Every sportsbook line includes the operator's margin — the 'vig' or 'juice.' On a -110/-110 spread, the margin is 4.5%. On a parlay, it's 12-22%. On a futures market, it can be 25%+. Understanding how the margin is constructed — and how to remove it to find the no-vig (fair) line — is the foundational skill of bet evaluation.
What vig is
Vig is the share of total market action the book retains as expected profit if action is balanced. On a 50/50 market with -110 on both sides, the book takes $110 from each side, returns $210 to the winner, and keeps $10 — a 4.55% hold.
Vig is not a 'tax' on your bet. It's a margin baked into the odds. You don't pay vig as a separate fee; the worse-than-fair odds are the vig.
The math: implied probability sums to >100%
On a -110/-110 spread:
- Side A implied probability = 110 / 210 = 52.4%
- Side B implied probability = 110 / 210 = 52.4%
- Total = 104.8%
The 4.8% over 100% is the vig (in a slightly different definition than the 4.5% hold — convention varies, but the concept is the same).
Removing the vig — the no-vig line
To find the fair (no-vig) probability of each side, divide each side's implied probability by the total:
- Side A no-vig = 52.4 / 104.8 = 50.0%
- Side B no-vig = 52.4 / 104.8 = 50.0%
That's the cleanest expression of what the market thinks: each side has a 50% chance.
Worked example: an unbalanced market
Cowboys -7 (-115) / Eagles +7 (-105). Implied probabilities:
- Cowboys: 115/215 = 53.5%
- Eagles: 105/205 = 51.2%
- Sum: 104.7%
No-vig probabilities:
- Cowboys no-vig = 53.5 / 104.7 = 51.1%
- Eagles no-vig = 51.2 / 104.7 = 48.9%
The market is saying Cowboys are 51.1% to cover. If you think they're 56%, you have a bet.
Why no-vig analysis matters
- Cross-book comparison. Different books offer different vig structures. Comparing raw implied probabilities is misleading; no-vig probabilities normalize the comparison.
- Estimating fair price. If a sharp book (Pinnacle, Cris) posts a market with low hold, the no-vig price at that book is a strong estimate of true probability.
- Detecting line moves. No-vig probability tracks market consensus better than implied probability over time.
- Computing required edge. To break-even at -110, you need 52.4% win rate. To break-even on -120 juice, you need 54.5%. The no-vig perspective shows the gap clearly.
Hold percentages by market type
| Market | Typical hold |
|---|---|
| Spread/total at -110 | 4.5% |
| Spread/total at sharp books | 2.5-3.5% |
| Moneyline (close prices) | 3-5% |
| Moneyline (heavy fav/dog) | 5-8% |
| Player props | 4-8% |
| Same Game Parlay | 14-22% |
| Futures | 15-30% |
| Live betting | 8-13% |
Knowing the hold tells you where to focus volume. Standard game lines have low structural cost; futures markets bake in 20%+ structural margin before you even start.
Two operator hold conventions
The industry uses two slightly different hold conventions:
- Theoretical hold: the margin per dollar of action, balanced (the 4.5% number).
- Actual hold: the operator's reported revenue / handle ratio. This is the post-results version, affected by which side won. Reported quarterly in operator filings.
Industry-wide actual hold has been 8.5-9.0% in recent quarters across all market types — higher than theoretical hold because of unbalanced action and structurally higher-margin products gaining handle share.
Practical workflow
- Look at posted odds on both sides.
- Compute each side's implied probability.
- Compute total — that's where the vig is hiding.
- Normalize each side: implied / total = no-vig probability.
- Compare to your estimate of the actual probability.
- Bet when your estimate exceeds no-vig by enough to overcome the vig (typically 2%+ edge).