Hold

Hold is the amount or percentage of gambling money that a casino retains after paying winning players, but the denominator changes by business area. In slots, hold is commonly measured against coin-in. In table games, reported hold is commonly measured against drop. In sportsbooks, hold is commonly measured against handle.

That means “the casino held 10%” is incomplete unless you know 10% of what.

Hold is also different from house edge. House edge is a mathematical property of a wager before results are known. Hold is a business result measured from actual gambling activity over a period.

One word can describe a dollar result or a percentage

Casino staff use hold in two related ways.

Dollar hold is the amount the casino retained:

Hold amount = Gambling inflow used by the report - Amount returned to players

The exact accounting inputs depend on the department and reporting system.

Hold percentage expresses that result relative to a selected base:

Hold percentage = Casino win / Reporting base × 100

The critical phrase is reporting base. A slot director, table-games manager, and sportsbook manager may all use the word hold while dividing by different forms of volume.

These percentages should not be compared as though the denominators were identical.

Slot hold uses repeated wagering volume, not cash inserted

On a slot machine, coin-in is the total amount wagered through the game. It is not the same as the cash a player inserted.

A player can insert $100, win credits, re-bet those credits, and generate $500 or $1,000 of coin-in before cashing out or losing the balance. The same dollar can therefore contribute to coin-in more than once as it is recycled through successive spins.

Suppose a bank of machines records:

Coin-in = $500,000
Casino win = $40,000

Then:

Hold percentage = $40,000 / $500,000
                = 8%

That does not mean players collectively inserted $500,000 in cash and lost $40,000 of it. It means $500,000 of wagering volume passed through the games and the casino’s net result for the period was $40,000.

The complementary percentage is sometimes described as actual payback for the period:

Actual payback = 100% - actual hold
               = 92%

Over a short sample, actual hold can differ substantially from a machine’s theoretical hold because jackpots, bonuses, volatility, and ordinary random variation do not arrive evenly.

Table hold usually uses drop rather than every dollar wagered

Table games create a different denominator problem.

Drop generally refers to money, instruments, or approved transactions entering the table’s drop system during the reporting period. It is closer to money brought into the game than to the total value of all bets made.

A player can buy in for $1,000 and place dozens of $100 wagers using the same chips. The table may generate many thousands of dollars of actual betting decisions even though the original drop associated with that player was only $1,000.

If a pit reports:

Drop = $100,000
Table win = $20,000

then the reported table hold is:

Table hold = $20,000 / $100,000 = 20%

A 20% table hold does not mean the underlying bets had a 20% house edge. The table may have cycled the drop through many rounds before the players left. Time on device, game speed, bet mix, player decisions, credit activity, and luck all affect the final win relative to drop.

This is why comparing a 20% table hold with an 8% slot hold without context is misleading. The table percentage is being measured against drop; the slot percentage is being measured against coin-in.

See Drop for the table-volume side of the calculation.

Sportsbook hold is usually win divided by handle

In sports betting, handle is the amount wagered. If bettors stake $1,000,000 and the book’s net win for the reporting basis is $70,000, a simplified hold calculation is:

Sportsbook hold = $70,000 / $1,000,000 = 7%

Again, the percentage is a realized result. It can move sharply with the outcomes of popular events, pricing, promotions, parlays, bettor mix, and the timing of settled wagers.

A month in which favorites win unusually often can look very different from a month in which heavily backed teams lose. That short-run change does not mean the sportsbook’s pricing model suddenly acquired or lost a fixed mathematical edge equal to the reported hold percentage.

Hold and house edge answer different questions

The cleanest distinction is temporal.

House edge asks:

Before the wager is resolved, what proportion of the amount wagered is the casino expected to earn in the long run under the stated rules and strategy assumptions?

Hold asks:

After gambling activity occurred, how much did the casino actually retain relative to the reporting base?

A roulette wager can have a known mathematical house edge while a particular table finishes the shift with negative hold because a player won a large amount. Conversely, a low-edge game can show unusually high hold during a favorable casino run.

A good casino report keeps those concepts separate rather than using actual hold as proof that the underlying game math changed.

