Cutting the clock-off allowance in online poker from 30 seconds to 15 seconds reduced the share of folds made after a player had already burned roughly 20 seconds of tank time by 12%, according to a session-level analysis of 140 tables. The finding, drawn from hand histories covering a single 24-hour window, suggests that a large portion of what looks like genuine deliberation is really a function of how much rope the client hands out, not how hard the decision is. If it holds up, time-bank policy is a bigger lever on pace of play than most room operators treat it as.
What the Session-140 Dataset Actually Contains
The label "Session-140" refers to a hand-history pull spanning 140 concurrent six-max No-Limit Hold'em tables at $0.50/$1.00, captured over one calendar day in the spring of 2025. That produced 186,412 hands, of which 41,038 reached a flop and 9,772 went to a turn with at least one player still holding a live time bank. The unit of measurement here is the tank fold: a fold made after the acting player had consumed 20 or more seconds of their allotted decision window, including time-bank seconds.
The baseline regime allotted 30 seconds of clock plus a 30-second time bank that refilled at a fixed rate. The treatment regime, applied to a matched set of tables in the same pool, allotted 15 seconds of clock plus the same 30-second bank. Everything else — blind structure, rake, HUD availability, the anonymization policy — was held constant. That matters, because time-bank research is unusually vulnerable to confounding: rooms that shorten clocks also tend to be the rooms tightening anti-collusion rules, and the two effects are hard to separate after the fact.
Why the 20-Second Threshold
The 20-second cutoff is a judgment call, but a defensible one. In the 30-second baseline, the median tank fold consumed 24.1 seconds, and the distribution had a visible shelf between 21 and 27 seconds — players who had clearly decided to fold but were running out the clock, either to disguise timing tells or simply because the client permitted it. Under the 15-second clock, that shelf largely disappears; the median tank fold falls to 16.8 seconds, and the 21-to-27 shelf thins out to a fraction of its former mass.
The 12% Figure and What It Is Measuring
The headline number is a 12% relative reduction in the tank-fold rate among hands that reached the turn. In absolute terms, tank folds fell from 14.9% of turn decisions to 13.1%. That is a modest absolute move and a meaningful relative one, and the gap between those two framings is where most of the argument lives.
Two caveats deserve airtime. First, the reduction is not uniform across player types. Players with a VPIP above 40% — the loose-passive cohort — showed almost no change, a 2.1% relative decline that sits inside the confidence interval. The effect is concentrated among regulars with VPIP between 22% and 28%, who cut their tank folds by 19%. That is consistent with a story about performed deliberation: strong players tank partly to manage their own timing-tell profile, and a shorter clock removes the runway for the performance.
Second, the fold rate overall did not fall. Players did not fold less; they folded faster. Total folds as a share of turn decisions moved by less than half a percentage point. The time-bank change redistributed when folds happened, not whether they did.
The Pace Effect Is Larger Than the Fold Effect
The more striking number in the dataset is not the 12% at all. Average decision time on the turn dropped from 9.4 seconds to 6.1 seconds, a 35% reduction, and hands per hour rose from 71 to 84 across the matched tables. That throughput gain is where the operator's interest lies, and it is roughly three times the size of the headline fold effect. Any room weighing a clock change should be honest that the revenue case rests on hands per hour, not on tank-fold reduction, which is mostly a player-experience argument.
Where the Design Gets Contested
The Refill Rate Problem
The Session-140 comparison held the time bank at 30 seconds with a fixed refill. That is not how most rooms actually run banks. Many tie refill to hands played or to time elapsed, and a 15-second clock paired with a fast refill can produce more total tanking than a 30-second clock with a slow one. The 12% figure is therefore conditional on the refill policy used in the trial, and it should not be generalized to rooms running hand-based refills without a fresh measurement.
Selection Into the Treatment
Players were not randomly assigned to clock regimes; they sat where seats were open. The matched-table design controls for stake and format but cannot fully control for the possibility that faster-format tables attracted a different regular population during the 24-hour window. The 19% effect among mid-VPIP regulars is the most fragile estimate in the set, precisely because that cohort is the most likely to sort itself across table types. A 24-hour window is also short enough that novelty effects — players behaving differently because the clock is new — cannot be ruled out. A four-week follow-up would be the obvious next step, and as of this writing it has not been published.
Practical Reading for Rooms and for Players
For operators, the defensible takeaway is narrower than the headline. A 15-second clock with a matched bank reduced post-20-second folds by 12% in one pool over one day, and the effect was concentrated among regulars rather than recreational players. If the goal is protecting recreational players from being tanked at — a common justification for clock reform — the data offer only partial support, since the loose-passive cohort barely moved. If the goal is throughput, the 35% decision-time reduction is the number to cite.
For players, the implication runs the other way. Timing tells are cheaper to fake when the clock is long, and a shorter clock compresses the information content of a tank. Regulars who built timing-based reads under a 30-second regime should expect those reads to decay, and the 19% drop in regular tank folds suggests many already have. The practical adjustment is to weight bet-sizing and line composition more heavily and timing less, which is unglamorous advice but follows directly from the numbers.
Responsible-gambling notes apply here as anywhere: faster hands per hour means faster bankroll churn, and a room that raises throughput from 71 to 84 hands per hour has effectively raised the hourly cost of play by roughly 18% at constant stakes and win rates. Players on fixed session budgets should recalibrate loss limits when a room shortens its clock, because the same two-hour session now buys more decisions and more rake.
The Question the Data Cannot Answer
The open question is not whether 15 seconds trims tank folds — it does, by 12% in this sample — but whether the trim survives contact with a live ecosystem. Time banks exist partly to protect genuinely tough spots, and the hands most affected by a shorter clock are, by construction, the ones where players took longest. Whether the 12% reduction is concentrated in performed deliberation or in real deliberation is the difference between a clean efficiency gain and a subtle degradation of decision quality that the fold-rate metric cannot see. A study that paired clock length with post-session hand-review accuracy would settle it. Until then, rooms are making a policy bet on a 24-hour snapshot, and players are making a reads bet on a clock that may not be there next quarter.