Systematic backtesting · Statistical validity

Backtesting that
doesn't lie to you.

The Deflated Sharpe Ratio. Walk-forward decay. CPCV. The statistical machinery quant funds use to know whether a backtest is real — now in a tool built for serious retail traders who are tired of being misled by their own Sharpe ratios.

Start free — no card requiredSee sample backtest →
DSR
on every backtest, every tier
SHA-256
hash-signed deterministic results
< 5 s
median backtest runtime

The problem with most backtests

You pick indicators, tune parameters, see an in-sample Sharpe of 2.4, and conclude you have a strategy. You don't. You have a curve-fit. The Sharpe ratio rewards you for looking at data until something works.

Institutional quant research corrects for this. Retail tooling pretends the problem doesn't exist. Kestrel Signal closes that gap — and shows you the validity metrics whether you want to see them or not.

What makes this different

Built around the metrics that matter

Deflated Sharpe Ratio

Your raw Sharpe ratio is optimistic by construction. DSR corrects for the number of strategies you tested, selection bias, and non-normal returns. Every backtest shows it — even on the free tier.

Walk-Forward Validation

Split your data into rolling in-sample / out-of-sample windows. See exactly how much edge survives each fold. Walk-Forward Efficiency tells you whether you have a real strategy or a curve-fit.

CPCV — Probability of Overfitting

Combinatorial Purged Cross-Validation gives you the probability that your backtest result is noise. Quant funds use it. Retail tools ignore it. We don't.

#
Deterministic, hash-signed results

Identical inputs always produce byte-identical results. Every backtest is fingerprinted with a SHA-256 hash. You can verify your results didn't change. No silent floating-point drift.

Who this is for

You write Python. You know the Sharpe ratio is misleading but you've been using it anyway because everything else does too. You trade your own capital, run real tests, and you're frustrated that every backtesting tool optimistically hands you a Sharpe of 2.1 and calls it a day.

Kestrel Signal is built for you — not for beginners who want signals, not for funds who have their own infra. For the serious retail quant who wants to know if their strategy is real before they trade it.

New to DSR or walk-forward? Start in /learn →Read the blog →

Pricing

Start free. Upgrade when you need more.

Full details →
Free
Free
3 symbols · 10 bt/day · DSR always shown
Start free
Researcher
S$19/mo
15 symbols · walk-forward · intraday data
Get Researcher
Most popular
Lab
S$49/mo
CPCV · Python upload · Firecracker sandbox
Start 14-day trial
Pro
S$99/mo
REST API · custom data · 3 team seats
Get Pro

All prices SGD · Annual billing saves ~20% · Lab includes 14-day trial · DSR free on all tiers

Know if your strategy is real.

Free tier includes DSR and validity metrics on every backtest. No credit card. No waitlist.

Start free