I built Kiploks after watching my own backtested strategies collapse the moment they hit live markets. The strategy looked great on paper 40% annual return, low drawdown but it was overfit to historical noise. Lost real money finding that out.
The core problem: most traders validate strategies on the same data they optimized on. Walk-forward analysis fixes this, but doing it properly is surprisingly hard to get right WFE calculation, OOS retention, parameter sensitivity, performance degradation across regimes. Most people skip it or get the math wrong.
Kiploks does this automatically:
- Splits your backtest into IS/OOS windows and measures how much performance survives into out-of-sample
- Detects parameter fragility (one parameter with R²=0.93 sensitivity = profit island, not an edge)
- Gives a deployment verdict: REJECT / RESEARCH ONLY / APPROVED
- Currently supports OctoBot exports; Freqtrade and raw CSV next
It's invite-only right now at kiploks.com. Happy to give HN access to anyone who wants to try it.
The biggest insight I kept seeing: strategies don't fail because the logic is wrong they fail because the backtest never tested what happens when the market regime changes. That's what I'm trying to make visible before you deploy, not after.
What would you want to see in a tool like this?