Win Rate

“My win rate is $80” is not a complete statement. Per hour, per session, per 100 hands, per $1,000 wagered, or per table day? The denominator determines what the number means.

In casino analysis, win rate can describe an observed result or a mathematical expectation. Those two figures should never be presented as though they are the same.

Four different rates that share one name

A rate without its period and base cannot be compared fairly. A $2,000 win over two hours is not the same performance as $2,000 over 200 hours. A table winning $6,000 per day may look strong until its limits, table hours, drop, theo, and labor cost are compared with another table.

Actual win rate records history

For a player:

Actual Win Rate per Hour = Net Win / Hours Played

Suppose a player records every buy-in, cash-out, chip balance, tip, and session duration. After 60 hours, the complete ledger shows a net win of $4,200:

Actual Win Rate = $4,200 / 60 = $70 per hour

That is a valid description of the sample. It does not, by itself, prove that the player has a $70 expected hourly edge. The sample may contain favorable variance, unfavorable variance, inaccurate time records, omitted expenses, or a mixture of games and rules.

An honest actual-rate record needs a defined scope. Decide in advance whether the ledger includes:

• tips and dealer tokes;
• travel, accommodation, and food;
• cash-equivalent promotions and comps;
• tournament entries;
• rebates or loss discounts;
• unresolved markers or online balances;
• hours spent scouting, waiting, or traveling.

Different scopes can be useful, but they must be labeled. “Table result only” and “net business income after all costs” are not interchangeable.

Expected win rate starts with a real edge

For a player with a demonstrated positive expectation:

Expected Hourly Win = Average Wager × Decisions per Hour × Player Edge

Where:

• average wager is the amount exposed on an average decision;
• decisions per hour is the realistic pace, not the table maximum;
• player edge is the expected return above break-even after the applicable rules and strategy.

Consider an advantage player averaging $100 per decision, receiving 70 decisions per hour, with a verified 0.8% edge:

Expected Hourly Win = $100 × 70 × 0.008 = $56

Over 60 hours, expected action is:

Total Action = $100 × 70 × 60 = $420,000
Expected Win = $420,000 × 0.008 = $3,360

The player's actual $4,200 win equals 1.0% of action, while the model expected 0.8%. That difference does not automatically mean the edge was underestimated. It may be ordinary fluctuation around the expected value.

For standard negative-edge casino play, the same structure produces an expected loss rate. A $50 average wager, 60 decisions per hour, and a 1.5% house edge imply:

Expected Player Result per Hour = $50 × 60 × -0.015 = -$45

Calling recent wins a “win rate” does not reverse the underlying House Edge. Read Loss Rate for the player-facing version of the same relationship.

Why a rate needs uncertainty, not just an average

Expected value gives the center of a distribution. It does not describe the width of the possible short-term results. Blackjack, baccarat, roulette, poker, and slots can all produce samples far from expectation.

A more complete performance statement might be:

Actual result: +$4,200 over 60 hours. Estimated expectation: +$56 per hour under the recorded conditions. Sample remains too small to treat the observed $70 hourly rate as the permanent rate.

The NIST explanation of measures of location and spread is useful background: an average should be interpreted alongside variability and sample size.

For repeatable bets, uncertainty is often modeled from Variance or Standard Deviation. The exact method depends on the game. A poker player's hourly distribution, a card counter's wager spread, and a slot's paytable require different models.

Converting between time and action

If the data are consistent, an action-based rate can be converted to time:

Expected Win per Hour = Action per Hour × Expected Win per Dollar Wagered

And:

Action per Hour = Average Wager × Decisions per Hour

This separation is useful because game speed changes the hourly cost without changing the house edge per dollar. A 1% disadvantage at $5,000 action per hour costs an expected $50 per hour. The same disadvantage at $12,000 action per hour costs an expected $120 per hour.

That is why Decisions Per Hour and Total Action belong in any serious rate calculation.

