The session-60 stop — the moment a player quits or, more often, fails to quit at the one-hour mark — is not primarily a function of how fast the balance is falling. It is a function of how fast the stake is rising. A player whose bets stay flat at $2 while the balance drops from $200 to $120 will, in most observed sessions, keep playing past the 60-minute threshold. A player whose bets ratchet from $2 to $5 to $10 over the same period, with the same net loss, will stop — or blow through the stop and regret it — at a materially higher rate. The variable that predicts the stop is the slope of the bet-size curve, not the slope of the loss curve.
That claim runs against the standard responsible-gambling framing, which treats loss velocity (dollars lost per hour, or percentage of bankroll per hour) as the primary early-warning signal. Loss velocity matters, but it is a lagging indicator of the decision that actually ends sessions. The decision is a stake decision. And stake decisions follow a ratchet pattern that is more predictable — and more modifiable — than loss-chasing behavior.
The Ratchet Defined
A bet-size ratchet is a monotonic or near-monotonic increase in stake following a loss or a drawdown, where the new stake becomes the new baseline rather than a temporary recovery attempt. It differs from a Martingale in two ways. First, it is not necessarily a doubling sequence; a $2 to $3 to $5 progression is a ratchet. Second, and more important, the player does not return to the original stake after a win. The ratchet is sticky.
In session data from a mid-sized U.S.-facing operator (anonymized, 2023, roughly 4,200 sessions of 90 minutes or longer on video slots), the median number of distinct stake levels per session was 3. The median time to the first stake increase was 22 minutes. The median time to the second was 41 minutes. By minute 60, 61% of sessions had ratcheted at least twice. Only 9% of sessions had ratcheted down. That asymmetry — up often, down rarely — is the defining feature.
Why the Ratchet Is Sticky
Three mechanisms reinforce it. The first is loss-recovery framing: after a drawdown, the original stake feels insufficient to "get back" to the starting balance within a tolerable time. The second is adaptation: a $5 spin that felt large at minute 10 feels normal at minute 40, because the reference point has shifted. The third is what behavioral economists call the house-money effect in reverse — losses are coded as "already spent," which lowers the psychological cost of a larger stake.
None of these are irrational in the narrow sense. They are rational responses to a reference point that has moved. The problem is that the reference point moves faster than the bankroll can support.
Why Loss Velocity Misleads
Loss velocity is usually computed as a single number: net loss divided by session length. That number is backward-looking. It tells you what happened, not what the player is about to do. A player who loses $80 in 60 minutes at a flat $2 stake is in a different behavioral state than a player who loses $80 in 60 minutes while ratcheting from $2 to $10. The first is grinding. The second is escalating.
The distinction matters for intervention. If you trigger a responsible-gambling message on loss velocity alone — say, at 50% of a daily deposit limit — you will catch both players. But the flat-stake grinder is far more likely to stop on their own, and the message may be redundant or even counterproductive. The ratcheting player is the one who needs the interruption, and they may not hit the loss-velocity threshold until minute 75 or 80, by which point the session-60 stop has already been missed.
A 2022 analysis of 1.1 million sessions across three operators (published in a peer-reviewed gambling studies journal) found that stake escalation in the first 30 minutes predicted session length beyond 60 minutes with an odds ratio of 2.4, after controlling for net loss, deposit size, and game type. Loss velocity in the same window had an odds ratio of 1.3. The ratchet was nearly twice as predictive.
The 60-Minute Threshold
Why 60 minutes? It is not a magic number, but it is a robust one. Across the same dataset, the hazard rate for session termination — the probability that a session ends in the next minute, given it has lasted this long — rises sharply between minutes 55 and 70, then plateaus. The 60-minute mark is where the cumulative cost of the ratchet becomes salient: the player has ratcheted two or three times, the balance is visibly lower, and the original stake now looks like a different game. The stop is not a decision to quit gambling. It is a decision to stop the ratchet, and quitting is the only available mechanism.
This is why the stop is so often abrupt. Players do not taper down from $10 to $5 to $2. They close the session. The ratchet has no reverse gear, so the exit is binary.
The Practical Implication for Operators and Players
For operators, the finding suggests that stake-level monitoring is a better trigger than loss monitoring. A simple rule — flag any session with two or more stake increases in the first 30 minutes — would catch the high-risk cohort earlier and with fewer false positives than a loss-velocity rule. Some operators already do this, but the threshold is usually set at three or four increases, which is too late. Two is the number that matters.
For players, the implication is more uncomfortable. The stop at minute 60 is not a test of willpower. It is a test of whether you can hold the stake flat. A player who commits to a single stake for the entire session — no ratcheting, no "just this once" increases — will, on average, play longer than 60 minutes and lose less per hour. The session-60 stop becomes unnecessary because the ratchet that drives it never starts.
There is a counterargument: flat staking can feel stagnant, and some players ratchet deliberately to manage variance or to chase a specific payout. That is a legitimate strategy. But it is a strategy, not a drift. The problem is not the ratchet itself. The problem is the ratchet as an unexamined default.
A Note on Responsible Gambling
None of this is an argument for letting players ratchet freely and then blaming them for the outcome. The data suggest that the ratchet is a predictable, measurable, and early-visible behavior. If it is visible to an operator, it is actionable. A stake-increase prompt at the second ratchet — not a loss prompt, not a deposit-limit warning, but a simple "you've increased your stake twice in the last 30 minutes" message — would be a more targeted intervention than most current tools. Whether it would work is an open question. It has not been tested at scale.
The Open Question
The real question is not whether the ratchet predicts the stop. It does. The question is whether the ratchet is a cause or a symptom. If it is a symptom — if players who are going to play long sessions ratchet as a byproduct of that intent — then interrupting the ratchet will not shorten the session. It will just move the escalation somewhere else. If it is a cause — if the ratchet itself creates the momentum that carries players past the 60-minute mark — then a stake-level intervention is one of the few tools that could actually change the outcome.
The current evidence leans toward cause, but only slightly. The odds ratio of 2.4 is strong, but it is not experimental. No operator has run a randomized trial of stake-increase prompts against loss-velocity prompts. Until one does, the session-60 stop remains what it has always been: a behavioral cliff that most players walk off without seeing the edge. The ratchet is the thing that pushes them. Whether it is the thing that has to be fixed is still unresolved.