A laboratory for honest trading research

Most backtests are profitable lies.

Build a strategy with no code. Drag it through the worst crashes in market history. Then let five statistical tests tell you — without mercy — whether your backtest survives the traps that fool everyone, or merely looked profitable.

Build a strategy, honestly → or score a backtest you already have

Exhibit A — we scored our own system first

Likely a lie

NIFTY option-selling system

1,265 trades · 10 years · built by us, killed by us

31.7 / 100 honesty

✗Loses money before costs are even applied (−₹99,444 net)
✗Edge vanishes out-of-sample — the curve was fitted, not found
✗50 parameter variations tried — the "winner" was data-snooped

This was our own product. The engine told the truth, so we shut it down and built the engine into a company instead.

How the lab works

Three steps. No code. No mercy.

Assemble a strategy

RSI, moving averages, MACD, SuperTrend, Bollinger — snap conditions together like sentences. Every component explains itself as you add it.

Choose the battlefield

Decades of real daily history — US giants since 1970, Indian indices since 2007. Then pick an era: a calm decade, or the 2008 collapse.

Receive the verdict

Out-of-sample decay, Monte-Carlo luck testing, real-cost survival, survivorship exposure, sample size — one honesty score, stamped. It measures whether the backtest is honest, not whether it will profit. Nothing can promise that.

⚡ Crash Labs

Bull markets flatter everyone. Crashes tell the truth.

Every other platform defaults to recent, friendly data. Here, the worst regimes ever recorded are one click away — because a strategy you wouldn't hold through 2008 is not a strategy.

2000–02Dot-com bust−49% over two years

2007–09Global Financial Crisis−56% peak to trough

2013Taper tantrumrates shock, EM rout

2020COVID crash−34% in 23 days

2022Rate-hike grindthe slow bleed

House rules

Built to stop you from fooling yourself.

Every tweak is counted.

Each parameter change you try is recorded and handed to the overfitting test. Fiddle twenty times and the engine demands twenty times the proof.

No peeking at tomorrow.

Signals are read on a day's close; trades fill at the next open. Look-ahead bias is impossible by construction, not by promise.

The stop fills first.

When a stop and a target are both touchable in one bar, we assume the worst. Optimistic fills are how fake equity curves are born.

No cherry-picking names.

Strategies run across a whole arena of symbols, never the one stock that happened to work. Selection bias is still bias.

We sell no signals.

No tips, no calls, no "join my channel." We build instruments that test your ideas. We win when you stop losing.

Our failures are public.

The first system we scored was our own — it failed, we published it, and it's the reason this company exists.

Would your strategy survive an honest score?

Build it in two minutes. The verdict is free.

Enter the Strategy Lab →

Educational only. Not investment advice. Backtested results are hypothetical, use end-of-day data, and will differ from live trading. The Honest Quant is not a SEBI-registered adviser and does not sell signals or recommendations — it builds tools that help traders test their own ideas honestly.
About · FAQ · Terms · Privacy