Loss rate is the expected amount of money a gambling activity costs over a unit of time or play. It converts a percentage such as house edge into a practical estimate such as $18 per hour or $45 over a session.
The term describes an average mathematical rate. It does not predict the path of one session, and it should not be confused with the amount a player actually loses.
The shortest useful definition
A game’s house edge tells you the expected cost per dollar wagered. Loss rate adds the amount wagered and, when needed, the speed of play.
For a repeating wager:
$$\text{Expected loss per hour}=\text{average stake}\times\text{decisions per hour}\times\text{effective house edge}$$
- Average stake is the average amount exposed on each decision.
- Decisions per hour is the number of resolved wagers, hands, spins, or comparable betting events.
- Effective house edge is the expected loss percentage for the rules, wager and strategy actually being used.
The same relationship can be written as:
$$\text{Loss rate}=\text{action rate}\times\text{house edge}$$
Action rate means total money wagered per hour. The formula measures expected cost, not a guaranteed deduction from the bankroll.
Three games, three very different hourly costs
Consider these simplified examples:
| Game and assumptions | Hourly action | Effective edge | Expected loss rate |
|---|---|---|---|
| Blackjack: $25 average bet, 70 hands, 0.50% edge | $1,750 | 0.50% | $8.75 per hour |
| Double-zero roulette: $10 per spin, 45 spins, 5.26% edge | $450 | 5.26% | about $23.67 per hour |
| Slot: $1.20 per spin, 500 spins, 96% RTP | $600 | 4.00% | $24.00 per hour |
The blackjack estimate assumes a specific rule set and competent play. A different payout, side bet, or strategy error changes the effective edge. The slot estimate uses 1 − RTP as the theoretical hold percentage.
This comparison shows why a percentage alone is incomplete. A low-edge game played quickly at a large stake may cost more per hour than a higher-edge game played slowly for small amounts.
Session loss rate and total expected loss
For a session of fixed length:
$$\text{Session expected loss}=\text{loss rate per hour}\times\text{hours played}$$
A player with an estimated $18 hourly loss rate who plays for 2.5 hours has:
$$18\times2.5=$45\text{ expected loss}$$
That $45 is the center of the long-run calculation. The player could finish the session ahead $400, down $600, or near the expected number. Variance determines how widely actual results can move around the expectation.
Mixed wagers need separate calculations
One average percentage can mislead when a player combines wagers with different costs. The cleaner method is to calculate each component and add the results:
$$\text{Total loss rate}=\sum_i (\text{action rate}_i\times\text{edge}_i)$$
Suppose a blackjack player makes:
- $25 main bets for 60 hands per hour at an assumed 0.60% edge; and
- $5 side bets on 30 of those hands at an 8% edge.
Main-game expected loss:
$$25\times60\times0.006=$9$$
Side-bet expected loss:
$$5\times30\times0.08=$12$$
The side bet uses only one tenth as much action as the main game, yet it creates the larger expected cost. Combined loss rate is $21 per hour.
Loss rate can also be measured per 100 decisions
Time is not always the best denominator. A crowded table, a novice dealer, a machine pause or a long dispute can change decisions per hour without changing the cost of each wager. For comparing the wagers themselves, calculate expected loss per 100 decisions:
$$\text{Expected loss per 100 decisions}=\text{average stake}\times100\times\text{effective edge}$$
A $20 wager at a 1% edge costs an expected $20 per 100 decisions. A $5 wager at an 8% edge costs an expected $40 per 100 decisions. The second wager is smaller, but its price per block of play is twice as high.
This measure is useful when pace is unknown. Once a realistic speed is available, it can be converted back into an hourly estimate.
Use a range when the inputs are uncertain
Loss-rate estimates often look more exact than the observations justify. If average bet may be $20–$30, pace 60–80 decisions per hour, and effective edge 0.5%–1%, the plausible hourly range is:
- low estimate: $20 × 60 × 0.5% = $6;
- high estimate: $30 × 80 × 1% = $24.
Reporting an estimated $6–$24 per hour is more honest than selecting one convenient input and presenting $13.47 as certainty. The range also shows which behavior changes the estimate most.
What loss rate is not
Loss rate is not any of the following:
- House edge: a percentage of action, not dollars per hour.
- Actual session loss: the observed result after wins, losses and cash-outs.
- Casino hold percentage: a reporting ratio that may compare casino win with drop or buy-in, depending on the game and jurisdiction.
- Risk of ruin: the probability that a bankroll reaches a specified loss boundary.
- Volatility: the size and frequency of swings around the expected result.
These measures answer different questions. Expected loss asks what repeated action costs on average. Risk of ruin asks whether the bankroll can survive the swings. Loss rate expresses expected cost against a clock or another unit of exposure.
Actual loss rate can be calculated, but it means something else
A player can divide a real net loss by time played:
$$\text{Observed loss rate}=\frac{\text{net session loss}}{\text{hours played}}$$
If someone loses $300 in two hours, the observed rate is $150 per hour. That figure describes what happened in that session. It does not prove that the game’s mathematical loss rate was $150 per hour.
A single jackpot can produce a large negative observed loss rate from the casino’s perspective. A rapid losing run can produce a large positive observed loss rate for the casino. Neither short result changes the underlying expectation.
How slot loss rate is estimated
For a slot or other game quoted by RTP:
$$\text{Expected loss rate}=\text{coin-in per hour}\times(1-\text{RTP})$$
If coin-in is $900 per hour and theoretical RTP is 95.5%:
$$900\times(1-0.955)=900\times0.045=$40.50$$
The calculation requires the correct RTP configuration and an honest estimate of coin-in. A player who increases the bet after a bonus, uses a faster spin mode, or plays two machines changes the action rate even when the displayed game is unchanged.
The UK Gambling Commission’s rules and likelihood standard explains that game information should make the rules, payouts and likelihood of winning available. Those details are the starting inputs; they do not remove the need to estimate actual betting volume.
How casinos use the idea
Casinos usually use the related term theoretical win or theoretical loss when rating play. A table-game estimate may combine average bet, time, game pace and a configured advantage. A slot system can use tracked coin-in and theoretical hold.
The casino is not claiming that the player lost the calculated amount. It is estimating the expected value of the recorded action. That distinction matters for player ratings, comp decisions and performance analysis.
A practical estimate for a planned session
A useful planning sequence is:
- Identify the exact wager and rules.
- Estimate the effective edge rather than using the best advertised edge for a different version.
- Estimate average stake, including side bets and raises.
- Estimate realistic decisions per hour.
- Multiply by the intended session length.
- Treat the answer as an average cost, not a maximum loss.
A session budget should still be based on money the player can afford to lose. Variance can exhaust the budget long before the expected-loss figure is reached.
Related reading
Continue with house edge, total action, decisions per hour, theoretical loss, and session length. The expected loss calculator can test different bet sizes and speeds without treating its output as a prediction of one session.