Forward Test
Past signals, tracked forward from the day they printed.
The tool runs at /receipts and needs a $49.99/mo membership. There is no free tier and no trial, so opening it signed out lands on the pricing page rather than the board. This page is the open documentation of what it does.
A backtest is a story you tell after choosing the ending. The Forward Test is the opposite arrangement: each time a top-ten sharps ranking is published, the ranking is written down along with every listed wallet's official all-time profit at that exact moment. Later, the same wallets are read again and the difference is reported.
That difference is the whole product. It is not a simulation, it is not re-optimised, and it cannot be improved after the fact, because the cohort was fixed before the outcome existed.
How a cohort is built and scored
A snapshot captures the top ten wallets by Sharp Score at publication time, each with its official all-time profit as Polymarket itself reports it. That figure is used rather than the sum of open positions, because a position that has already been redeemed disappears from the open book entirely and would quietly vanish from any total built that way.
Forward profit for a wallet is today's official all-time profit minus the figure recorded at snapshot. The cohort's number is the sum across its wallets, and the newest thirty cohorts are shown.
The twenty-four-hour rule and the exclusions
A cohort does not count until it is more than twenty-four hours old. Before that there is nothing forward to measure, and reporting a number would be reporting noise. On a fresh deployment the page therefore has nothing to show, which is the correct output rather than a failure.
Any wallet without an honest baseline — no official profit reading at snapshot time, or none available now — is excluded from the sum and counted as excluded on the row. Nothing is guessed at to make a cohort look complete.
Where the history is stored
When a database is configured, snapshots are written to it and survive restarts. Without one, the module falls back to an in-process buffer that resets to empty on every restart, redeploy or new server instance — preview behaviour, useful for exercising the interface and useless as durable history. The page reports which backend it is reading, so a short history is never mistaken for a bad one.
Capture runs on a schedule, with an opportunistic floor of one capture per kind per hour, so ordinary traffic accrues history without stampeding the upstream APIs.
What a green cohort actually means
It means those wallets, as a group, made money in the period after we ranked them. It does not mean the ranking caused it, and it does not mean copying them would have produced the same figure — you would have entered later, at worse prices, on a subset of their trades.
The honest use is as a sanity check on the ranking method over time. A method that consistently crowns wallets who then go flat is telling you something, and this is the page where that would show.
What it cannot tell you
Every tool has a boundary, and knowing where it sits is the difference between using one well and being misled by it. For Forward Test:
- It needs stored snapshots and a cohort older than twenty-four hours before it can display anything. An empty board on a young deployment is correct.
- It measures the wallets' whole accounts, not only the trades we surfaced. Their forward profit includes everything they did, related to our ranking or not.
- It is not a record of your results. Nothing here models entering after the sharp, paying a worse price, or missing trades entirely.
- Without a configured database the history is in-process and resets on redeploy, so the sample can shrink to nothing without anything being wrong.
- Wallets missing an official profit reading at either end are dropped from the sum. The count of dropped wallets is shown, but their outcome is unknown.
- Past performance does not guarantee future results, and a forward test is still a description of a period that has already happened.
What it is built on
- Snapshots of the published top-ten Sharp Score ranking, written at publication time.
- Polymarket's official all-time profit endpoint, read at snapshot and read again now — the same number Polymarket shows on a profile.
- A database table when one is configured; otherwise a non-durable in-process buffer, reported as such on the page.
Related reading
- Sharp Score: How to Find Skilled Traders on Polymarket and Kalshi — Guide
- Trader Risk Metrics: Reading Polymarket and Kalshi Leaderboards Beyond Raw PnL — Guide
- How to Copy Trade (Tail) Sharp Wallets on Polymarket — Strategy
Tools that pair with it
Wallets ranked by repeatable skill, not bankroll or luck.
Where tracked sharp capital leans, market by market.
The session's tracked flow, written up.
WhaleTracks is informational analytics, not financial advice. Market data comes from Polymarket, Kalshi and Manifold's public APIs; WhaleTracks is not affiliated with any of them. Past performance does not guarantee future results.