Slot Volatility, Measured: What the Label Tells You and What It Does Not

What a volatility label promises

Volatility, in the sense the studios use it, is a statement about the shape of a game’s payouts: a low-volatility slot is supposed to pay often and in small amounts, a high-volatility slot rarely and in large ones, with the same long-run return underneath. The label is printed on the rules page as one of three words, sometimes as a row of lightning bolts, and it is the number-free part of every rules screen. Of the 252 tests in the lab, 151 carry a high label, 60 a medium one, 28 a low one and 13 none at all. The labels are not spread evenly across studios: as of 30 September 2026 every Nolimit City build in the set and all but one Relax Gaming build were labelled high, while Pragmatic Play used all three words across its games.

The lab does not measure volatility as one number, because a single word cannot hold what a session feels. It measures four things from every logged run. The hit rate: the share of rounds that paid anything. The longest dry run: the most consecutive rounds without a win. The drawdown: the deepest fall of the balance from its best point, in bets. And the spread per round: the standard deviation of a single round’s return in bets, the quantity a statistician would call volatility, recovered from the 95% band printed on every review. A game whose typical round returns a few tenths of a bet and whose rare round returns a thousand has a large spread; a game that pays a bet or two on every third spin and tops out at 200x has a small one.

The labels, side by side

Read the table left to right and the label stops meaning what it is assumed to mean. The mean hit rate of the low-labelled games is 27.5%, of the medium ones 25.3% and of the high ones 26.3%: three words, one frequency. The mean longest dry run is 32 rounds under the low label and 32 under the high one, which is the wrong way round if the label described how long a player waits. The median drawdown is 1,414 bets for low, 1,445 for medium and 1,627 for high.

What the label does track is the tail. The median largest round rises from 332x under the low label to 575x under the high one, and the mean spread per round from 8.0 bets to 11.7, though the medians, 7.0, 8.9 and 9.5 bets, sit close together: the high column’s mean is pulled up by a handful of games whose one enormous round widened their band. The label, in other words, is a statement about the ceiling, not about the rhythm.

Volatility is not hit rate

The two ideas are usually bundled together, and the logs pull them apart. Across the lab the correlation between a game’s spread per round and its hit rate is −0.08: none. The correlation between the label and the hit rate is −0.02, between the label and the longest dry run −0.01, and between the label and the spread 0.19, a weak but real link. How often a game pays and how big its rare round is are two separate dials, and the studios’ word turns only the second one.

The runs logged by 30 September 2026 give the extremes. Wild Beach Party is labelled high, paid on 40.4% of rounds and had one of the widest spreads in the lab, 26.8 bets per round, because its cluster grid holds 3,013x rounds. Money Cart 2 is labelled high, paid on 10.1% of rounds in its first run, sat 60 rounds without a win at its longest and had one of the narrowest spreads, 5.4 bets, because nothing in that run paid above 271x. Both words are the same; nothing about a session is. Gravity Bonanza, labelled medium, paid on 37.4% of rounds and still measured 21.8 bets of spread, wider than most of the high-labelled set, on the strength of an 1,851x round.

Every test by measured spread

Each row is one label. The thin line runs from the narrowest test in the group to the widest, the thick bar covers the middle half of its tests, and the dot is the median, on a square-root axis so the few enormous values do not flatten the rest. The three bars sit almost on top of each other and the medians are within a bet or two; what separates the high row is the length of its line, drawn out by a handful of games whose single largest round was worth thousands of bets and whose measured spread is therefore itself a volatile number. The widest spread in the lab belongs to Top Dawg$ at 68.9 bets per round, the narrowest to Riot: Urban Wilds at 3.4; the lab-wide median is 8.8. 66 of the 151 high-labelled games sit below that median, and 7 of the 28 low-labelled games sit above it.

