Slot Coin-In Tracks Floor Music Key, Not Tempo, Study Finds
A slot machine's coin-in moves with the key of the background music, not its tempo, according to a study set to appear in the March 2026 issue of the Journal of Gambling Studies. Researchers tracking 412 players across three Las Vegas locals casinos over 14 months found that shifting a cabinet's soundtrack from D major to B♭ major raised average coin-in per session by 9.4% — while speeding the same track from 92 to 128 beats per minute changed it by less than 1%. The finding cuts against two decades of slot-floor convention, in which tempo has been treated as the primary audio lever on player pace.
What the study actually measured
The research team, led by a behavioral economist at the University of Nevada, Las Vegas, worked with a regional operator that agreed to let them modify the audio firmware on 60 identical video slots — same cabinet, same math model, same 94.1% RTP configuration, same $1.50 minimum bet. The machines sat in three properties: two in the Las Vegas valley, one in a tribal casino outside Phoenix. None of the properties are named in the paper; the operator's agreement required anonymity in exchange for access to machine-level accounting data.
That data is the study's real strength. Rather than relying on self-report or lab play, the authors pulled actual coin-in, session length, and spin count from the operator's slot accounting system, matched to the exact minute each audio condition was active. Over 14 months they logged 1.83 million sessions and roughly 214 million spins.
The design was a rotating crossover. Each machine cycled through four audio conditions on a fixed schedule:
- Control: the cabinet's original soundtrack, 92 BPM, D major
- Tempo shift: same melody, 128 BPM, D major
- Key shift: same melody, 92 BPM, B♭ major
- Both: 128 BPM, B♭ major
Conditions rotated every 72 hours. The rotation schedule wasn't random — the operator wanted to avoid mid-weekend switches that might confuse floor staff — but the authors controlled for day of week, time of day, property, and player card status in the regression. They also excluded the first 20 minutes after every switch to wash out carryover.
The headline numbers: key shift alone produced a 9.4% increase in coin-in per session (95% CI: 6.1% to 12.7%). Tempo shift alone produced a 0.8% increase, which the authors describe as "not statistically distinguishable from zero at conventional thresholds." The combined condition produced a 10.2% increase — barely more than key alone, suggesting the two variables aren't additive and that tempo is doing almost none of the work.
Session length moved in the same direction but less sharply. Key shift added about 3.1 minutes to the median session; tempo added 0.4 minutes. Spin rate per minute was essentially flat across all four conditions, which is the detail that most complicates the obvious explanation.
Why key and not tempo
If players aren't spinning faster, the coin-in increase has to come from somewhere else. The data points to three channels, none of them fully separable.
First, bet size. Average bet per spin rose 6.8% under the key-shift condition, from $2.41 to $2.57. The authors can't say why from the accounting data alone, but their working hypothesis is arousal: a key change alters the perceived brightness of the same melody without changing its pace, and brighter perceived sound has been linked in earlier lab work to higher risk tolerance in simple wagering tasks. A 2023 study in Psychology of Music found that major-key passages raised participants' willingness to accept a 60/40 coin flip by roughly 11 percentage points compared to minor-key passages at identical tempo.
Second, session continuation. Players in the key-shift condition were 7.2% less likely to end a session in any given five-minute window. That's a hazard-rate effect, and it compounds. A player who stays 3% longer per decision point ends up staying far longer than 3% by the end of a session. This is the mechanism most likely to matter to operators, because it's the one that scales.
Third, and least comfortably, loss-chasing. The researchers flagged players who had dropped below 50% of their session-starting balance and continued playing. Under key shift, that group's continued coin-in was 14.9% higher than control. The authors are careful here — they note that "loss-chasing" is a contested construct and that their proxy is crude — but the effect size is large enough that they flag it explicitly in the discussion as a harm-relevant finding.
The tempo result is the more surprising one, and it deserves context. The idea that faster music speeds up play is old and mostly comes from a 1982 study of supermarket shoppers, not gamblers. Slot-floor audio design inherited it anyway. In this dataset, tempo did almost nothing. One plausible reason: at 128 BPM, the loop is fast enough that players stop hearing it as music and start hearing it as texture — a wash of sound rather than a melody with a shape. The key change, by contrast, alters the character of the loop while leaving its structure intact, so it stays perceptible.
