The Gap #5: the book and the crowd disagree on Steveson, Gable vs Ellison, Elisha

Prediction markets and crypto books disagree most sharply in the cage this week, with one MMA underdog priced so differently across the two sources that the gap is impossible to ignore.

The fights

The standout number in this entire digest is Elisha Ellison at +1351 (14.51) on Cloudbet for her bout against Gable Steveson on 11 July. Polymarket has her winning at 15.9%, which strips back to a fair-price implied of roughly 6.29 — meaning the crowd thinks she is almost two and a half times more likely to win than the book's vig-stripped 6.6% implies. The book price translates to a 6.6% chance; the crowd says 15.9%. That 9.3 percentage-point gap is the widest in the dataset by some distance, and at a price of 14.51 it has genuine money significance: you are being offered what the market treats as a 6.29 fair price when the prediction crowd says the correct price should sit closer to 6.30. If Polymarket's sharper crowd has the edge here, the book is mispricing this by a substantial margin.

Conor McGregor to beat Max Holloway on 12 July shows a 3.7pp gap — Polymarket at 33.7% against a vig-stripped 30% at Duelbits. The crowd fair price comes out at 2.97 and the book offers 3.15. That is a real, if modest, edge. McGregor fights carry enormous public money that often distorts bookmaker lines in one direction; here it appears the book has drifted slightly generous on the Irishman relative to where the prediction market sits. Three points is not a screaming overlay, but it clears the threshold and the price is live.

The football

Argentina versus Cape Verde on 3 July produces the widest football gap in the set: 3.3pp, with Polymarket at 84.1% and Stake's vig-stripped line at 80.8%. The crowd fair price is 1.19; the book offers 1.16. The direction is correct — the market thinks Argentina more likely than the book — but at these short prices a 3pp gap barely moves the needle in absolute terms. Still, it clears the 2pp threshold and is worth noting. France to beat Sweden kicks off on 30 June with a 3.1pp gap, Roobet offering 1.28 against a crowd fair price of 1.29. The gap is directionally real but the prices are nearly identical in decimal terms, which illustrates the limitation of percentage-point analysis at high probability: 3pp at 75% converts to a much smaller absolute edge than 3pp at 50%.

Germany versus Paraguay and Spain versus Austria both sit at 2.3pp and 2.2pp respectively. They are genuine gaps on the right side of the threshold, but at odds of 1.36 and 1.31 the practical edge per unit staked is thin. Both are worth considering only if you are comfortable staking volume at short prices to extract small theoretical edges.

The skips

Several markets look appealing on the surface but do not hold up. USA to beat Bosnia and Herzegovina shows just 1.7pp, Norway to beat Ivory Coast 1.5pp, and Pinas over Almeida only 1.4pp — all below the 2pp minimum we require before calling something an edge. The draws in Portugal versus Croatia and Netherlands versus Morocco each come in at 0.9pp, which is noise. Congo DR to beat England is the one entry in the dataset where the book is actually shorter than the crowd fair price, meaning the market implies worse value than Polymarket suggests — a clear skip regardless of the gap direction.

Odds move constantly and the price quoted at any book is the only price that matters when you place the bet.