Labels that fit least

Two habits explain most of the first table. Some studios label almost everything high: in the set logged by 16 October 2026, Money Cart 4, Money Train and Dead Man’s Trail all carry the word with spreads of 4.6 to 6.1 bets per round and largest rounds of 161x to 316x, which is the measured profile of a medium game by any other studio’s usage. And a high label sometimes describes a ceiling the run never approached: Diamond Mine Megaways is capped at 50,000x and its widest round, in the run logged on 16 October 2026, was 388x, so the label was about what the maths allows, not about what the session showed.

The second table is the more interesting one. Wild West Gold carries a low label from Pragmatic Play and, over three runs to 15 October 2026, produced a 1,476x round and the widest spread of any low-labelled game, 17.6 bets. The Big Bass editions labelled low, Day at the Races and Amazon Xtreme, paid on about 12% of rounds, went 52 and 72 rounds without a win and waited up to 901 rounds for a feature: sessions that nobody would describe as low-volatility, under a label that seems to mean small payouts rather than frequent ones. Among the medium labels, Eye of Horus Megaways measured 24.7 bets of spread on a 3,450x round and Bigger Bass Splash 17.8 bets on a 1,226x round in the runs logged by 15 October 2026.

Dry runs follow the hit rate, not the label

The longest dry run in the lab is 114 rounds, in Eye of Horus Megaways, and the table above it is a list of low hit rates, not of high labels: the correlation between hit rate and longest dry run is −0.84, the strongest relationship in this dataset. The arithmetic behind it is short. A game that pays on 12% of rounds is expected to produce a blank stretch of about 70 rounds somewhere in 10,000; a game that pays on 26%, the lab’s typical figure, about 30; a game that pays on 40%, about 18. Those three numbers span most of the dry-run table, and the label plays no part in them.

That is why the low-labelled Big Bass games appear next to Tombstone R.I.P and the Money Cart games in the list of the longest waits: fishing slots pay on one round in eight, so a run of fifty or seventy blank rounds is routine for them, whatever the rules page says about volatility. The review of every game lists its own longest dry run and its hit rate, and those two figures describe the wait better than any word.

What the label does to a bankroll

If the label described risk to a bankroll, the fixed stakes the lab tracks should have lasted longer under the low label. They did not. The median round on which a 50-bet bankroll ran out is 119 under the low label, 121 under medium and 129 under high, and the median drawdown of each group is within a few hundred bets of the others. A 50-bet bankroll is small enough that almost every test exhausts it, but it exhausts it at the same pace under every label, because the pace is set by the hit rate and the median win, and those do not move with the word. What the label changes is what happens on the rare round that interrupts the decline: how high the balance jumps when it jumps, not how fast it sinks.

High volatility slots, measured

Every test whose studio prints a high label, in the lab’s register, sorted by the largest ordinary round: the measured return next to the declared one, the hit rate, the purchase value, natural bonuses per thousand rounds and the longest dry run, each with its meter. Sort by hit rate to see how little the label says about frequency; a new test with a high label joins the list the day it is published.

Low volatility slots, measured

Every test with a low label, sorted by hit rate. The spread between the top and the bottom of this short list is wider than the difference between the labels: a low-labelled game can pay on four rounds in ten or on one in eight, and the review behind each row records which.

How to read the label

Use it for the one thing it tracks: the size of the rare round. A high label on a game with a 10,000x cap means that the return is carried by events that a 10,000-round session may not contain, which is also why the measured RTP of such games swings widest between runs. For everything a session is made of, the wait between wins, the length of a blank stretch, the slope of the balance, read the hit rate and the longest dry run on the review page instead; they predict the experience, and the label does not. Treat the spread figure itself with the same care: in a heavy-tailed game it is measured by the rounds that happened to land, and one round can double it, which is not a flaw in the measurement but the definition of what is being measured.

What changes as tests are added

Every table and figure above recomputes from the current test set on each load. Two things are expected to hold as the lab grows: the three hit-rate means should stay within a point of each other, because nothing in the labelling relates to frequency, and the high column of the spread chart should keep its long upper tail while its median stays close to the others. If a studio’s labels start to sort the hit rates, that studio is using the word differently from the rest, and that will be worth a paragraph here.