Hold Percentage vs. Vig: How Sportsbooks Actually Profit

Published October 10, 2026 · 9 min read · By the OddsLighthouse Editorial Team

Most bettors who've been around long enough to know what vig is also assume it answers the bigger question: how much does the sportsbook actually make? It doesn't, not by itself. Vig tells you the cut built into a single bet's price. Hold tells you what the book actually banked once every bet on an event settled. Those two numbers are related but they are not the same calculation, and the gap between them is where a lot of sportsbook profit — and plenty of bettor confusion — actually lives.

A quick refresher on what vig prices in

Vig (also called juice) is the markup baked into a bet's odds before anyone places a wager. On a standard -110/-110 spread, each side requires you to risk $110 to win $100, which means the two prices, if you converted both to implied probability, add up to roughly 104.76% instead of a clean 100%. That extra 4.76% is the theoretical edge the book has built into that specific market, assuming it takes an equal amount of money on both sides.

Vig is a pricing concept. It exists the moment the odds are posted, before a single dollar is wagered. If you want the full mechanics of how it's calculated and how to find your own break-even win rate against it, we cover that in our guide to vig and break-even win rate. This article picks up where that one leaves off, because vig alone doesn't tell you what a book's revenue looked like once a game actually happened.

Hold percentage: a different measurement entirely

Hold percentage is a backward-looking accounting number: total money the book kept, divided by total money wagered (handle), across some pool of bets — a single game, a weekend, a full month of NFL action, an entire sport for a quarter. If a book took in $1,000,000 in bets on an NFL Sunday and paid out $940,000 to winners, its hold for that day was 6%, regardless of what any individual line's vig happened to be.

This is also the number you'll see in state gaming-commission reports, because it's the one regulators and operators actually track for revenue purposes. "Hold" in those reports almost always means revenue divided by handle for a given period, and it's reported after free bets, promotional credits, and some adjustments are accounted for — which is one more reason a state's monthly hold figure and a single game's theoretical vig rarely line up on paper.

Worked example: why a 50/50 book and a lopsided book hold differently

Here's where the two numbers actually diverge, and it's easier to see with real figures than with definitions alone. Say a book takes $100,000 on a game at standard -110 both sides: $50,000 on Team A, $50,000 on Team B.

If the action really is split 50/50 and Team A wins, the book pays Team A bettors $50,000 × (100/110) in winnings on top of their stake — roughly $45,455 in profit paid out, for a total return to winners of about $95,455 — and keeps the other $4,545 of the $100,000 handle. That's a 4.55% hold, which lines up almost exactly with the theoretical vig on that market. This is the case sportsbooks actually want: balanced action means the vig alone does the work, with no outcome risk left over.

Now change only one thing: the betting public pours $80,000 onto Team A and just $20,000 onto Team B, still both at -110. Two very different outcomes are now possible:

• Team A (the heavily bet side) wins: the book owes Team A bettors about $72,727 in winnings. Total paid out is about $152,727 against only $100,000 collected — the book loses roughly $52,727 on this game. Hold percentage: deeply negative.
• Team B (the lightly bet side) wins: the book owes Team B bettors about $18,182 in winnings, paying out about $38,182 total against $100,000 collected. The book keeps roughly $61,818 — a 61.8% hold on this one game.

Vig didn't change at all between these two scenarios — every single bet was still priced at -110. What changed was which side the money piled onto and which side actually won. That's the entire reason hold percentage swings so much more than vig ever could: hold is a function of both the pricing and the outcome, while vig is a function of pricing alone.

Why reported hold numbers vary so much by bet type

Over a large enough sample — a full season, a whole state's monthly handle — the lucky and unlucky outcomes above tend to even out, and a book's overall hold percentage starts converging toward something closer to its blended theoretical vig. But it rarely converges to the same number across every bet type, because different bet types carry very different vig to begin with.

Straight bets on spreads and totals at -110 carry a relatively thin, well-understood vig around 4.5%–5%. Moneylines on heavy favorites or underdogs often carry more vig once you account for the rounding books build into lopsided prices. Parlays and same-game parlays carry dramatically more built-in margin per bet, because the vig compounds across legs, which is a major reason state-level hold percentages have climbed in recent years as parlay betting has grown as a share of total handle. Player props, similarly, tend to carry wider margins than mainstream spread and total markets because they're priced on thinner data with less competitive pressure from sharp action.

This is also why a single sportsbook's overall monthly hold can look completely different from its hold on, say, just NFL point spreads that same month — the blended number reflects the entire mix of bet types customers chose, not any one market's pricing.

What hold percentage is actually useful for

For bettors, hold percentage isn't something you can calculate or act on before a game — it's an after-the-fact number, and mostly a tool for understanding the business of sportsbooks rather than for finding an edge in any specific bet. Where it is genuinely useful is as a sanity check on how a book or a market is trending: a sportsbook whose state-reported hold is rising quarter over quarter, even as handle stays flat, is usually a signal that its customer base is shifting toward higher-margin products like parlays and props rather than straight bets — which tells you something about how that book is likely to market and price going forward.

It's also a useful corrective against a common misconception: that a sportsbook's profit is simply "vig times total handle." It isn't. A book's actual revenue in any given period depends heavily on which side won, and across thousands of simultaneous games that risk mostly — but never perfectly — cancels out.

Where the comparison breaks down

Don't read too much into a single game's hold, positive or negative, the way you might be tempted to read into a single bet's closing line value. One lopsided outcome tells you almost nothing about whether a book's pricing is sound — it mostly tells you that one specific game broke a specific way for that book's specific liability on that day. The number only becomes meaningful at scale: a full week, a full month, a full season of a given market.

It's also worth remembering that hold, as reported publicly, is usually calculated after promotional credits and free bets are stripped out in some jurisdictions and included in others, which means comparing hold percentages across two different states' gaming-commission reports isn't always an apples-to-apples exercise. If you're trying to understand a specific number you've seen in a state revenue report, check that report's own methodology notes before assuming it means the same thing a different state's report means.

For the mechanics of the number that actually does drive a single bet's price, start with our guide to how vig works and your break-even win rate, and for how sharp bettors evaluate whether they're beating that pricing over time, see our guide to closing line value.