Expected Loss

Expected loss is the average mathematical cost of a wager or group of wagers. For a simple game with a stable house edge, it is calculated by multiplying total amount wagered by the house edge.

Expected loss = total action × house edge

If a player makes $2,000 of total wagers on a game with a 2% house edge, the expected loss is $40. That does not mean the player will lose exactly $40. A real session can finish far ahead or far behind that amount. Expected loss describes the long-run average of repeated comparable play, not the result of one visit.

The term is useful because it connects a percentage such as house edge to actual money and shows why bet size, game speed, and session length matter.

Expected loss is based on wagering, not the original buy-in

A common mistake is to multiply the house edge by the amount of cash brought to the casino.

Suppose a player buys in for $200 and makes 100 wagers of $20. Total action is:

100 × $20 = $2,000

If the wager has a 1% house edge, expected loss is:

$2,000 × 1% = $20

The $200 buy-in tells you how much money initially entered the session. The $2,000 tells you how much was actually put through wagers. Wins returned during play can be bet again, so total action can be many times larger than the original cash.

That is why total action is central to expected-loss calculations.

House edge is the price rate; expected loss is the money amount

House edge and expected loss are related but not interchangeable.

A low edge can still create a meaningful expected dollar loss when total action is large. A high edge can create a small expected dollar loss if very little money is wagered.

For example:

• 1% edge on $10,000 action = $100 expected loss;
• 10% edge on $500 action = $50 expected loss.

This is why comparing games only by percentage can miss the effect of speed and bet volume.

When every wager has the same edge

If the bet size and house edge stay constant, the formula is straightforward.

A baccarat player makes 80 $25 Banker wagers under a standard game with an approximate 1.06% house edge:

Total action = 80 × $25 = $2,000

Expected loss ≈ $2,000 × 1.06% = $21.20

The calculation is an average. The actual session can easily swing by several hundred dollars because baccarat outcomes are variable.

The baccarat flat betting page shows why a fixed unit makes this kind of estimate easy to audit.

When wagers have different edges, calculate them separately

A single blended percentage can be misleading when the player makes different bets.

Suppose a player makes:

• $2,000 of main-game action at a 1.5% house edge;
• $400 of side-bet action at a 7% house edge.

The expected loss is the sum of the two components:

Main wager: $2,000 × 1.5% = $30

Side wager: $400 × 7% = $28

Total expected loss = $58

The side bet created only one-fifth as much action as the main game, yet almost matched its expected cost because the edge was much higher.

This component method is better than multiplying all $2,400 by one guessed edge. It preserves the real economics of each wager.

Expected loss per hour adds game speed

For a stable average wager and edge, expected hourly loss can be estimated as:

Decisions per hour × average wager × house edge

Example:

• 70 decisions per hour;
• $20 average wager;
• 2% house edge.

Then:

70 × $20 × 2% = $28 expected loss per hour

If the same game speeds up to 140 decisions per hour while the wager and edge remain unchanged, the expected hourly cost doubles to $56.

Game speed does not change the edge per dollar. It changes how quickly dollars are exposed to that edge. The decisions per hour page explains that operational link.

Slot expected loss uses coin-in and theoretical RTP

For a slot, the corresponding house-edge percentage can be written as:

House edge = 100% − theoretical RTP

If a slot has a theoretical RTP of 94%, the theoretical house edge is 6%.

A player makes 500 spins at $1 each:

Coin-in = $500

Expected loss = $500 × 6% = $30

Again, the $30 is not a predicted ending balance. Slot volatility can produce a large win, a total bankroll loss, or anything between. The expected value emerges over repeated action, not as a smooth deduction from the meter.

The RTP glossary distinguishes theoretical return from actual observed session return.

Strategy-dependent games require the correct edge assumption

Expected loss is only as good as the house-edge input.

In blackjack, the effective edge depends on rules and player decisions. A calculation that assumes competent basic strategy will understate expected loss for a player making frequent costly mistakes.

In video poker, return depends on the exact paytable and strategy used. Two machines with the same game family name can have different mathematical returns, and poor holds can reduce the player’s effective return further.

So “expected loss = action × edge” remains valid, but the chosen edge must match the actual rule set, paytable, and decision quality being modeled.

