What we measured
We measured the time from a withdrawal request to the funds landing in the bank account, across the rummy platforms we have used. The sample is small, but the sample is honest: every data point is a withdrawal we actually requested, on an account we actually hold, to a bank account in our name.
We measured three methods: UPI, IMPS, and bank transfer (NEFT). The methods are not equally fast, and the methods are not equally common. UPI is the most common; IMPS is the second; bank transfer is the third.
UPI, in numbers
UPI withdrawals on the platforms we have used settle in 5 to 30 minutes during business hours, with a median of about 12 minutes. The fastest withdrawal we have seen was 3 minutes; the slowest was 90 minutes. The 90-minute withdrawal was a single outlier, and the cause was a network issue on the platform's end, not the bank's.
UPI withdrawals outside business hours are slower. A withdrawal requested at 11pm may not settle until the next morning. The platforms we have used are honest about this; the cashier shows an estimated time, and the estimate is conservative.
IMPS, in numbers
IMPS withdrawals on the platforms we have used settle in 30 minutes to 2 hours, with a median of about 45 minutes. IMPS is the second-fastest method, and it is the right method for a withdrawal that needs to settle outside UPI's business hours.
IMPS has a per-transaction limit, which is usually 2 lakh rupees. Withdrawals above the limit are usually routed through NEFT, which is slower.
Bank transfer (NEFT), in numbers
Bank transfer (NEFT) withdrawals on the platforms we have used settle in 2 to 24 hours, with a median of about 6 hours. NEFT is the slowest method, and it is the right method for a withdrawal above the UPI or IMPS limit.
NEFT has batched settlement windows, which is why the median is 6 hours rather than 1 hour. A withdrawal requested just after a batch will not settle until the next batch.
The first withdrawal, in detail
The first withdrawal on a new account is usually slower than subsequent withdrawals, because the platform runs an additional verification check. The check is the penny drop, and the check is real. The first withdrawal can take 24 to 48 hours, even on UPI.
The slowdown is real, and it is the reason we recommend completing KYC and the first deposit well before the first withdrawal is needed.
What slows a withdrawal down
The most common reasons for a slow withdrawal are: KYC is still being reviewed, the withdrawal is the first on the account, the bank is slow on the receiving end, or the platform is running a manual review (triggered by volume, name mismatch, or new device). Each of these has a fix: complete KYC, wait for the first withdrawal, retry the bank, or contact support.
The least common reason is platform-side delay. A platform that delays a withdrawal without a documented reason is a platform to leave.
What to do when a withdrawal is slow
The first thing to do is to read the cashier's estimated time. The estimate is conservative, and the withdrawal may settle inside the estimate. The second thing to do is to check the bank account directly; some banks show a "pending credit" entry before the credit clears. The third thing to do is to contact support with the transaction ID ready.
A useful rule: do not panic for 24 hours. A withdrawal that is slow inside 24 hours is a withdrawal that is in the platform's queue. A withdrawal that is slow past 48 hours is a withdrawal that needs support intervention.
The state machine, in detail
A normal cashier session has six states: logged in (KYC pending), KYC complete (first deposit), requesting withdrawal, withdrawal pending review, withdrawal rejected, bonus balance locked. Each state has a clear next step, and each state has a clear time-to-next-state.
The state machine is the same across the platforms we have used, with small variations in the time-to-next-state. A reader who understands the state machine is a reader who is not surprised by a slow withdrawal.
What we do not measure
We do not measure the withdrawal limits (the maximum amount per transaction or per day). The limits are platform-specific, and the limits change. We do not measure the bonus balance withdrawal restrictions, because the bonus balance is not a withdrawal method. We do not measure the platform's customer support response time, because the support response time is not the withdrawal time.
The shelf, in three lines
The shelf is the same three measures for every platform: lobby and table finding, cashier and verification, and responsible-play controls. The withdrawal timeline is part of the cashier and verification measure, and the numbers above are the honest averages across the platforms we have used.