Original research · Bankroll Guardian
What the vig actually costs, by sport: 315 games of measured bookmaker margin
Published · Methodology and limitations below · No picks, no affiliate links
Key findings
- Across 315 games and 3,437 price observations in six two-way markets, the consensus bookmaker margin clustered tightly between 4.54% and 4.78% — with one clear exception.
- Boxing was the most expensive market we measured at 5.98%, more than a point above every other sport, and it was also the hardest to improve by shopping.
- Taking the best available price instead of the consensus removed about half the margin in most sports: 57% in UFC/MMA, 55% in the WNBA, 53% in the CFL, 50% in MLB — but only 29% in boxing.
- In practical terms, shopping moved the average break-even win rate from about 52.3% to about 51.1% — almost exactly the difference between betting −110 and betting −105 on every wager.
Every bettor knows the sportsbook takes a cut. Far fewer know how big it is, or how much it varies between the sports they bet. We had the raw material to check: our own odds pipeline captures live prices across books every 15 minutes for line-movement history. This study measures what that margin actually was, sport by sport, over roughly six weeks — and how much of it disappears if you simply take the best available price.
Methodology
Source. Our own odds_snapshots table: live moneyline (h2h) prices captured from The Odds API across our tracked set of regulated North-American sportsbooks, bucketed into 15-minute intervals, 25 June – 20 July 2026.
Measure.For each game-and-bucket we converted both sides' prices to implied probabilities and summed them. A margin-free market sums to 100%; the excess is the bookmaker's margin (the overround). We report two versions: the consensus margin, using the average price across books — what you pay betting at a typical book — and the best-available margin, using the highest price on each side across all books — what a disciplined line shopper pays.
Sample. 315 games, 3,437 game-bucket observations, six sports. Only observations with both sides priced were counted. Means and medians agreed closely in every sport (e.g. MLB 104.60 vs 104.37 total implied probability), so these figures are not being dragged by outliers. The NFL was measured but excluded — see limitations.
The measured margin, by sport
| Sport | Games | Consensus margin | Best available | Removed by shopping |
|---|---|---|---|---|
| MLB | 150 | 4.60% | 2.28% | −50% |
| Boxing | 62 | 5.98% | 4.26% | −29% |
| UFC / MMA | 49 | 4.75% | 2.02% | −57% |
| NCAAFthin | 27 | 4.54% | 2.36% | −48% |
| WNBAthin | 19 | 4.78% | 2.13% | −55% |
| CFLthin | 8 | 4.55% | 2.14% | −53% |
“Thin” marks samples under ~30 games — directionally useful, not conclusive. See limitations below.
What stands out
The consensus margin is remarkably uniform.Five of six sports landed between 4.54% and 4.78% — baseball, college football, women's basketball, Canadian football and mixed martial arts all priced within a quarter of a point of each other. That consistency is itself a finding: the market has converged on a standard price for a two-way bet, and it is very close to the familiar −110 on both sides.
Boxing is the outlier, and it is expensive. At 5.98% it carries roughly a third more margin than the next-priciest market. It was also the least responsive to shopping: the best available prices still left 4.26%, removing only 29% of the margin against 48–57% everywhere else. Both facts point the same way — fewer books quote boxing, they agree with each other, and they charge more for it. It is the one market in our sample where the price you are offered is close to the only price available.
Shopping is worth about half the vig. This is the number we did not expect to be so consistent. Across MLB, UFC, NCAAF, WNBA and CFL, simply taking the best available price on each side cut the margin by roughly half. Expressed as a break-even win rate, the average two-way bet in those five sports needs about 52.3% at consensus prices and about 51.1% at the best available ones. (Boxing sits above both figures.)
That 1.2-point gap deserves a moment. It is almost exactly the difference between habitually betting −110 and habitually betting −105 — the textbook illustration of why line shopping matters, which we had been using as an example. Here it falls out of six weeks of real prices without being put there.
Limitations
Sample sizes vary a lot. MLB (150 games), boxing (62) and UFC (49) are solid. NCAAF (27), WNBA (19) and especially CFL (8) are thin — directionally useful, not conclusive on their own.
The NFL was measured and then excluded. Its only captured prices were 31 observations from a single day in late June: early-season lines posted in the offseason, at low limits, which do not price like in-season markets. Including it would have added a sport-sized row built on one day of untypical data, so we dropped it rather than let it anchor a comparison. (For the record it read 4.47% consensus / 1.86% best available — plausible, but not evidence.) We will measure the NFL properly once the season is under way.
Our book set is not the whole market. Consensus means consensus across the regulated North-American sportsbooks we track, which excludes offshore books, exchanges and prediction markets. A wider net — particularly one including low-margin books — would likely find a lower best-available margin, making the shopping benefit larger, not smaller.
Moneylines only, two-way markets only. We did not measure spreads, totals, props, parlays or futures, all of which carry their own (usually higher) margins. And soccer is excluded entirely: its three-way match-result market was being captured without the draw, a bug we found and fixed while preparing this study. Rather than publish a soccer figure we knew was wrong, we left it out — and will measure it properly once the corrected pipeline has collected a full sample.
Six weeks is a window, not a verdict. Margins move with competition, season and event profile. These figures describe mid-2026 pricing at the books we track, not an eternal constant.
What to do with this
The practical reading is not “bet the low-vig sports.” A cheaper market is not an easier one — the biggest markets are cheap precisely because they are the most efficiently priced. The reading is that roughly half of the house edge on a normal bet is optional, and it is removed by a behaviour that requires no handicapping skill: comparing prices before you bet, every time. In boxing, where shopping recovers far less, the honest conclusion is that the market itself is expensive — worth knowing before you make it a staple.
Common questions
- What is the bookmaker's margin (the vig) on a typical bet?
- In our measurements the consensus market margin was about 4.5% to 6% of a two-way market, depending on the sport. MLB, NCAAF, WNBA, CFL and UFC all clustered near 4.5–4.8%; boxing was the most expensive at 5.98%.
- Which sport has the highest vig?
- Boxing, in our sample: a 5.98% consensus margin, well above the ~4.5–4.8% of every other market we measured. It was also the least improvable by shopping — the best available prices still left a 4.26% margin.
- How much does line shopping actually save?
- Taking the best available price instead of the market consensus removed roughly half the margin in most sports — 57% in UFC, 55% in the WNBA, 53% in the CFL, 50% in MLB and 48% in NCAAF. Boxing was the exception at 29%.
- Does a lower vig mean a sport is easier to beat?
- No. A lower margin means a smaller headwind, not an edge. The most heavily bet markets tend to be the cheapest precisely because they are also the most efficiently priced — the discount and the difficulty arrive together. The margin sets how much you must overcome; it says nothing about whether you can.
- Why isn't soccer in this study?
- Soccer's match-result market is three-way (home/draw/away), and our odds pipeline was capturing only the two team prices — a bug we found and fixed while preparing this study. Any soccer margin computed from that historical data reads below 100%, which is impossible, so we excluded it rather than publish a number we knew was wrong.
Related: What is the vig · No-vig & hold calculator · glossary: overround
Cite this study
Bankroll Guardian (2026). What the vig actually costs, by sport: 315 games of measured bookmaker margin. https://www.bankrollguardian.com/research/vig-by-sport
Free to cite with attribution and a link. Questions about the data or method: support@bankrollguardian.com.
See your own numbers instead
Bankroll Guardian tracks your bets and grades every one against the closing line — the honest measure this research points to. Free to start.