutually exclusive outcomes, all from the same bookmaker, time, match, line, and settlement rules. The embedded calculator uses the proportional method.

## Which odds belong in the same calculation?

Enter every possible outcome of one market, with no overlap between outcomes. An incomplete or overlapping selection cannot provide a valid no-vig estimate for the whole market.

Suitable examples include:

- **Football 1X2:** Home, Draw, and Away. You must include the draw.
- **Over/Under 2.5:** Over 2.5 and Under 2.5, with identical settlement rules.
- **A four-way market:** Four genuinely disjoint outcomes that cover every possible result—not four shortlisted winners from a larger field.

Do not combine double-chance outcomes 1X, 12, and X2: they overlap. A match-winner market including overtime and a winner market settled after 90 minutes are also different markets.

Pushes and refunds require a separate model. This tool does not adjust for betting-exchange commission. All numerical examples below are hypothetical, not current bookmaker offers.

## How to enter a three-way market

Select **3-Way**, switch to **Decimal**, and enter Home `2.70`, Draw `3.20`, and Away `2.80`. Results update automatically; there is no Calculate button.

1. Select **3-Way**.
2. Select **Decimal**. The default format is American.
3. Enter **Home**, **Draw**, and **Away** in the outcome-name fields.
4. Enter `2.70`, `3.20`, and `2.80` in their respective odds fields. Use a decimal **dot**.
5. Read the automatically generated results in the same row order.

![No-vig calculator inputs with 3-Way and Decimal selected, showing Home 2.70, Draw 3.20, and Away 2.80](https://toolsgambling.com/media/article/2026/10/5a845a0bbf74e9c58cc95720066b952f.webp)
Localhost screenshot of a complete hypothetical three-way market, not a live bookmaker offer.

Fill every odds field in the selected mode: two, three, or four. Placeholder text in an empty field is not entered data. **Reset** clears the fields for the selected mode.

### How to read the displayed results

For this example, the calculator displays no-vig odds of **2.808 / 3.328 / 2.912** and adjusted probabilities of **35.6% / 30.0% / 34.3%**, in Home, Draw, Away order.

![Last two result columns showing no-vig odds 2.808, 3.328, and 2.912, with adjusted probabilities 35.6%, 30.0%, and 34.3%](https://toolsgambling.com/media/article/2026/10/0091d808669f35d52ee8fae883951d15.webp)
Localhost screenshot showing only the last two result columns; the rows remain Home, Draw, Away.

The **Fair Prob** label means an adjusted estimate under the proportional method—not a measured true probability.

Two other display details matter:

- **Total Overround** shows **104.00%**, which is actually the **booksum**, or total implied probability. The excess above 100%, **4.00%**, appears separately as **Margin**.
- **Efficiency** is not expected return or theoretical hold. Ignore it for betting decisions, and do not treat the verdict badge as verified evidence of profitability.

## How to calculate no-vig odds manually

Under the proportional method, convert each decimal price to an implied probability, then divide each probability by their sum. This applies the same relative reduction to every outcome—not the same subtraction in percentage points.

For decimal odds $O_i$:

$$
q_i=\frac{1}{O_i},\qquad S=\sum_i q_i
$$

Here, $q_i$ is the raw implied probability and $S$ is the dimensionless booksum. For example, $S=1.04$ represents 104%, not 4%.

The overround—the excess above 100%, conventionally displayed as a margin percentage—is:

$$
100(S-1)
$$

The adjusted probability and corresponding no-vig odds are:

$$
p_i=\frac{q_i}{S},\qquad O_i^*=\frac{1}{p_i}=O_i\times S
$$

This normalization is explained in [Jonas C. Lindstrøm’s implied package vignette, dated June 11, 2023](https://cran.r-project.org/web/packages/implied/vignettes/introduction.html).

![Original proportional margin-removal formula from the implied package documentation, using rᵢ for raw implied probability](https://toolsgambling.com/media/article/2026/10/99630d4e57ed8defb98230572dadb45f.webp)
Original formula image from the documentation dated 2023-06-11. Its notation $r_i$ corresponds to $q_i$ in this guide.

### Full calculation for 2.70 / 3.20 / 2.80

The booksum is **1.0400132275**, or **104.00132275%**, and the excess above 100% is **4.00132275%**.

With greater displayed precision, the values are:

- Raw implied probabilities: **37.037037% / 31.25% / 35.714286%**.
- Adjusted probabilities: **35.612083% / 30.047695% / 34.340223%**.
- No-vig odds: **2.8080357 / 3.3280423 / 2.9120370**.

| Outcome | Entered odds | Raw implied probability | Adjusted probability | No-vig odds |
|---|---:|---:|---:|---:|
| Home | 2.70 | 37.04% | 35.61% | 2.808 |
| Draw | 3.20 | 31.25% | 30.05% | 3.328 |
| Away | 2.80 | 35.71% | 34.34% | 2.912 |

This example assumes a complete, mutually exclusive 1X2 market with matching settlement rules and no push or commission adjustment. The table rounds percentages to two decimal places and prices to three. Retain precision during calculation and round only for display.

