The vig, not the guesswork
A bookie does not make money by betting against players and hoping. They make money from the vig, also called juice or the margin, which is the cut built into the price. Merriam-Webster defines vigorish as a charge taken by a bookie or a gambling house on bets, and that is exactly what it is.
The standard example is a point spread priced at -110 on both sides. The bettor has to risk $220 to win $200. When the money is roughly even on each side, the bookie pays the winners out of the losers and keeps what is left over.
What that looks like on one game
Two players take the same game, one on each side, at -110.
The loser pays $220. The winner collects $200. The bookie keeps $20, and it does not matter at all which team won.
Spread that across a full slate and a full roster and the vig stops looking like a bet and starts looking like arithmetic. That is the whole design of the business.
The hold: what a book actually keeps
The vig is the theory. The hold is what survives contact with real players.
Real players do not bet in neat opposing pairs. Some are sharper than the rest. Results run hot and cold for weeks at a stretch. The hold is the share of everything wagered that a book actually keeps once all of that has happened, and it is always lower than the vig on a single game suggests.
For scale, Nevada's regulated sportsbooks held 7.75% of everything wagered over the twelve months to 30 June 2026, on the Gaming Control Board's own monthly figures. Those are licensed books at state scale rather than a private book of forty players, so treat it as the order of magnitude rather than a target. On that number, a book taking $40,000 of action in a week keeps about $3,100 of it.
Why limits are the margin
If one player is allowed to bet far more than everyone else, a single game can move the entire book and erase the vig collected from every other player that week.
That is the whole argument for limits, set per player and per sport rather than as one number for everybody. They are not a way of being difficult with players. They are how the operator makes sure the margin the arithmetic is supposed to hand over actually arrives. A book with no limits is not running on the vig; it is gambling with extra steps.
The casino, and the cut some providers take
For a lot of operators a real share of the money comes from the casino rather than the sportsbook. Which makes how the software provider charges for it worth reading closely.
Some providers advertise a low per-player rate and then take a percentage of the casino on top. If players lose $3,000 in the casino in a week, a 10% cut is $300 gone that week on a single line. Across a month that is roughly $1,300 leaving the operator for nothing visible. A provider that includes the casino at no percentage leaves that money where it was.
BookiePerHead takes no percentage of casino or sportsbook action on either tier. What the casino includes sets out what that covers.
How much do bookies make?
There is no fixed number, and anyone quoting one is selling something.
What an operator, or agent, keeps comes down to four things: how many players are active in a given week, how much those players actually bet, how well the limits are set, and what the platform costs. Two agents carrying the same number of players can finish a season a long way apart on all four.
Quality of roster matters more than length of roster. A small, steady group of players who bet every week will usually leave more behind than a long list that barely moves. And the software cost comes straight off the top, which is why a flat per-head fee is easier to plan around than a percentage: it scales with the number of players rather than with the size of the action.
Why the money is rarely balanced
The clean example has one player on each side. Real books almost never look like that.
Players in the same town tend to like the same teams. A local favorite can pull most of a book onto one side of one game, and when that happens the bookie is no longer collecting a margin on that game. He holds a position on it, like any bettor.
There are a few ways to live with that. Limits cap how large any one position can get. A price can move to make the other side more attractive. And over enough games the lopsided ones cut both ways, because the popular side does not win every week. What a small book cannot do is expect each Saturday to balance by itself.
Parlays, and why books offer them
A parlay ties two or more picks into one bet, and every pick has to win. Players like them because a small stake can pay a large amount. Books like them because the margin compounds.
The arithmetic is short. Two legs at -110 pay about 3.64 times the stake. A fair price for two even chances would pay 4 times. That gap is roughly 8.9 percent, close to double the 4.55 percent on a single straight bet. Add a third leg and the payout is about 6.96 times against a fair 8, which is a margin of about 13 percent.
Those figures assume each leg is a true even chance, which real games are not, so read them as a guide to the direction. The direction is reliable, though. The more legs a player adds, the more of the stake stays with the book on average.
The other side of that is the size of a single payout. A parlay that does hit pays far more than the stake, so a book needs a cap on what one ticket can win.
A month in a small book
Averages hide how uneven the weeks are. Take a book that handles 40,000 a week for four weeks.
In the first week the favorites all cover and the book loses 2 percent of handle, which is 800 out. In the second week it holds 9 percent and keeps 3,600. In the third it holds 5 percent, which is 2,000. In the fourth it holds 12 percent and keeps 4,800.
Add the four together and the book kept 9,600 on 160,000 of handle. That is a hold of 6 percent for the month, from weeks that ran from a loss to twice the average.
An operator who judged the business on week one would have quit. One who judged it on week four would have raised every limit. Both would have been wrong, and the month is the smallest stretch worth reading.
What comes off before you count profit
Hold is what the book keeps from the betting. It is not what the operator takes home.
The platform fee comes off first, and it follows the number of active players. If you carry sub-agents, their share of their own players' results comes off next. Then there is the money that was held on paper and never collected. A player who loses 600 and pays 400 has cost you 200 of hold that the weekly figure still shows.
That last line is the one new operators leave out. A book that holds 6 percent and collects all of it is in better shape than one that holds 8 percent and chases a third of it.
Estimating your own week
Rather than chasing a number somebody else promises, run it:
- Count the players who would realistically be active in a normal week, not the ones on the list.
- Estimate what they bet in that week, and apply a hold you would actually accept rather than a hopeful one.
- Subtract the platform: $8 per active player per week on Standard, $11 on Pro, plus $13 for live dealer casino if you take it.
- Subtract any casino percentage your provider takes, if it takes one.
- Do it over a few months. One hot week tells you nothing, and one cold week tells you nothing either.
What the bill is made of works through the cost side properly, and the rate card has the full set of numbers.