Return to player and house edge are complements when they use the same wager base. A slot designed for 96% RTP has a 4% theoretical house edge. One figure describes the share returned as prizes; the other describes the share retained by the game model.
[ \text{House edge}=1-\text{RTP} ]
Using percentages:
[ \text{House edge %}=100%-\text{RTP %} ]
The conversion is simple. Interpreting it correctly requires separating theoretical design, actual recorded results, total coin-in, and short-term volatility.
One game, two ledger views
Imagine $100 of total wagering moving through a theoretical 96% slot model.
| Viewpoint | Amount | Percentage |
|---|---|---|
| Expected prizes returned to players | $96 | 96% RTP |
| Expected casino win before other costs | $4 | 4% house edge |
| Total wagering base | $100 | 100% |
The two percentages do not describe separate charges. Adding a 4% house edge on top of a 96% RTP would double-count the same mathematical difference.
This relationship is exact only when both figures use the same denominator and scope. A loyalty rebate, tax, jackpot contribution, fee, or promotion may be reported separately and can change a player’s net economics without changing the underlying game RTP.
Conversion table
| Theoretical RTP | Theoretical house edge | Expected loss per $1,000 coin-in |
|---|---|---|
| 98% | 2% | $20 |
| 96% | 4% | $40 |
| 94% | 6% | $60 |
| 92% | 8% | $80 |
| 88% | 12% | $120 |
The final column is not a session forecast. It is the average price implied by the model over repeated wagering.
Coin-in is the denominator—not the amount deposited
A player deposits $200 and spins $2 a total of 600 times. The relevant wagering volume is:
[ \text{Coin-in}=\text{Bet per spin}\times\text{Number of spins} ]
[ \text{Coin-in}=$2\times600=$1,200 ]
At 96% RTP:
[ \text{Expected return}=$1,200\times0.96=$1,152 ]
[ \text{Expected loss}=$1,200\times0.04=$48 ]
The player did not need to bring $1,200 in cash. Wins may have been wagered again, causing the same money to circulate. This is why a modest deposit can produce much more coin-in and why session speed matters.
The $48 figure is an expectation, not a guaranteed stopping balance. A high-volatility game can lose the full $200 quickly or produce a large win even though its long-run edge remains 4%.
Theoretical RTP and actual RTP are different records
Theoretical RTP comes from the approved probability-and-prize model. Actual RTP is measured from recorded play during a selected period.
[ \text{Actual RTP}=\frac{\text{Prizes paid}}{\text{Coin-in}} ]
Suppose a machine records $10,000 coin-in and $9,450 in prizes:
[ \text{Actual RTP}=\frac{9,450}{10,000}=94.5% ]
The matching actual hold is:
[ \text{Actual hold}=1-0.945=5.5% ]
A 94.5% actual result does not prove the theoretical setting is 94.5%. It may be an ordinary fluctuation around a 96% design. The sample size and game volatility determine how surprising the result is.
The UK Gambling Commission’s actual-RTP guidance uses the same prizes-divided-by-turnover method and specifically notes that volatility and play volume affect the acceptable distance from theoretical RTP.
House edge is not the casino’s final profit margin
The theoretical edge estimates gaming win before many business costs. A casino still pays for:
- machine purchase or participation fees;
- progressive contributions;
- labor and maintenance;
- loyalty rewards and promotions;
- taxes and regulatory fees;
- property, utilities, security, and surveillance.
A 4% slot edge therefore does not mean the casino earns 4 cents of final profit from every dollar wagered. It means the game model expects 4 cents of gaming win before those wider expenses and adjustments.
Similarly, actual hold is an operating result, while house edge is a theoretical parameter. They may converge over suitable volume, but they are not interchangeable on a daily report.
What RTP and house edge do not tell you
Two games can both have 96% RTP and 4% house edge while producing entirely different sessions.
| Feature | Game A | Game B |
|---|---|---|
| RTP | 96% | 96% |
| House edge | 4% | 4% |
| Hit frequency | High | Low |
| Volatility | Low | High |
| Typical pattern | Frequent small returns | Long dry periods, occasional large wins |
The pricing is equal in expectation. The distribution is not. Read slot volatility and hit frequency before treating RTP as a complete description of risk.
RTP also does not reveal the probability of a particular jackpot, the number of losing spins in a row, or how quickly the player will wager. Those require the full paytable, outcome probabilities, and play pattern.
Comparison mistakes that change the answer
Comparing displayed RTP with one session’s result
A player who gets back $40 from $100 of coin-in experienced 40% actual RTP for that sample. The machine’s theoretical RTP did not become 40%.
Applying edge to the starting bankroll
Expected loss is based on total wagering, not merely the original deposit. Recycled wins increase coin-in.
Treating a higher RTP as cheaper at any bet size
A 98% game at $5 per spin can create more expected dollar loss than a 94% game at $0.50 per spin if the number of spins is the same.
For 500 spins:
- 98% RTP at $5: (500\times$5\times0.02=$50) expected loss;
- 94% RTP at $0.50: (500\times$0.50\times0.06=$15) expected loss.
The better percentage does not cancel the much larger action.
Confusing RTP with payout frequency
A return can be frequent but smaller than the bet. RTP counts value, not merely the number of spins showing a prize.
The useful way to use both numbers
Use RTP when comparing how much a game is designed to return. Use house edge when estimating the theoretical cost of total action. Convert between them to check that a claim is internally consistent.
Then add the missing dimensions:
- bet size;
- number of spins;
- volatility;
- hit frequency;
- progressive or side features;
- rebates, promotions, or fees.
For the base concepts, read slot RTP and slot house edge. The expected-loss calculator can translate wager size, speed, time, and edge into an estimated cost, but the result remains an average rather than a prediction of the next session.