5 Devig Methods Compared: Which One Should You Use?

The Problem

Bookmakers build a margin (vig) into their odds. A "fair" -110/-110 market has implied probabilities summing to 104.76% instead of 100%. To find the true probability, you need to remove the vig — but how you remove it matters.

There are five common methods, each with different assumptions about how bookmakers distribute the margin.

The Five Methods

1. Equal Margin (EM)

The simplest approach: subtract an equal share of the margin from each outcome.

$$p_{\text{fair},i} = p_{\text{implied},i} - \frac{\text{margin}}{n}$$

Assumption: The bookmaker adds the same absolute margin to each outcome. This rarely holds — favorites and longshots usually get different treatment.

2. Margin Proportional to Odds (MPTO)

The most common method: divide each implied probability by the total.

$$p_{\text{fair},i} = \frac{p_{\text{implied},i}}{\sum_j p_{\text{implied},j}}$$

Assumption: The margin is proportional to each outcome's probability. This is the default "normalize" approach and works well for most markets.

3. Shin Method

Based on Shin's (1993) insider trading model. The idea is that some fraction $z$ of bettors are "insiders" who know the outcome.

For each outcome, the fair probability satisfies:

$$p_{\text{fair},i} = \frac{\sqrt{z^2 + 4(1-z) \cdot q_i / S} - z}{2(1-z)}$$

where $q_i$ is the implied probability, $S$ is the sum of implied probabilities, and $z$ is found via bisection such that fair probabilities sum to 1.

Assumption: Part of the margin comes from the book protecting itself against informed bettors. The Shin method typically gives fair odds that are slightly more extreme than MPTO — favorites become slightly bigger favorites, and longshots become slightly bigger longshots.

4. Odds Ratio (Power Method)

Find an exponent $c$ such that $\sum_i p_i^c = 1$:

$$p_{\text{fair},i} = p_{\text{implied},i}^{\ c}$$

Since $\sum p_i > 1$ at $c = 1$, we need $c > 1$. This is solved via bisection.

Assumption: Margin is applied through a power transformation. This method tends to produce results between MPTO and Shin.

5. Logarithmic

Adjust in log-odds space by subtracting a constant $k$:

$$p_{\text{fair},i} = \text{logistic}\!\left(\log\frac{p_i}{1-p_i} - k\right)$$

where $k$ is found such that the fair probabilities sum to 1.

Assumption: The margin is additive in log-odds space. This preserves the relative log-odds ratios between outcomes.

When Do They Differ?

For balanced markets (e.g., -110/-110), all five methods produce nearly identical results. Differences emerge in unbalanced markets:

Market	EM	MPTO	Shin	OR	LOG
-300/+250	72.50%	73.17%	73.35%	73.25%	73.21%
-110/-110	50.00%	50.00%	50.00%	50.00%	50.00%
+500/-700	14.64%	14.63%	14.50%	14.57%	14.60%

The differences are small (typically < 1%), but when you're making decisions at the margin, even 0.5% matters.

Which Should You Use?

• MPTO is the safest default — it's simple, well-understood, and widely used
• Shin is theoretically grounded and preferred by academics
• When methods agree, you can be more confident in the fair probability
• When methods disagree significantly, the market may be unusual — investigate further

The Bettor Calculator Devig Calculator shows all five methods side-by-side so you can compare.