Slot Volatility Simulator

Play out a slot session at any RTP and volatility, and see how often a bankroll actually doubles or busts.

This prices the spread of outcomes for a stylized game at the RTP and volatility you set, and it is not a model of any particular machine you can go and play.

Across 1,000 sessions

Simulating 1,000 sessions…

Where the 1,000 sessions ended

$0start2x or more

below the startat or above it

Same 96% RTP, four different rides

All four rows are priced at the same $1 bet and 96% RTP so only volatility moves. They are this simulator’s own four settings, a stylised model, not four named real slot machines.

Hits	Average win	Top win	RTP	Edge
38%	$2.53	$17.461 in 263	96%	4%
25%	$3.84	$35.561 in 133	96%	4%
15%	$6.40	$91.041 in 222	96%	4%
9%	$10.67	$224.371 in 370	96%	4%

Average win is the mean payout on a spin that pays something, not on every spin. RTP and house edge do not move down the table: all four rows cost the same 4% of what gets staked. Volatility only trades hit frequency for win size: same cost, a different ride, never a safer or cheaper one.

Set the game

RTP %Bet / spinStart balance

Spins in the session300

Set by these settings, before a spin

House edge

4.0%

Top win at this bet

$91.04

How often it lands

1 spin in 222

What 1,000 sessions pin down

The four figures here are read off the sweep above, and arrive with it.

One session, spin by spin

Balance

$69.47

spin 300/300

Biggest win

$91.04

Spins that paid

28 of 300

Each tick under the curve is a spin that paid, taller for a larger win.

Example session: 300 spins at $1.00, high volatility, 96% RTP. Press Run a session to play one with the settings above.

What a full session would cost

Spins300Bet per spin$1.00Staked in full$300

Staked in full$300House edge4.0%Cost$12

The loss follows the money actually staked, not the number of spins you planned. Playing all 300 spins at $1.00 would stake $300 and cost about $12 at 96% RTP. A session that runs out of balance stops sooner and stakes less than that. Volatility moves the busts and the big wins further apart without changing what a spin returns on average.

Loss disguised as a win

A real multi-line machine stakes several lines on one spin and pays each independently. A line can pay while the spin as a whole returns less than it cost, and the machine still shows it as a win: credits added, a line lights up, the same reinforcing display a genuine win gets. Academic literature calls this a loss disguised as a win (LDW). Below, each of the 20 lines pays independently about 1 time in 6, usually for a modest multiple of that one line’s own bet rather than a large one, which is what lets several lines pay on one spin while the spin still costs you money.

Lines playedBet per line

Lines20Bet per line$0.05Cost per spin$1.00

What the machine’s screen shows

20 lines staked and nothing spun yet.

What you actually got back

The spin cost$1.00

It paid backnot spun yet

Spin, and what the spin paid back is set against what it cost.

The tally

Spin to start the tally.

A stylised demonstration, not a schedule for any real machine: the per-line hit rate and payout table are chosen to show the mechanic at a realistic order of magnitude, not transcribed from one game. Dixon, Harrigan, Sandhu, Collins & Fugelsang, “Losses disguised as wins in modern multi-line video slot machines,” Addiction (2010) found that players’ physiological arousal during an LDW spin is statistically indistinguishable from a genuine win and that players systematically overestimate how often they actually won a session. Templeton, Dixon, Harrigan & Fugelsang, “Using an educational animation to reduce erroneous win overestimates,” International Gambling Studies (2017) tested a short educational reveal of exactly this mechanic and measured it reducing that overestimate, which is the design this block borrows.

Written and reviewed by Alexander Korsager · last reviewed September 26, 2026 · corrections

The idea

Two slots can advertise the same 96% return to player and feel completely different to play. RTP only says how much of all money wagered is paid back over millions of spins; volatility says how that return arrives. A low-volatility game feeds back small wins steadily; a high-volatility game takes your balance down for long stretches and occasionally returns everything at once.

This simulator lets you set both.Each spin is drawn from a payout distribution whose average is exactly your chosen RTP, with the shape controlled by the volatility setting. A thousand sessions are played at your settings as soon as you change one, and they are what the panel reports first: what fraction bust before a set number of spins, what fraction double, and where the middle outcome lands. Underneath them you can run a single session and watch one balance curve unfold against that spread.
The lesson is usually the same.Even at a generous 96% RTP, a fixed number of spins at a fixed bet leaves most sessions down, because the house edge is relentless and the big wins that carry the average are rare. Higher volatility widens both tails, so more sessions bust and more sessions hit something large, without changing what a spin returns on average. It does change how long a balance lasts, and a session that runs out of money early stakes less than you planned and so loses less in total.

FAQ

Questions players ask

Does a higher RTP mean I will lose less?

On average, yes, over a very long run. In a single session variance dominates, so a 96% game and a 94% game can easily produce the opposite of what the RTP suggests. The edge only shows reliably over tens of thousands of spins.

What does volatility actually change?

The spread of outcomes, not what a spin returns on average. High volatility means longer losing runs punctuated by rarer, larger wins. Your expected loss per spin is set by RTP and does not move when you change the volatility setting. Your total loss over a session can still come out lower at high volatility, because a balance that busts early stops staking, which is why the simulator reports what was actually staked rather than what a full session would have cost.

Is this a real slot machine?

No. It is a statistical model with the same RTP and volatility characteristics as a real game, built to show the distribution of outcomes. It pays nothing and costs nothing.