Actual loss can be much larger or smaller

Expected loss is not a cap, guarantee, or prediction interval.

Suppose a player has $100 of expected loss over a session. Possible actual results include:

• win $600;
• win $50;
• lose $40;
• lose $100;
• lose $800.

The existence of a $100 expectation does not make any one of those outcomes impossible. Short-term results are driven by the distribution and variance of the game.

This distinction is essential:

• expected loss is a mathematical average;
• actual loss is the real cash result;
• variance describes how widely actual results can move around expectation.

Read variance for the short-term swing component.

Expected loss can exceed the starting bankroll

At first this sounds impossible: how can expected loss be $150 if the player only brought $100?

The answer is that expected loss is based on action, and action can recycle wins.

Imagine a player starts with $100 but repeatedly wins small amounts and re-bets them, eventually creating $3,000 of total action on a 5% edge game:

$3,000 × 5% = $150 expected loss

The player cannot lose more cash than was actually available without adding funds or credit, but the expected-loss model describes the average cost attached to the accumulated action. Across repeated comparable sessions, many will end before reaching that action and others will survive long enough to generate much more.

This is one reason session-path modeling is more complex than a single expectation figure.

Pushes and ties are normally already embedded in the published edge

Players sometimes try to “correct” the formula by manually removing pushes or ties after using a published house edge. That can double-count an adjustment.

A properly calculated house edge for a wager already incorporates its win, loss, push, tie, payout, and rule probabilities according to the convention used for that bet.

If the edge is reliable for the exact wager, expected loss is normally calculated from action using that edge. Do not subtract pushes again unless the underlying metric is defined in a way that specifically requires it.

Expected value and expected loss are two views of the same average

Expected value can be positive or negative. From the player’s perspective, a casino wager with a 2% house edge has an expected value of approximately −$0.02 per $1 wagered.

Expected loss states the same negative expectation as a positive cost amount:

• player EV on $1 = −$0.02;
• expected loss on $1 = $0.02.

Across $1,000 of action:

• player expected value = −$20;
• expected loss = $20.

The sign convention changes, but the underlying average is the same.

Casino theoretical loss is related but may be an operational estimate

Casinos often use theoretical loss when rating players, forecasting revenue, or estimating reinvestment such as comps.

Mathematically, theoretical loss and expected loss are closely related. Operationally, a casino rating system may estimate the inputs rather than measure every wager exactly. A table-games rating might use:

average bet × decisions per hour × time played × house-advantage assumption

If the average bet or game-speed assumption is imperfect, the casino’s recorded theo can differ from a precise mathematical reconstruction of the session.

That is why “expected loss” in a probability calculation and “theoretical loss” in a casino player-rating system should not automatically be treated as identical database fields.

Comps do not erase the expected cost automatically

A player might receive meals, rooms, free play, points, or discretionary benefits based partly on theoretical value. Those benefits have value, but they should not be assumed to offset expected gambling loss dollar for dollar.

If a session has $200 of theoretical expected loss and the player receives a benefit worth $30 to them, the benefit changes the net entertainment economics, not the underlying game edge. The wager still carries the same negative expectation unless the promotion itself changes the effective return enough to alter the full calculation.

The theoretical loss page covers the casino rating side, while expected value is the better framework for combining all components of a promotion.

A compact way to calculate variable sessions

For a session containing several wager types, use:

Total expected loss = Σ(action on wager i × house edge of wager i)

Example:

The table shows why a small amount of high-edge side action can materially change the session cost.

The term is most useful as a planning and comparison tool

Expected loss answers questions such as:

• What is the average mathematical cost of this amount of play?
• How does doubling the bet affect expected dollars lost?
• How does playing twice as fast affect hourly exposure?
• How much does adding a high-edge side bet change the session?
• Why can a low-edge game still become expensive over a long session?
• Why is total action more informative than buy-in when estimating cost?

It does not answer “How much will I lose tonight?”

A precise expectation can coexist with a very uncertain short-term outcome. That is not a contradiction. It is the difference between an average and an individual realization.

Use house edge for the percentage behind the calculation, total action for the wager volume, and theoretical loss for casino-rating context. The expected loss calculator can compare bet size, game speed, and edge assumptions without presenting the result as a prediction of one session.