Before you risk it
Every bot shows you a backtest.
We show you whether it means anything.
A backtest is easy to make look good. We re-test the same setup on periods it has never seen, charge it the fees you actually pay, and say plainly whether it holds up.
We tested our own strategy first
Our founder's own live strategy looked like it made +54% a year. Run through this tool, it turned out to make nothing — the gain came from a flaw in how the backtest filled its exit orders.
We publish that. A tool that can only say yes is a marketing tool.
What we check that others skip
Periods it has never seen
We pick the settings on older data, then score them only on what came after. Picking and scoring on the same data is writing the exam after reading the answers.
Your fee rate, not a nice one
The same setup lost money 3 years out of 5 at one fee level and made money every year at another. The gap was 8 basis points.
How many settings were tried
Try a thousand combinations and the best one always looks good — even on pure noise. We account for how many were tried.
Whether the data itself is sound
One exchange re-used a ticker for a different coin. A naive test read that as a +224% gain on an asset that had lost 99.99% of its value.
What this is not
- Not a promise of profit. A setup that held up before can still fail.
- Not investment advice, and we never hold your money.
- Not a claim that we found a way to beat the market. Most setups fail here.