Someone shows you a strategy with great results: a fund, a trading bot, a signal group. We test whether those results are skill or chance, the same way we tested our own methods, and tell you plainly.
We replay the same trades at random times and on random coins. If luck matches it, there is no skill to pay for.
Rules are frozen, then judged on years the strategy never saw. Most curve-fitted results fall apart here.
On big coins that rose, simply holding often wins. A strategy must beat the lazy alternative, not zero.
Trading fees, spreads and the funding paid on leveraged positions, settlement by settlement.
Real results from our research, written in the form a client receives. Rules were fixed before each test was run.
| Strategy | Tested on | What it claimed to do | What the check found | Verdict |
|---|---|---|---|---|
| Retest entry wait for price to return to a broken level |
848 perpetuals 2022-2026 · 48,626 trades |
Win 70% of trades at +10% target / -30% stop | It did win 70%, but needed 70% just to break even. Random entries with the same exits did as well. | No edge |
| Better exits 104 target, stop and trailing rules |
Same 48,626 entries chosen on 2022-24, judged 2024-26 |
The right exit turns a flat entry into a profitable one | Every exit shape won about as often as its own break-even. Exits changed the shape of results, not the outcome. | No edge |
| Crash warning hedge half when traders are crowded |
BTC 2020-2026 and a 5-coin basket |
Cut the deep drawdowns of a long-term holder | Missed the two worst falls (-77%, -53%). Holding a bit less all the time did as well. | No edge |
| Exchange inflows coins moving to exchanges = selling ahead |
324 tokens on-chain flows |
Big deposits warn of a price drop | Flows followed price instead of leading it. Spikes marked days that had already moved. | No edge |
| Retest entry vs buying the breakout | 848 perpetuals same period |
Waiting for the retest beats buying at once | Waiting earned about 3 percentage points more per trade than buying the breakout at market. | Holds up |
Most ideas fail an honest check. That is the point of doing it before money is committed. Full studies →
The strategy's rules, or its full trade history with dates. Results alone (a chart, a monthly return table) are not enough to test.
Usually one to two weeks, depending on the data.
A fixed price per strategy, agreed before we start. The verdict does not change the price.