Hit frequency vs. RTP: a payout is not always a profit

Hit frequency counts how often a defined event occurs; RTP measures payout value relative to wagers. When “hit” means a round with any payout, it can include a payout smaller than the round’s cost. It is not automatically the chance of increasing your balance.

Establish what counts as a hit and what counts as a round before comparing percentages. This guide uses a fully specified fictional model, not frequency estimates for an actual game.

Define the event first

A frequency needs an event and a denominator. “Rounds with any payout divided by completed paid rounds” answers a different question from “rounds returning more than their full cost divided by completed paid rounds.”

Harrigan and colleagues’ 2011 study, in its Reinforcement Rate section, counts spins with any award as hits, including awards below the amount wagered. The authors describe such payouts presented as wins as “losses disguised as wins” (LDWs). That historical multi-line study supports the distinction, not a universal definition for every developer or current U.S. game. We adopt an explicit positive-payout definition for the teaching model below.

Four useful events are:

• Any payout: the completed round returns more than zero.
• Breaking even: its complete payout equals its complete cost.
• A positive net result: its complete payout exceeds that cost.
• A bonus trigger: a specified feature starts, regardless of its eventual payout.

These categories overlap. A positive net result is also a positive payout. A bonus trigger may occur in a round that eventually pays something. Adding overlapping percentages would count some rounds twice.

One model, three different payout counts

This is an invented probability model, not a real slot, simulation, or recorded session. Each trial costs exactly 1 unit. The payout is the entire amount returned, including any return of that cost. There are no extra charges or separate feature rounds.

Complete payout	Probability per trial	Net change after the 1-unit cost
0 units	50%	−1 unit
0.50 units	30%	−0.50 units
1 unit	10%	0 units
6.50 units	10%	+5.50 units

The probabilities add to 100%. The events have these probabilities:

• Any positive payout: 30% + 10% + 10% = 50%.
• Exactly breaking even: 10%.
• A positive net result: 10%.
• A negative net result: 50% + 30% = 80%.

Half of trials have some payout, yet only one in ten has a positive net result in probability terms. Eight in ten have a negative net result. These are model probabilities, not a promised arrangement of the next ten trials.

RTP weights the payout amounts

The model’s expected payout is:

(0.50 × 0) + (0.30 × 0.50) + (0.10 × 1) + (0.10 × 6.50) = 0.90 units

Dividing by the 1-unit cost gives 0.90 / 1 = 90% theoretical RTP. Expected net change is 0.90 − 1 = −0.10 units per trial. The 50% positive-payout frequency and 90% theoretical RTP describe different properties of the same model.

For a limited what-if calculation, change only the last payout from 6.50 to 5.50 units. The positive-payout frequency stays 50%, and positive-net frequency stays 10%, because the qualifying rows have not changed. Expected payout becomes (0.50 × 0) + (0.30 × 0.50) + (0.10 × 1) + (0.10 × 5.50) = 0.15 + 0.10 + 0.55 = 0.80 units, giving 80% theoretical RTP.

Thus a known payout frequency cannot establish RTP without the payout amounts and probabilities. This event-counting exercise differs from comparing volatility labels. For that broader distinction, read RTP versus volatility.

Bonus frequency needs a separate definition

The model has no bonus feature, so its bonus-trigger probability is zero by construction. This says nothing about a real game’s bonus frequency. A large payout is not, by itself, evidence that a feature was triggered.

For a real document, identify which trigger is counted and in which mode. Does the denominator count paid base rounds, events within a feature, or individual free spins? Is the complete feature payout attributed to its initiating round? Resolve these questions from the source rather than assuming one convention everywhere.

Counting prize combinations also differs from counting rounds with at least one payout. If ten rounds pay and two contain several prize combinations, the combination count could exceed ten while the paying-round count stays ten.

Observed frequency and observed return use different arithmetic

For a defined sample of complete rounds:

observed positive-payout frequency = rounds with payout > 0 / total completed rounds

observed return = total payouts / total wagers

In a separate invented sample, 12 of 40 rounds have a positive payout, totaling 28 units. Each round costs 1 unit. Observed positive-payout frequency is 12 / 40 = 30%; observed return is 28 / 40 = 70%. Neither establishes the underlying theoretical value. With variable wagers, use their actual sum for observed return, not the round count.

The UK Gambling Commission’s monitoring glossary distinguishes designed RTP from measured return and defines turnover as total stakes, including reused winnings. These are definitions in a UK context, not U.S. legal requirements. A deposit cannot substitute for total wagers.

What a useful frequency claim should tell you

Record the event definition, unit of play, included features, configuration, and whether the figure is theoretical or observed. For observations, also record sample size, time period, and wager conditions.

If those details are unavailable, keep the interpretation limited. Do not turn an unexplained percentage into “profitable spins,” a fixed waiting time, or an assurance of breaking even.

The paytable reading guide helps locate award units and feature conditions. This article has not tested a real game or measured an operator’s results. A frequent payout can accompany a negative net result; it is not a reason to expect gambling income.