Methodology

How this is calculated

Every spin is one draw from a five-outcome table with a sixth outcome of nothing, and the two controls do separate jobs. Volatility sets the hit frequency and the relative sizes of the wins: 38% of spins return something on Low, 25% on Medium, 15% on High, 9% on Extreme.

RTP then rescales every win size so the average return per dollar staked lands exactly on the figure you typed, leaving all six probabilities untouched. That is why moving the volatility control changes how the return arrives and not how much of it does.

The 1,000-session sweep replays the same rules the live session uses, stopping when the balance cannot cover another bet or when the spin cap runs out. It is seeded, so it holds still while you read it: leave the controls alone and the same 1,000 sessions stay on screen. Re-run, above the sweep, deliberately breaks that.

It draws a different 1,000 sessions of the same game on settings you have not touched, and the figures move when you press it, which is the fastest way to see for yourself which of them a thousand sessions actually pins down.

The figure the page does not print, and the one that explains the rest: at the shipped defaults the middle session finishes on about $55, and moving volatility alone to Low takes that to about $87. Same 96% RTP, same 4% edge, same $12 theoretical cost.

Common misreadings4

The top win beside the settings is the largest the pay table can pay at your bet size, worked out before any spin rather than found by the sweep. It sits with the settings for that reason. It is also not the jackpot the size of it suggests: it is already counted inside the RTP the sweep is measuring, not an upside on top of it. The four-way comparison above the controls shows the same thing across all four settings, at the same $1 bet and 96% RTP: the top win gets bigger as volatility rises, never a better one, since every setting shares the same 96% return and 4% edge.
A higher volatility setting showing a smaller average loss is not the higher volatility game being cheaper. Averaged over 200 sweeps at the defaults, Extreme finishes about $8.65 down against Low's $11.97, and the reason is that the money is gone sooner: the average Extreme session stakes about $200 where the average Low session stakes about $300. The edge is 4% of what gets staked either way. What Extreme does to the typical session is take the median end balance from about $87 to about $0.77.
The double-up rate is not the odds of a session going well. At the defaults about 14% of sessions finish on twice the starting balance and about 37% cannot afford a spin before the cap, out of the same 1,000.
Finishing ahead of most of the thousand says the draw went your way, not that the session was played better. Every session in the sweep uses the same bet, the same stopping rule and the same table as the one you just watched, and nothing in this tool responds to how you play it.

Assumptions4

Spins are independent draws from one fixed table. There are no free-spin rounds, no feature that carries state between spins, and no change in the odds after a long losing run.
Changing RTP moves the win sizes and never the probabilities. The hit frequency you get at 99.5% is the same one you get at 80%, which is what keeps the two controls independent and is the claim the whole page rests on.
The bet is the same on every spin, and the session ends the first time the balance cannot cover another one or the spin cap is reached, whichever comes first.
RTP is the long-run return on every dollar staked, not a promise about a session. At 96% the edge is 4% of what you actually stake, and 300 spins at $1 stakes $300 only if the balance survives all 300.

Limits4

The four volatility settings are a stylized model of a category, not a measurement of a real game. What they are built to reproduce is the ordering: a low volatility game hits over a third of spins for small multiples, an extreme one hits under a tenth and carries its return in a tail worth 1,000 times the bet before scaling. Real machines have bet levels, feature buys and multi-stage bonus rounds that none of this represents. Paylines are the one exception, since #667 added a second, separate model of exactly that below the sweep: this simulator's own single-line table cannot pay less than 1.819 times the RTP you set (the smallest reachable multiplier, at Low volatility), which only drops under a full bet back past a 55% RTP nothing here or on a real floor reaches, so it cannot represent a spin that pays something and still returns less than it cost. The block below can, on its own smaller model, because that is the one thing a single-line table structurally cannot show.
One thousand sessions is a sample, so every figure under the sweep carries the range that sample pins it to, printed under the figure itself. The average result is much the loosest of them, and the reason is the shape of the game rather than anything about the code: nearly all of a high volatility game's return arrives in rare large wins, so a sweep that draws a few more of them than usual reports a much smaller loss, while the bust rate only counts sessions and barely moves. Run the shipped defaults 200 times with different seeds and the average lands anywhere between $4.60 and $15.54 down, nine times in ten, around a center near $10.20, while the bust rate stays between 34.9% and 39.9%. That is what the printed ranges are measuring, one sweep at a time, and Re-run is the same experiment by hand: press it a few times without touching a control, and the bust rate barely moves while the average result will not settle.
The theoretical loss in the closing note is the cost of a session played to its full length. Any session that runs out of balance first stakes less than that and so loses less than that, which is arithmetic about a shorter session rather than a better one.
Nothing here is a schedule for a real machine's RTP. A published return figure is the manufacturer's long-run average over millions of spins across every player, and an operator can often configure the same game to several different ones.

Evidence

The four profiles, the scaling step and the sweep are in math.ts, and math.test.ts checks them against a reference it writes out by hand rather than against the module's own output.

The per-spin return is a discrete mixture with a point mass at zero, so its mean, its hit frequency and the scale factor that sets them are all closed forms a reader can check with a calculator, and the test file shows that arithmetic in full.

The relation between the two figures the panel prints side by side is pinned separately: Wald's identity gives the average result as the average amount staked times the house edge, whatever stopped each session, so no volatility setting moves the center. A passing test means the code matches this model.

It does not mean the model matches a machine on a real floor.

Gambling can cause serious harm. The National Problem Gambling Helpline is free, confidential and open 24/7, call or text 1-800-MY-RESET (1-800-697-3738), or chat at ncpgambling.org/chat.

Read up

Go deeper

Related tools