Start Free Trial →Free Trial →

Devigging methods compared — which to trust when

Lesson #4 taught the standard way to strip vig from sportsbook odds — proportional devigging. It works on most markets. It doesn't work on all of them. This lesson is the longer story: three devigging methods, when each is right, and the one you should default to.

The setup

Every devigging method does the same job: takes the implied probabilities from a vigged market and converts them into a fair-probability estimate that sums to 100%. They differ in how they remove the vig.

The three you'll see in practice:

  1. Equal-multiplier (proportional) — the simplest, what Lesson #4 used

  2. Additive (logit / log-odds shift) — a small refinement

  3. Power method (Shin or similar) — used for asymmetric markets

For most main markets, all three give answers within a few tenths of a percent. The differences matter at the extremes — heavy favorites, longshots, and three-way markets.

Method 1 — Equal-multiplier

The default. Take the implied probabilities, normalize them so they sum to 1.

Example: Lakers -180 (64.3% implied), Celtics +160 (38.5% implied), sum 102.8%.

  • Lakers no-vig = 64.3 / 102.8 = 62.5%

  • Celtics no-vig = 38.5 / 102.8 = 37.5%

Assumption: the book applies vig evenly across both sides. Real for most markets. Slightly wrong at extremes.

Use it for: spreads, totals, moneylines at reasonable odds, props on main markets.

Method 2 — Additive (logit shift)

Slightly more accurate at the extremes. Instead of scaling probabilities multiplicatively, it shifts them in log-odds space.

The intuition: at -1000 odds (90.9% implied), shaving 1% off costs much less in payout terms than shaving 1% off a 50/50 line. A multiplicative shave under-corrects the favorite.

The math is messier — typically computed in code rather than by hand. The result: heavy favorites devig a little lower (more favorable for fading them), heavy underdogs devig a little higher (less generous payouts implied).

Assumption: vig is uniform in log-odds, not in probability.

Use it for: lines where one side is heavily favored — moneylines on big-favorite games, alt-line markets, futures with one dominant outcome.

Method 3 — Power method (Shin)

The most sophisticated of the three. Designed for markets where the vig is asymmetric — for example, when the book is structurally biased against the underdog because of "informed bettor" risk.

The Shin method introduces a parameter (typically called z) that adjusts how much extra correction is needed for the longshot side. For three-way markets (soccer 1X2, fight outcomes with draw), Shin tends to produce noticeably more accurate fair-probability estimates than the simpler methods.

The math is genuinely involved — solver-based, not closed-form. You don't compute this by hand. Sharp's odds-comparison tool gives you all three devigging options side-by-side.

Assumption: the longshot side carries a structural premium that needs separate correction.

Use it for: three-way markets (soccer, hockey regulation result, MMA with draw), longshot props, futures.

Which one to default to

For 90% of day-to-day betting, the equal-multiplier method is fine. Lower friction, fast enough to compute mentally, accurate enough to find real edge.

Switch to additive when you're working a heavy-favorite line or a moneyline at -300 or below.

Use the power method for three-way markets and longshots where the proportional method visibly under-devigs the favorite. Or just always use it for those — there's no cost to being more accurate beyond a click in the tool.

Worked comparison

Soccer EPL: Manchester City -250, Draw +320, Arsenal +600.

Implied:

  • City: 71.4%

  • Draw: 23.8%

  • Arsenal: 14.3%

  • Sum: 109.5% (about 9.5% vig — typical for three-way)

Equal-multiplier:

  • City: 71.4 / 109.5 = 65.2%

  • Draw: 23.8 / 109.5 = 21.7%

  • Arsenal: 14.3 / 109.5 = 13.1%

Power method (Shin, z ≈ 0.05):

  • City: ~67.0%

  • Draw: ~21.4%

  • Arsenal: ~11.6%

The proportional method over-estimates Arsenal by about 1.5 points. On a futures bet that's a small but real distortion of your edge calculation. The Shin estimate is closer to what the market actually thinks.

When devigging breaks entirely

A few honest limits.

  • Stale markets — if the line hasn't moved in hours on a market that should be reacting to news, no devigging method will produce a useful fair price. The market is wrong, period.

  • Single-book devigging — devigging Pinnacle is a strong fair-price estimate. Devigging the only book that posts the line (a small-market prop, a niche international event) just removes their margin but doesn't tell you what fair is — that book might be 5% off true.

  • Promotional lines — boosted prices, free-bet conversions, and welcome-offer odds aren't honest market prices. Don't devig them — back them out separately.

The bottom line

Default to proportional. Know that the additive and power methods exist for the edge cases. Use Sharp's odds-comparison tool to compute all three at a glance so the method choice never slows you down. The goal is a fair-price estimate you trust — once you have it, the EV math is the same regardless of which method got you there.

Related lessons