The exact model sums to 100%. Display rounding can produce totals such as 99.9% or 100.1% without indicating an error. The embedded tool calculates its displayed probability from the no-vig price already rounded to three decimals, so its one-decimal probability may differ from directly rounding the precise $p_i$.

## How American odds and four-way markets work

American odds must first be converted without intermediate rounding; a four-way market uses the same proportional normalization as a two- or three-way market. In either case, the outcomes must form a complete market.

### American example: -180 and +155

Use the original American inputs **`-180`** and **`+155`**, not rounded decimal equivalents, for this hypothetical two-outcome market.

The exact conversions are:

$$
O_1=1+\frac{100}{180},\qquad O_2=1+\frac{155}{100}
$$

The raw implied probabilities are **64.285714%** and **39.215686%**. Their total is **103.501401%**, giving an overround of **3.501401%**.

After proportional margin removal:

| Measure | First outcome | Second outcome |
|---|---:|---:|
| Adjusted probability | 62.110961% | 37.889039% |
| Decimal no-vig odds | 1.6100218 | 2.6392857 |
| Approximate American no-vig odds | -164 | +164 |

These displayed values assume two exhaustive, mutually exclusive outcomes and no push or commission. The American outputs are approximate whole-number prices.

**1.556** is only a rounded representation of the first original price. Treating it as an exact input changes the calculation. Use the [odds converter](https://toolsgambling.com/betting/odds-converter) when you need to check formats.

### Four-way example: 2.5 / 3.5 / 5 / 8

For a hypothetical complete four-outcome market priced at **2.5 / 3.5 / 5 / 8**, the booksum is **101.071429%** and the overround is **1.071429%**.

| Outcome | Original odds | Adjusted probability | No-vig odds |
|---|---:|---:|---:|
| First | 2.5 | 39.575972% | 2.526786 |
| Second | 3.5 | 28.268551% | 3.5375 |
| Third | 5 | 19.787986% | 5.053571 |
| Fourth | 8 | 12.367491% | 8.085714 |

Values are shown at the stated display precision. This calculation only represents the whole market if these four outcomes are disjoint and exhaustive, with the same settlement rules.

## Why a 4% margin does not mean 4% bookmaker profit

Overround and theoretical hold under balanced liabilities are different measures. Neither tells you the bookmaker’s realized profit.

Keep these three quantities separate:

| Measure | Formula, expressed as a percentage | Meaning |
|---|---|---|
| Booksum | $100S$ | Total raw implied probability |
| Overround | $100(S-1)$ | Excess above 100%, conventionally displayed as margin |
| Theoretical hold under balanced liabilities | $100(S-1)/S$ | Conditional turnover share when stakes are allocated to produce equal payouts across outcomes |

For hypothetical odds of **1.91 / 1.91**:

- Booksum: **104.7120419%**.
- Overround: **4.7120419%**.
- Adjusted probabilities: **50% / 50%**.
- No-vig odds: **2.00 / 2.00**.

Under equal-payout allocation, the theoretical hold is **4.5%**, while the conditional return share $100/S$ is **95.5%**. These figures are not promised returns to an individual bettor, guaranteed losses on a particular bet, or the bookmaker’s actual revenue.

In the main 1X2 example, theoretical hold is **3.8473768%**, versus an overround of **4.0013228%**. The [margin calculator](https://toolsgambling.com/betting/margin-calculator) can help check these quantities; it also provides no-vig odds.

## Why margin-removal methods give different estimates

Different methods make different assumptions about how margin is distributed across outcomes. Their estimates can therefore differ for identical odds, and choosing a method does not turn its output into a true probability.

The embedded calculator uses **only the proportional method**. The separate [implied probability calculator](https://toolsgambling.com/betting/implied-probability) supports four methods, up to eight outcomes, and decimal, American, and fractional odds.

For **2.70 / 3.20 / 2.80**, the comparison is:

| Method | Home | Draw | Away |
|---|---:|---:|---:|
| Proportional / multiplicative | 35.61% | 30.05% | 34.34% |
| Additive | 35.70% | 29.92% | 34.38% |
| Power | 35.70% | 29.93% | 34.37% |
| Shin | 35.68% | 29.95% | 34.37% |

Percentages are rounded to two decimal places. Small differences may disappear at this display precision; they do not establish which estimate is more accurate.

### What assumptions do the methods make?

Each method models margin allocation rather than verifying how the bookmaker actually set its prices.

| Method | Assumption | Limitation |
|---|---|---|
| Proportional / multiplicative | Divide each $q_i$ by $S$ | Preserves relative implied probabilities without proving their accuracy |
| Additive | Subtract $(S-1)/n$ from each $q_i$, where $n$ is the number of outcomes | An equal percentage-point reduction can produce negative probabilities |
| Power | Find $k$ such that $\sum q_i^k=1$ | Not a universally best margin-allocation method |
| Shin | Apply a pricing model with an assumed informed-bettor component | A fitted parameter does not independently prove insider activity |

In the standard two-outcome formulation, Shin equals additive—not universally proportional. This equivalence and other methodological limits are discussed in [Clarke, Kovalchik and Ingram’s paper, published December 25, 2017](https://www.sciencepublishinggroup.com/article/10.11648/j.ajss.20170506.12). Its method-comparison findings apply to the research dataset, not every betting market.