Theoretical hold and realized hold should not be confused

Casino systems may use terms such as expected hold, theoretical hold, actual hold, or realized hold. The exact labels vary, but the distinction is useful.

A theoretical figure is based on the configured game math, expected player behavior, known pricing, or a management model. A realized figure is what happened in the measured period.

Suppose a slot product has a long-run theoretical hold of 8% under its approved configuration. A short reporting period might show 2%, 14%, or even a negative result if a large jackpot dominates a small sample. As coin-in accumulates, the actual percentage may move closer to the theoretical value, but there is no rule saying it must do so smoothly.

Table games can be even noisier because the denominator is drop rather than true wager volume. One high-value player can materially change a pit’s daily hold percentage without changing the rules of any game on the floor.

The related terms Expected Hold and Realized Hold are useful when a report needs that distinction explicitly.

A negative hold is possible over a short reporting period

If players win more from the casino than the reporting calculation treats as the relevant inflow, the casino can show a negative win or hold result for that period.

That does not imply the game became player-favorable in mathematical expectation. It means actual outcomes landed on the player side strongly enough to overwhelm the expected advantage during the sample.

For a table, a large winning player can turn a normally profitable shift negative. For a slot bank, a jackpot can dominate the day’s result. For a sportsbook, one heavily backed event can move the book from positive to negative for the period.

This is why experienced operators do not investigate every low or negative hold as though it proves an operational failure. They first ask whether the result is plausible variance, and then whether any operational evidence suggests something more.

High hold is not automatically good news

Management can also misread a very high hold percentage.

A strong result may simply be favorable short-run variance. It can also coexist with warning signs:

• unusually low volume;
• one large player losing heavily;
• a change in game mix toward higher-edge wagers;
• reduced playing time because customers lost too quickly;
• data-capture problems in drop, coin-in, or win;
• settlement errors;
• promotions or credits recorded in a different accounting period.

A casino therefore reads hold with volume, theoretical expectation, player mix, game speed, and sample size.

High hold is revenue, but it is not automatically evidence of excellent operations. Low hold is disappointing revenue, but it is not automatically evidence of weak operations.

Hold is especially dangerous in cross-department comparisons

Imagine a management meeting with three statements:

• “Slots held 8%.”
• “Blackjack held 19%.”
• “Sports held 6%.”

Those figures look comparable because all are percentages. They are not.

The slot figure is likely based on coin-in. The table figure is likely based on drop. The sportsbook figure is likely based on handle. Each denominator captures a different kind of customer activity.

A meaningful comparison asks:

1. What is the numerator—actual win, adjusted win, or another revenue measure?
2. What is the denominator—coin-in, drop, handle, or something else?
3. What period is being measured?
4. Are jackpots, free play, promotional credits, taxes, or unsettled wagers treated consistently?
5. Is the figure actual or theoretical?

Without those answers, a hold percentage is only a number with a percent sign.

Hold can flag a question without proving the cause

Operators use hold as an exception signal because unusual results deserve explanation. A table-games manager may ask why baccarat hold is far below expectation; a slot manager may ask why one bank is holding far above its configured range; a sportsbook manager may ask why a month finished negative.

But the investigation should move from result to evidence.

For table games, relevant context may include player concentration, average bet, decisions per hour, credit, fills and credits, game protection events, and unusually large wins or losses. For slots, it may include coin-in, jackpots, denomination mix, theoretical configuration, meter data, and machine status. For sports, it may include event mix, promotional pricing, parlays, and major bettor exposure.

Hold tells management where to look. It does not tell management what happened.

The practical definition to remember

When someone says hold, translate the word into a full sentence:

The casino retained this amount, or this percentage, from a stated base of gambling activity during a stated period.

Then identify the base.

• Slot hold: usually win relative to coin-in.
• Table hold: commonly win relative to drop.
• Sportsbook hold: commonly win relative to handle.
• House edge: not hold; it is the long-run mathematical price of a wager.

That one denominator check prevents most hold-related confusion.

Continue with Hold Percentage for the ratio itself, House Edge for wager mathematics, Actual Win for realized casino result, and Gross Gaming Revenue for the broader revenue vocabulary.