A £5 minimum buy-in sounds like a rounding error next to the £50 and £100 tables most UK cash-game regulars think of as "real" poker. But on the sites that enforce it — 888poker's UK-facing lobby, several iPoker skins, and a handful of micro-stakes rooms on the MPN before its closure — the £5 floor is the number that decides who is still sitting at 6:55pm and who has already gone to make dinner. The effect is sharpest at 7pm, the point where the after-work session and the pre-dinner exit collide, and it shows up not as a wave of bust-outs but as a quiet thinning of tables that the lobby never reports.

The 7pm cliff is a liquidity event, not a skill event

Cash-game traffic on UK-facing sites follows a predictable curve. Volume climbs from around 4:30pm as the commute ends, peaks somewhere between 8pm and 10pm, and troughs hard in the early hours. The 7pm slot sits on the rising edge, which is why it is misleading. Tables are being created, not destroyed, so the aggregate numbers look healthy.

What the aggregate hides is the churn underneath. At a £5 minimum, the median buy-in on a 5p/10p table is not £5 — it is £5 to £8, because most players top up to the minimum or one step above and then stop. That gives a short-stacked player roughly 50 to 80 big blinds, which at 5p/10p is 50 to 80BB, and that is a workable but fragile stack. Two standard coolers and it is gone. At a £20 minimum, the same player has 200BB and can absorb the same two coolers without leaving.

The consequence at 7pm is that the £5 cohort — which by most room-side estimates is 30% to 40% of the player pool at the lowest stakes — exits on a timer set by variance rather than by choice. A player who sits down at 6:15pm with £5 has, on average, between 45 minutes and 90 minutes of play before the stack is either doubled or gone. 7pm is when the first wave of that coin-flip resolves.

Why the exit is asymmetric

Losses at a £5 minimum are capped at £5. Wins are not. A player who runs £5 up to £25 at 5p/10p has a decision: keep playing with a now-respectable stack, or bank the £20 profit and leave. In practice, the majority leave. The £5 minimum does not just limit downside; it manufactures a class of player whose session length is determined by whether they hit their stop-win before their stop-loss, and at micro-stakes the stop-win is hit more often than the raw math suggests, because the field is soft enough that a competent player doubles up roughly one session in three.

That asymmetry is why 7pm thinning is worse on £5-minimum tables than on £20-minimum tables at the same blind level. The £20 player has no reason to treat a £40 stack as a life-changing result. The £5 player does.

What the lobby data does and does not show

Poker room lobbies report tables and players, not session length. A table that shows "6/6" at 6:50pm and "4/6" at 7:10pm looks like two players left. What actually happened is that four players left and two new ones sat down — a 66% turnover in twenty minutes, invisible in the snapshot.

This matters because the standard argument in favour of low minimums is that they lower the barrier to entry. True, but the barrier argument stops at the deposit screen. It says nothing about retention. A £5 minimum recruits players efficiently and then loses them at a rate that a £20 minimum does not, and the loss is concentrated in the 6:45pm to 7:30pm window because that is when the first short-stack resolution lands.

There is a second-order effect. Regulars who multi-table avoid £5-minimum tables at exactly this hour, because table-breaking mid-session is a real cost — you lose your seat, your reads, and your position in the queue. A regular with a 6-table setup will drop a £5 table the moment it drops below four players and pick up a £20 table instead. So the £5 pool loses not just the short-stacked recreational players but the regs who would otherwise backfill the seats. By 7:30pm, a £5-minimum 5p/10p table is often down to two or three players, all of whom are waiting for the same thing.

The rake arithmetic makes it worse

UK-facing cash games typically rake 5% capped at £1 to £1.50 at micro-stakes, with some rooms taking a flat 2p or 3p per pot regardless of size. At 5p/10p, a 5% uncapped rake is brutal — a £2 pot pays 10p, which is a full big blind. Rooms that use a flat per-pot rake are gentler on small pots but harsher on the smallest ones.

The £5 minimum interacts with rake in a way that is easy to miss. A player with a £5 stack who wins a £3 pot pays 15p in rake at 5%. That is 3% of their entire bankroll on the table, gone in one hand. Over a 60-hand session, even a break-even player is paying the equivalent of their full buy-in in rake across multiple pots. The £5 stack is not just short; it is short and being taxed at a rate that a £50 stack barely notices.

This is the strongest argument that the £5 minimum is not a neutral accessibility feature but a structural tax on the least experienced players. It is also why some rooms have quietly moved to £10 minimums at 5p/10p and £20 at 10p/25p — not to exclude players, but because the rake math at £5 produces a churn rate that costs more in liquidity than the extra £5 of buy-in costs in sign-ups.

A note on the 2023 Gambling Act review

The white paper published in April 2023 and the subsequent consultations have focused heavily on deposit limits, financial risk checks, and slot stake caps. Cash-game minimums have not been a headline item. But the same logic that produced the £2 slot stake cap — that small-stake products can produce disproportionate harm through volume — applies awkwardly to a £5 poker minimum. The difference is that poker harm is not usually a function of stake size but of session length and loss-chasing, and a £5 minimum arguably encourages exactly the short, repeated, high-churn session that the safer-gambling framework is trying to discourage.

What rooms could measure, and mostly do not

The data to settle this exists. Any room can pull, for a given stake and minimum, the distribution of session lengths, the hour-by-hour table count, and the ratio of new seats to returning seats. Most do not publish it, and the affiliate-facing numbers they do publish are deposits and first-time sign-ups, which flatter low minimums.

A room that wanted to test the claim would compare two 5p/10p tables, one at a £5 minimum and one at £20, over a fortnight, and look specifically at the 6:45pm to 7:30pm window. The prediction is straightforward: the £5 table shows higher seat turnover, lower average session length, and a larger drop in player count across that window, with the difference concentrated among players who bought in for exactly the minimum.

Whether any UK-facing room has run that comparison and chosen to keep the £5 minimum anyway is the open question. The barrier-to-entry argument is easy to make in a marketing deck. The retention argument requires admitting that the cheapest seat on the site may be the one that empties fastest — and that the players it empties out are the ones least able to absorb the rake while it happens.