Casino uses of win rate

Casino departments use several operating denominators:

• win per unit per day for slot or electronic-game comparisons;
• win per table day or win per table hour for table capacity;
• win per occupied seat-hour for utilization analysis;
• actual win versus theoretical win for result review;
• hold percentage for win relative to the approved drop or coin-in base;
• contribution per labor hour for a bounded operating-cost view.

A high short-term casino win rate can be luck, just as a high player rate can be luck. Managers should compare actual results with Theoretical Win, operating volume, game mix, limits, and a suitable rolling period. One large player can dominate a day's result without changing the quality of the operation.

How to test a claimed win rate

Before accepting a rate, ask:

1. What is the numerator? Net gaming result, gross cash-out, casino win, or profit after costs?
2. What is the denominator? Hours, decisions, sessions, action, tables, units, or days?
3. How was time counted? Seat time only, full trip time, or all work time?
4. Are losing and winning records complete? Selective records create a false rate.
5. Is the edge mathematical or inferred from results? Winning alone does not prove positive expectation.
6. How large is the sample relative to the game's variance? A rate can remain unstable for a long time.
7. Did conditions change? Rules, limits, strategy, game speed, promotions, and expenses can all change the rate.

A claimed $100 hourly win rate based on eight favorable hours is a result, not a reliable estimate. A rate supported by a complete ledger, defensible edge model, realistic pace, and sufficient sample is much more informative.

Choose the denominator before looking at the result

The safest way to analyze a win rate is to define the denominator before seeing whether the result is flattering. If a player wins $6,000, the same session can be described as $3,000 per hour, $120 per 100 decisions, or 1.2% of $500,000 in action. Each statement may be arithmetically correct, but each answers a different question.

For player-performance analysis, action-based rates are often more portable than hourly rates because table speed, breaks, and game pace can change sharply. Time-based rates remain useful for bankroll planning and income comparisons, but they should be paired with the amount actually wagered. For casino operations, win per table hour can reveal capacity productivity, while win as a percentage of drop or handle answers a different financial question.

This is why a denominator should never be chosen after the fact simply because it creates the most impressive number.

A winning sample does not create a player edge

Observed results and expected results answer different questions. A roulette player can post a positive hourly win rate for a weekend while still facing a negative expectation on every properly priced spin. An advantage player can post a negative rate for weeks while still having a defensible positive expectation. The direction of a short sample does not identify the underlying edge.

To estimate an expected rate, the analyst needs a model that exists independently of the recent result: rules, paytable, strategy, promotion terms, wager distribution, pace, and any genuine player advantage. The historical ledger is then evidence about what happened around that model, not a replacement for it.

That distinction also protects against survivorship bias. If only the winning players publish their rates, the visible average will be distorted even when every individual calculation is correct.

Casino win rate needs volume context

Casino managers can make the same mistake in reverse by celebrating a strong win day without asking how much business produced it. A baccarat pit that wins $250,000 from one short high-limit session may have an exceptional actual day but a poor indicator of normal production. Another pit may win only $80,000 while producing far more stable theoretical value across many rated players.

A practical operating review separates at least four layers: actual win, theoretical win, wagering volume, and the resource base used to generate the action. That resource base can include table hours, occupied seats, dealer hours, machine days, or marketing reinvestment. A rate that ignores the relevant base can reward luck instead of operating quality.

Rates become more useful when the definition does not change

Consistency is more important than choosing one universal formula. A player can track result per hour and per $100,000 of action. A casino can track win per table day and actual-to-theoretical performance. The problem begins when the definition changes between periods, games, or people without being disclosed.

Before comparing two rates, confirm that the numerator, denominator, treatment of comps and expenses, session boundaries, and currency are the same. If they are not, normalize the data first. A precise percentage built from inconsistent definitions is still misleading.

The useful definition

Win rate is net win divided by a stated base. Actual win rate reports what occurred. Expected win rate estimates the long-run average from edge and volume. Casino win rate measures operating output against a chosen unit. None of them can be interpreted responsibly without the denominator, the time period, and the uncertainty around the result.

Continue with Expected Value, Positive Expectation, Sample Size, Win Per Unit, and Win Per Day.