Likewise, one odds snapshot cannot establish favourite–longshot bias or prove that a particular method is universally superior.

### How to compare methods in the separate calculator

Open the separate calculator, enter the same complete market, and switch method buttons while keeping the inputs unchanged. Results update automatically.

1. Open the [implied probability calculator](https://toolsgambling.com/betting/implied-probability).
2. It initially has two outcome rows. Click the add-outcome button to create a third row.
3. Select decimal format and enter Home `2.70`, Draw `3.20`, and Away `2.80` in the outcome and odds fields.
4. Click **Multiplicative**, **Additive**, **Power**, and **Shin** in turn. **Multiplicative** is the interface label for the proportional method.
5. Compare the probability columns.

Check the inputs carefully when the margin is low or negative, or the market may be incomplete. This implementation’s additive method clamps probabilities to **1e-9**, a very small positive floor, and renormalizes them. That changes the model; it does not establish validity.

The power search is restricted to $k\geq1$, so it is not valid for normalizing a booksum below 1. For an underround, use **only proportional normalization** and investigate why the sum is below 100%. The [implied package documentation](https://cran.r-project.org/web/packages/implied/vignettes/introduction.html) explains broader method and market-scope limitations.

## How to use a no-vig estimate when assessing EV

Removing margin alone does not prove that a bet offers value. Expected value requires a defensible probability estimate and the **offered price**, not merely a comparison between your estimate and a no-vig benchmark.

For offered decimal odds $O$, the break-even probability is:

$$
p_{\text{break-even}}=\frac{1}{O}
$$

Expected net return per unit staked, with no fees or push, is:

$$
\mathrm{EV}=p\times O-1
$$

Suppose the main example’s Home estimate is **35.612083%**, and a hypothetical price of **3.00** is available elsewhere for the same event, outcome, and settlement rules.

Then:

- Break-even probability: **33.3333%**.
- Probability gap: **2.278749 percentage points**.
- Conditional expected net return: **+6.836248%**, or approximately **+6.84 units per 100 staked**.

This uses the underlying calculation precision, with values rounded only for display. It holds **only if the benchmark probability estimate is valid**, with no fees or pushes. A probability gap of 2.278749 percentage points is not the same quantity as 6.84% EV.

At the original offered price of **2.70**, using its own proportional no-vig estimate gives conditional EV of **−3.847377%**. The calculator has not created an edge.

The [value bet calculator](https://toolsgambling.com/betting/value-bet-calculator) requires you to enter a probability estimate; it does not establish that estimate’s truth. For further context, see [what an edge means in betting](https://toolsgambling.com/blog/what-does-edge-mean-in-betting).

## What to check when a result looks wrong

First check the input format, completed fields, and market coverage. An unusual total calls for checking the data—not assuming a profit opportunity.

- **Decimal odds:** Enter values greater than 1 using a dot, such as `2.70`.
- **American odds:** Use a sign and a standard absolute value of at least 100, such as `-180` or `+155`.
- **Required fields:** Fill every odds field in the current mode. Placeholders do not count.
- **Complete coverage:** Do not omit a draw, an “other” outcome, or a participant needed to complete the market.
- **Matching scope:** Do not mix bookmakers, timestamps, lines, or settlement rules.
- **Outcome count:** For more than four outcomes, use the separate tool for up to eight, or an appropriate full-market tool. Eight selected runners do not form a complete 30-runner market.
- **Rounding:** Small deviations from 100% in displayed probabilities are normal. Input validation and invalid-value filtering in the separate calculator do not prove that the market is complete.

### What does a total below 100% mean?

A booksum below 100% can reflect incomplete, mixed, or stale data—or a genuine underround in the quoted market. It does not automatically establish an accessible arbitrage opportunity.

For hypothetical odds of **2.10 / 2.10**:

- Booksum: **95.238095%**.
- Overround: **−4.761905%**.
- Proportional probabilities: **50% / 50%**.
- Normalized no-vig odds: **2.00 / 2.00**.

This assumes two complete, mutually exclusive outcomes. Values are rounded for display. The normalized prices become **shorter**, not longer: this is normalization to 100%, not the usual removal of a positive bookmaker margin.

The calculator helps explain prices, but it does not remove betting risk. If you are an adult and choose to bet, set a limit you can afford to lose and do not chase losses.

## FAQ

### What if my odds are in fractional format?

The separate implied probability calculator supports fractional odds. For the embedded calculator, first use the odds converter to convert them to decimal or American odds. Do not enter a fraction in a decimal field.

### Does the calculator provide betting signals?

No. It calculates method-derived probability estimates and no-vig odds. This benchmark alone does not establish that a bet offers value or predict the event's result.