696JILI: Hit Frequency vs Payout Size, With Worked Examples

The identity behind every paytable

RTP = hit frequency × average win size, where the average win is measured in stake units and both figures are taken across the whole outcome distribution. That single identity explains almost everything about how a slot feels. If a designer wants a game to pay more often, the average win has to shrink; if the designer wants a 15,000× ceiling, the frequent small wins have to be thinned out to pay for it.

Hit frequency is the percentage of spins that return anything at all, including returns smaller than the stake. That definition is doing a lot of quiet work, and we come back to it below, because a great many advertised 'wins' are net losses.

All three cost an identical expected ₱350 per ₱10,000 wagered. Which one you should play is a question about your bankroll and your tolerance for drawdown, covered on our volatility page, not a question about value.

A full worked distribution: 1,000 spins at ₱10

Here is a complete, internally consistent model of a 96.50% slot at a ₱10 stake over 1,000 spins, which is ₱10,000 of turnover. The probabilities sum to 100% and the returns sum to exactly ₱9,650. This is representative of how a modern feature-driven slot is built, not a claim about one specific title.

Three numbers fall straight out of that table. Hit frequency is 265 ÷ 1,000 = 26.5%. The base game alone returns ₱5,900, which is 59.0% RTP. The feature returns ₱3,750, which is 37.5 percentage points, or 38.9% of everything the game gives back. Miss the feature and you are playing a 59% game.

Most 'wins' are losses

Of the 265 winning spins in the table, 170 pay either half the stake or exactly the stake. That is 64.2% of all celebrations on screen returning ₱5 or ₱10 on a ₱10 bet, and together those 170 hits account for only ₱1,200 of the ₱9,650 returned, or 12.4% of the game's payback. The sound, the animation and the counter are identical whether you won ₱5 or ₱500.

This matters when you compare games on hit frequency alone. A slot advertising a 45% hit frequency may simply be paying 0.2× and 0.5× returns constantly, which registers as a win in the statistics and as a loss in your balance. Look for the net hit frequency, meaning the share of spins that return more than the stake. In the model above that is 95 spins in 1,000, or 9.5%. Roughly one spin in ten actually puts you ahead on that spin.

• Hit frequency (any return): 26.5% of spins.
• Net hit frequency (return above stake): 9.5% of spins.
• Spins returning 10× or more: 1.00%, i.e. one in a hundred.
• Spins returning 100× or more: 0.08%, i.e. one in 1,250.
• Share of total payback delivered by the feature: 38.9%.

Why the median session is far worse than the average

Take a 100-spin session at ₱10, so ₱1,000 of turnover. The expected return is 96.50% of ₱1,000 = ₱965, an expected loss of ₱35. That figure is technically correct and practically misleading, because the feature fires once in 200 spins.

The chance of triggering it at least once in 100 spins is 1 − (199 ÷ 200)^100 = 39.4%. So 60.6% of these sessions contain no feature at all and return only the base game's 59%, which is about ₱590 on ₱1,000 staked: a loss of ₱410, or 41% of turnover. The other 39.4% of sessions average about 1.27 triggers, worth roughly ₱952 of feature money on top of the ₱590 base, for about ₱1,542 returned.

Check the arithmetic: (0.606 × ₱590) + (0.394 × ₱1,542) = ₱357.5 + ₱607.5 = ₱965. The average is real, but it is an average of two very different experiences, and the more common one loses 41% of turnover. This gap between the mean and the median is the single most misunderstood thing about slots, and it widens as volatility rises.

How long the dry runs are supposed to be

With a 26.5% hit frequency, 73.5% of spins return nothing. The probability of ten consecutive blanks is 0.735^10 = 4.6%, which sounds rare until you count how many ten-spin windows a session contains. The expected longest blank run in 1,000 spins is approximately ln(1,000) ÷ −ln(0.735) = 6.91 ÷ 0.308 ≈ 22 spins.

So a stretch of about 22 spins with nothing at all is the typical worst run in a normal thousand-spin session on a medium game. On a higher-volatility title with a 19% hit frequency the same calculation gives ln(1,000) ÷ −ln(0.81) ≈ 33 spins. Those runs are the design working correctly, not a machine going cold, and they are already fully priced into the 96.50%.

Hit frequency and house edge are independent: the table-game proof

If you want proof that winning often and winning value are unrelated, roulette settles it. Every bet on a European wheel carries the identical 2.70% house edge because every bet is paid as though the wheel had 36 pockets while it actually has 37, yet the hit frequencies range from 2.70% to 48.65%.

Note the two extremes in that table. Baccarat's banker bet wins nearly half the hands dealt and costs 1.06%; the tie bet on the same table wins less than one hand in ten and costs about 14.4%, more than thirteen times as much per peso. Frequency told you nothing about price. Only the edge did.

Why no staking system converts this into a profit

The best-known system is the Martingale: double after every loss so the eventual win recovers everything. On European roulette red or black, starting at ₱100, seven consecutive losses require an eighth bet of ₱12,800 and total staked of ₱25,500 to net ₱100. Seven consecutive losses occur with probability 0.5135^7 = 0.94%, roughly once in every 106 attempts, and most tables cap even-money bets well before you get there: with a ₱10,000 maximum you can only cover six doublings from a ₱100 base.

The deeper reason it fails is not the table limit. Expected loss equals the house edge multiplied by total turnover, and no rule for choosing the size of the next bet changes the edge on any individual bet. A system that stakes more after losses simply raises turnover, and raising turnover on a 2.70% or 3.50% game raises the expected loss in exact proportion. The same applies to every progression, every 'due' theory and every purported pattern on a slot: outcomes are independent, so past results carry no information about future ones.

What the numbers on this page genuinely support is narrower and more useful: choose the highest certified RTP tier you can access, read the paytable to find where the return is concentrated, size your stake so you can reach the feature that holds it, and treat the difference between the mean and the median as the real cost of volatility. The final page in this series follows that logic into progressive jackpots, where a slice of the headline RTP is diverted into a prize most players will never collect.