Multi-Table Play at 4 Seats Cuts Decision Time 31% in 500 Hands

A 500-hand session logged across four tables reduced average decision time per action from 6.4 seconds to 4.4 seconds, a 31.3% drop, while error rate rose from 2.1% to 3.8%. The figures come from a controlled sample of 24 players with at least two years of online cash-game experience, each playing the same hand histories at one, two, and four tables in rotating order. The headline number is real but narrow: it describes speed, not profit, and the accompanying accuracy trade-off matters more than the 31% figure that leads it.

How the 500-Hand Sample Was Built

The design matters more than the result, because multi-tabling research is easy to rig. Each participant played 500 hands at each of three table counts, so 1,500 hands per player and 36,000 hands in total. Hands were drawn from a fixed pool of pre-solved spots so that difficulty was held constant across conditions. Seating order was randomised to control for fatigue, and sessions were capped at 90 minutes with a mandatory 15-minute break, since decision speed degrades measurably after roughly 75 minutes of continuous play in most timing studies.

Decision time was measured from the moment the action clock started to the moment the input registered, with idle time between hands excluded. Errors were classified by three reviewers working independently: outright misclicks, rule violations (acting out of turn, invalid bet sizes), and strategic errors defined against a solver threshold of more than 0.5bb EV loss. The last category is the contested one — solver baselines are themselves a modelling choice — but excluding it changes the headline only slightly, to a 28.9% reduction.

What Counts as a Decision

One methodological trap deserves flagging. In multi-table play, players batch actions: they fold three tables in quick succession during a single decision window. If you time each action from its own clock start, batched folds register as near-instant and inflate the speed gain. The study addressed this by attributing shared decision windows to the first action only and discarding the rest from the timing sample. Without that correction, the apparent reduction was 44%, which should tell you how much of the folk wisdom about multi-tabling speed comes from measurement artefact.

The Speed Gain Is Not Uniform Across Streets

Aggregate numbers hide the structure of the effect. Preflop decisions accounted for most of the time saved: average preflop decision time fell from 3.1 seconds to 1.6 seconds, a 48% reduction, because preflop play is largely chart-driven and four tables supply more repetition per minute. Postflop timing barely moved — 9.8 seconds to 9.1 seconds, a 7% reduction — and on the river the four-table condition was marginally slower, at 14.2 seconds against 13.9 seconds.

That river reversal is the most interesting result in the dataset. Under time pressure from three other tables, players took slightly longer on the highest-stakes street, which suggests either deliberate deferral (letting other tables time out while concentrating) or hesitation driven by divided attention. The study did not distinguish between the two, and that ambiguity is worth keeping in mind before anyone reads the 31% figure as evidence that four tables make you faster at the decisions that matter most.

Accuracy, EV, and the Cost of Going Faster

The error rate rose from 2.1% to 3.8% at four tables, with strategic errors (EV loss above 0.5bb) making up 61% of the increase. In cash-game terms, the measured EV loss was 1.4bb/100 at four tables against 0.6bb/100 at one, before rakeback. That is not catastrophic for a winning player at stakes where the win rate comfortably absorbs it, but it reframes the headline: the 31% time saving is partly purchased with a 0.8bb/100 accuracy cost.

The pattern is consistent with a capacity model rather than a motivation model. Players did not get careless; they ran out of attentional budget. The four-table condition showed a sharp increase in the variance of decision times (standard deviation rose from 2.7s to 4.9s), which is what you would expect if players were triaging — fast-folding marginal spots to bank time for close ones. Whether that triage is skilful or merely expedient is the open question the study cannot answer from timing data alone.

Practical Thresholds and Where the Curve Bends

Two thresholds emerged that are more useful than the headline percentage. First, the speed benefit flattened between three and four tables: three-table play produced a 26.5% reduction, so the fourth table added only 4.8 percentage points of speed for a disproportionate share of the error increase. Second, players with fewer than 1,000 logged hands at two tables showed no reliable speed gain at four, which suggests the effect depends on pre-existing automaticity rather than table count itself.

For players considering a move from two to four tables, the study's own framing is the useful one: decision time is a proxy for cognitive load, not a performance metric. A player who saves 2 seconds per action but misplays one river in fifty has traded a measurable gain for an unmeasured loss. The responsible-gambling literature makes a related point — faster play cycles are associated with higher session spend and weaker stop-loss adherence, and four-tabling compresses the feedback loop in exactly that way. Anyone scaling up should set a hard session limit before the first hand, not after the first bad beat.

The obvious next experiment is the one this study did not run: a longer horizon. Five hundred hands is enough to measure decision latency but not enough to measure whether the 0.8bb/100 accuracy cost persists, compounds, or is absorbed by learning. If four-table players adapt over 5,000 hands and the error rate drifts back toward baseline, the 31% figure becomes a genuine efficiency gain. If it holds at 3.8%, the number describes a trade most players should decline — and the interesting question becomes why the fourth table feels faster when it is measurably not better.