The rebuilt “funding returns” page in the console is live. The changes concentrate on two things: the reporting basis, and drilling down.
Four bases for the return. The return column now switches between all-time, daily, monthly and annual, with the return rate and sample count switching alongside it. Previously there was only the cumulative view, which made it impossible to see what a given month actually produced.
Two levels of grouping. The same table reads per account (one row per trading account) or per operator (accounts under one operator collapsed into a single row). Each level runs its own server-side aggregation rather than summing in the browser.
Drill down to the month. Expand any row for the cell-level detail: the annual view breaks into months, and the monthly view is the floor. The operator view has one extra level, so you can first see which accounts made up a given cell.
Disabled accounts now report consistently. A disabled account still appears row by row in the detail — it did earn that money — but it no longer counts toward any total. The fact table and the operator table draw on the same list, so the two reconcile.
Every row carries its effective sample count. The denominator of the return is account equity at the settlement instant, which can only be read by standing on the settlement point — three of them a day on the standard schedule, 00:00 / 08:00 / 16:00 UTC, or 1,095 a year, though contracts moved to 4-hour settlement do not sit on that number. Rows carrying equity snapshots for under 50% of their settlements are flagged: those figures are indicative only, because a rate computed from three samples looks identical on screen to one computed from three hundred.
The figures come from the rebuilt collector. The public, same-source breakdown by symbol is the funding rate dataset — 450 symbols, 878,142 settlement records — and the platform’s own year-by-year table is on performance, worst cell included: ETHUSDT returned 7.873% amplified across 2022.