The denial rate,
published whatever it is

Where the counter is today

The alternative was to hold this page back until there was a real number to put on it. The number is here instead, in the same counter format, starting where an honest counter starts: at zero, before launch, with the rule for how it is computed already attached.

Soft prior preview (synthetic)

The priors are deliberately low for the balanced persona on every published ground: the audit’s job here is to show the shape of a denial-rate breakdown — a row per refusal clause, with the same denominators the live ledger will publish — not to forecast a specific number. Persona shifts (conservative / aggressive / grid-farm / abuser) move the priors; the breakdown shape does not.

How we will compute it

The definition matters as much as the number, because a denial rate is trivially manipulated by moving cases out of the denominator. Ours does not exclude requests refused pending correction, requests refused for identity reasons, or requests that were withdrawn after we asked a question. Anything decided is counted.

This expands the commitment already stated on /payouts/#denials rather than duplicating it: the method lives here, the live counts live there, and the two pages point at each other. If a figure on this page ever disagrees with the ledger, the ledger is the source of truth.

No firm in this industry publishes this number

Payout totals are advertised everywhere in this category. The share of payout requests refused is published nowhere, by anyone — and it is the one figure that would tell a buyer the most about what actually happens at the payout gate.

The research behind those statements is in the pass rates and payout rates dossier and the when a payout can be denied article in the research library. Both cite each firm’s own published terms; nothing there is a view about a strategy.

What we will count as a refusal

Five reasons, listed in advance. A reason not on this list cannot be used, and adding one is a rulebook change with a version number attached — dated, with the previous wording visible. The fifth was added exactly that way, at rulebook v0.10, and its review mechanics were revised the same way at v0.11. This is the set of criteria that decides what goes into the numerator above, and it is published before there is a payout gate to apply it to.

1. 1. Copying another funded account. Established on execution timestamps across accounts linked to one identity, not on a view about strategy. The evaluation fee is returned.
2. 2. Account holder is not the payout beneficiary. The name on the account and the name on the destination must match. Refused pending correction rather than permanently, and payable once they do.
3. 3. Identity verification not completed. Refused pending completion, not forfeited. The amount stays payable indefinitely and is paid the day verification clears.
4. 4. The request exceeds the accrued balance. An arithmetic refusal. The payable part is paid; the remainder is not created by asking.
5. 5. Account multiplication under rule 10. Opposite positions netting to roughly zero across linked accounts, an account transferred or traded by a third party, or the same trades duplicated across a cluster to multiply payouts. Established on the trade-correlation review that runs through the whole evaluation, not invented at the payout gate; the refusal row cites rule 10. A request under review is marked so in the ledger, with the review’s start time, and the review is capped at 72 hours — at the cap it either settles with the +2% penalty or is refused under this clause.

One ground that appears in this industry is deliberately absent here: a “risk management review” with no stated criteria. It is not in the rulebook, so it cannot be used to refuse a request — which means it cannot be moved into the numerator. Where other firms apply a consistency score after the fact, upme publishes the cap in advance instead: no single day and no single trade above 40% of total profit, stated as a number rather than judged at the gate.

Corrections

If a figure here ever disagrees with the ledger on /payouts/, the ledger is the source of truth and this page is corrected, dated, with the previous wording visible — rather than a silent edit. Corrections go to hello@upme.com, and the same standard applies to the research library: anything we write about another firm is drawn from that firm’s own published terms, and if a term has changed since we read it we would rather be told than be stale.

Sources this page rests on:

1. Prop firm pass rates and payout rates
2. When a prop firm payout can be denied
3. The payout ledger (currently empty)
4. The challenge rulebook