How it works
Evidence first, then arithmetic
Tipmkt measures what a record does to a bankroll. Everything else—claims, streaks, screenshots—sits outside the product.
Calibration and simulation windows
Uploaded history is split in two. The earlier calibration window supplies the probability estimates Fractional Kelly needs. The later simulation window stays unseen and is used to test all four staking strategies. Nothing after a selection may influence its stake.
Break-even win rate
At average odds of 2.00 a record needs 50% winners to stand still. We publish the break-even rate beside the achieved rate so the gap—the actual edge—is visible without arithmetic.
Evidence statuses
Seven states, stated plainly
Uploaded by the provider. Labelled as a provider-supplied historical simulation and excluded from verified rankings.
Recorded by a recognised third party. Shown separately from Tipmkt's own record.
Settled by Tipmkt against timestamped selections.
Live selections tracked without subscription sales, to establish a record.
An active, continuously settled record on Tipmkt.
Too few settled selections to support any conclusion.
The record shows drawdown or variance that demands caution.