Sort-code deposits fail roughly 2.3 times as often as card deposits between 01:00 and 02:00, according to a year of transaction data pooled from three UK-licensed operators. Card attempts at that hour fail about 4.1% of the time — already worse than the 2.7% daytime baseline — while bank transfer attempts via sort code and account number fail at 9.4%. The gap is not uniform across the night: it narrows to roughly 1.6x by 05:00 and disappears entirely between 10:00 and 16:00, when both rails sit near 2.4%.

That pattern tells you something more useful than "bank transfers are unreliable." It tells you the failure is structural, tied to who is awake, which systems are staffed, and how the two rails handle a transaction that needs a human or a second system to say yes.

Why the two rails fail differently

A card deposit is largely a synchronous decision. The acquirer sends an authorisation request, the issuer's fraud and balance systems respond, and a result comes back in a few hundred milliseconds. If the issuer declines, you get a decline code and the player can be told immediately. Most failure modes are terminal and fast.

A sort-code deposit is closer to a bank transfer with a confirmation layer. Depending on the provider, it may route through open banking APIs, a pay-by-bank scheme, or a faster-payments-style flow with a matching step. The player leaves the casino, authenticates with their bank, and the operator waits for a webhook or a poll to confirm. That introduces several windows where the transaction can stall rather than fail cleanly: the bank's authentication step times out, the callback never arrives, the session expires, or the deposit lands but the operator's ledger doesn't reconcile it before the player gives up and tries again.

At 01:00, every one of those windows is more likely to be a problem.

The night-shift problem nobody prices in

Three factors stack up between midnight and 04:00.

Bank-side risk scoring runs hotter. Several UK banks apply stricter step-up authentication to outbound transfers during unsociable hours, on the reasonable logic that a 01:30 transfer from a new payee is more likely to be authorised under duress or by someone who has just been talked through it. That means more players hit an authentication challenge, and a meaningful share abandon at that point rather than complete it. In the pooled data, 38% of failed sort-code attempts at 01:00–02:00 were abandoned at the bank authentication screen — not declined, just left.

Fraud and AML queues are thin. Operators batch manual reviews overnight. A deposit flagged for review at 02:00 may not be touched until 07:00 or 08:00. From the player's perspective that is a failure, even though the money may eventually arrive. Card deposits that trigger a review tend to be blocked outright and the player is told, which is a worse experience in the moment but a cleaner one.

Player behaviour is worse. Late-night depositors retry faster and with less patience. The median time between a failed sort-code attempt and a second attempt in the sample was 47 seconds, versus 2 minutes 10 seconds after a card decline. Rapid retries are more likely to trip velocity rules, which then produce a second failure that looks like a rail problem but is actually a risk-rule problem.

The 3DS comparison is misleading

It's tempting to blame strong customer authentication for the card figure. It shouldn't carry all of it. Card failure at 01:00 sits at 4.1% against 2.7% in the day — a 1.4 percentage point rise. Sort-code failure rises from 2.4% to 9.4%, a 7 point rise. SCA is a constant across the day; it doesn't explain why the bank transfer rail degrades three times as much overnight. What changes overnight is staffing, bank-side risk appetite, and the completion rate of a multi-step journey the player has to finish themselves.

What operators actually do about it

The obvious response — hide sort-code deposits at night — is worse than it looks. Bank transfers are the cheaper rail for most operators, and for a slice of players they are the only working option, because their card is blocked, expired, or issued by a bank that declines gambling merchants outright. Removing the option at 01:00 pushes those players to a rail that will fail them anyway, or off the site entirely.

The responses that show up in the data as effective are less dramatic:

• Show the expected completion time before the player starts. Operators that surfaced "this usually takes under 2 minutes" on the bank transfer screen saw abandonment at the authentication step drop by around 11 percentage points in A/B tests. The player who knows the flow is normal is less likely to bail when it takes 40 seconds.
• Stop treating a pending transfer as a failure in the UI. If the deposit is in flight, say so, and suppress the retry button for 90 seconds. Most duplicate-attempt velocity flags in the sample came from players who had no idea their first attempt was still live.
• Route around the queue, not the rail. Flagging a transfer for manual review at 02:00 and clearing it at 08:00 is a policy choice. A small number of operators run a reduced-value auto-approval band overnight — say, up to £150 for accounts with a clean six-month history — which cuts overnight sort-code failures by roughly a third without a measurable rise in chargebacks.

That last point is the one worth arguing about. Auto-approving more overnight transfers means accepting more fraud risk to reduce friction. Whether that trade is worth it depends on your chargeback rate and your average deposit value, and the two operators in the sample that tried it had very different results — one saw fraud losses rise by 0.04% of deposit volume, the other by 0.19%.

The number that should worry you

The 2.3x figure is an average across three operators, and the spread matters more than the mean. The worst-performing operator in the sample had a 12.6% sort-code failure rate at 01:00; the best had 5.8%. Same rail, same hour, same regulator. That's a 6.8 point gap that has nothing to do with the banks and everything to do with how the deposit journey is built and staffed.

If your overnight sort-code failure rate is closer to 12% than 6%, you don't have a payments problem. You have a conversion problem that happens to be wearing a payments costume, and it's costing you the players who are least able to use anything else.

The open question is whether the gap is actually widening. Two of the three operators saw overnight sort-code failures rise year on year, while card failures stayed flat. If that trend holds, the 2.3x becomes 3x, and the case for keeping bank transfers visible at 01:00 — with all the friction that implies — gets harder to make. The alternative is telling a player at 01:00 that their only working deposit method is unavailable until morning, which is a decision most operators have so far been unwilling to make out loud.