There's a second possibility the authors raise and don't resolve: the original soundtrack was already in D major, a bright key. Moving to B♭ major is a shift of a minor third down, which in absolute pitch terms is darker, not brighter. Yet coin-in rose. That contradicts the simple brightness story and suggests the effect may be about change rather than direction — a fresh key breaking auditory habituation. The study can't distinguish these, because it only tested one key move.
What the numbers don't settle
Three limitations are worth stating plainly, because the study's framing invites overreading.
The operator is one operator. Three properties, one math model, one cabinet type. Video slots with 94.1% RTP and a $1.50 floor are a specific product. Nothing here tells you whether the same key effect shows up on a $0.01 denom penny machine with a 88.5% RTP, or on a mechanical reel stepper, or in a sportsbook, or on an online slot where the player controls the volume slider and probably has it at 20%.
The effect is measured in coin-in, not revenue. Coin-in is handle — total wagered — not house win. A 9.4% coin-in lift on a 94.1% RTP game translates to a much smaller change in hold, and the paper doesn't report net win by condition. If the key shift also changed bet patterns in a way that shifted players toward higher-variance choices, the revenue effect could be larger or smaller than the handle effect implies. The authors say they're preparing a follow-up on this; it isn't in the current paper.
Habituation is unmeasured over the long run. Fourteen months sounds long, but each individual player in the sample encountered the rotating conditions, not a fixed one. The study measures the effect of a change in key within a player's experience. It does not measure what happens if a cabinet plays B♭ major for six months straight. Based on the habituation hypothesis, the effect would decay — possibly to nothing. That's the single most important open question for anyone thinking about applying this.
There's also the question of whether the effect survives when players can hear it. Slot floors are loud, and the study's machines used cabinet speakers at the operator's standard volume, which the paper describes as "moderate." Headphone use was not tracked. If the effect depends on the loop being perceptible as melody, then a noisy floor is quietly working against it.
The regulatory angle nobody asked for
Here's where this gets awkward for the industry. Nevada Gaming Control Board regulations already govern slot machine sound, but the relevant provisions are about volume and about not simulating the sound of coins or jackpots in misleading ways. Regulation 14 doesn't address key, mode, or melodic content, and there's no mechanism for it to. A key change is not a payout, not a display, and not a misleading signal in any conventional sense.
But the study's loss-chasing finding raises a question regulators have been circling for years without a clean answer: if a design feature measurably increases continued play among players who are already down, is that a feature or a harm? The UK Gambling Commission has moved toward treating "dark patterns" in online products as a regulatory category. US regulators have mostly stayed out of audio and interface design, with the partial exception of some tribal gaming commissions that have their own internal review processes.
The practical problem is that "B♭ major" is not a thing a regulator can easily write a rule about. You could regulate tempo, because it's a number. You could regulate volume, because it's a number. Key is a categorical property with 24 common options and no obvious bright line. Any rule would either be arbitrary or would amount to "don't change the music," which is not a rule anyone is going to write.
Which may be why the more likely outcome is voluntary. Slot audio is outsourced to a small number of sound design shops, and those shops take briefs from manufacturers. If key selection becomes a known performance lever, it becomes part of the brief. The question is whether it becomes part of the brief as "optimize for key" or as "optimize for key, and here's the harm data."
What to watch next
The follow-up on net win will matter more than this paper does. If the coin-in lift doesn't translate to hold, operators will shrug. If it does, the audio briefs change within a product cycle or two.
The habituation study matters more still, and it's the one nobody has run. Everything in this finding depends on the effect being about change rather than about any particular key. If a cabinet can only deliver the lift by rotating keys, then the design implication is a rotating soundtrack — which is a strange thing to build into a slot machine and a stranger thing to explain to a regulator.
And then there's the question the study can't answer because it didn't ask: did any of the 412 players notice? A 9.4% coin-in difference is large enough to show up in accounting data and small enough that no individual player would feel it. The whole finding lives below the threshold of conscious attention, which is exactly where slot design has always done its most effective work — and exactly where the case for disclosure gets hardest to make.