The Shin method is one way to strip the bookmaker's margin from a market and read fair chances from what is left. It comes from Hyun Song Shin's model, in which part of the margin is the bookmaker's cover against insiders. It leaves favourites more of their implied chance than plain scaling does, and long shots less.
How to remove the bookmaker margin sets Shin beside the multiplicative, additive and power methods on a soccer market and a 10-runner race.
What the Shin method does, in plain terms
Picture a bookmaker who knows a share z of every dollar bet comes from punters who already know the result. Those punters only back winners, and the costliest winner is a long shot: it pays many times the stake, with little ordinary money on it to cover them. So in Shin's model the bookmaker shortens long shots the most.
Run backwards, the method finds the z that makes the fair chances add up to 100%, then reads each fair chance from its price:
fair chance = (square root of (z x z + 4 x (1 - z) x implied chance x implied chance / market total) - z) / (2 x (1 - z))
With three or more outcomes, z is found by trial, which the fair odds calculator does.
How Shin shifts probability towards favourites
Scaling, the multiplicative method in how betting odds work, divides every implied chance by the market total, so every runner gives up the same share. Shin moves margin off the front of the market and onto the back. With an illustrative 8-runner race at one bookmaker adding up to 114.11%, z comes out at 0.02107:
| Price | Fair price, scaling | Fair price, Shin | Fair chance, Shin minus scaling |
|---|---|---|---|
| $3.10 | $3.54 | $3.39 | +1.19 points |
| $3.90 | $4.45 | $4.31 | +0.74 points |
| $5.00 | $5.71 | $5.59 | +0.35 points |
| $7.50 | $8.56 | $8.63 | -0.10 points |
| $10.00 | $11.41 | $11.84 | -0.32 points |
| $16.00 | $18.26 | $20.27 | -0.54 points |
| $26.00 | $29.67 | $36.78 | -0.65 points |
| $36.00 | $41.08 | $56.69 | -0.67 points |
Working for the favourite: 1 / 3.10 = 0.32258, and 0.32258 x 0.32258 / 1.14106 = 0.09119. Then 4 x 0.97893 x 0.09119 + 0.02107 x 0.02107 = 0.35752, whose square root is 0.59793, and (0.59793 - 0.02107) / 1.95786 = 0.29464, a fair price of $3.39. Scaling gives 3.10 x 1.14106 = $3.54.
Prices and points are worked from unrounded chances. The three shortest prices gain fair chance and the rest lose it: the favourite shortens from $3.54 to $3.39, the outsider lengthens from $41.08 to $56.69.
When the Shin method is used
Shin suits markets with a big margin spread over long shots: race fields and other many-runner markets, such as tournament outrights. If long shots there return less per dollar than favourites, the favourite-longshot bias, scaling leaves their fair prices too short. Erik Štrumbelj's 2014 study in the International Journal of Forecasting found Shin probabilities forecast results more accurately than scaling on the markets it tested.
With only two outcomes, Shin and the additive method land on identical fair prices. The power method also loads margin onto long shots, without the informed-money story.
Risk: Betting involves risk. A method that forecast past markets better does not make any one bet win, and a price above a Shin fair price can still lose. See responsible gambling for limits and support.
The de-vig behind B337's +EV screener
B337's +EV screener takes the margin out of prices across sports markets, exchanges and global market references included, and ranks bookmaker prices above the resulting fair price by edge. You can adjust how its fair price is built, and every edge updates to match. A free account shows live counts with the rows locked, and you place any bet yourself.
Risk: Betting involves risk. Each edge the screener lists depends on an estimated fair price, and there is no guarantee of profit. See responsible gambling for limits and support.