Independent statistical audit of backtests and track records

Real edgeor just luck?

Upload the backtest or track record you already have. Rigor tests it against costs, repeated trials and broken data, and gives you a class from A to D.

  • Never connects to your broker or recommends trades
  • Every report can be checked
  • First full report free with your account
Illustration with synthetic data
6audited dimensions
35red flags checked on every file
16prop-firm challenges to simulate
34platforms it recognises

Reads the file you already have

  • MetaTrader 5
  • MetaTrader 4
  • TradingView
  • NinjaTrader
  • QuantConnect
  • backtesting.py
  • vectorbt
  • Myfxbook
  • FX Blue
  • MQL5 Signals
  • CSV

It also recognises the export format of Interactive Brokers, Tradovate, TopstepX, thinkorswim, TradeStation, Charles Schwab, Fidelity, E*TRADE, Webull, Robinhood, tastytrade, eToro, XTB, DEGIRO, Trading 212, Revolut, Zerodha, cTrader, Sierra Chart, Rithmic, Binance, Kraken, Coinbase and KuCoin.

Who it is for

The same rigour, whatever your market.

Forex, stocks, futures or crypto; your own strategy, a robot you bought, a challenge or money you hand to someone else. Find your case.

You buy or run a robot (EA)

The vendor's backtest looks perfect and you cannot tell whether it is overfit.

See what it checks for your case

You trade stocks, futures, forex or crypto

You do not know whether your edge is real or you found it by trying enough.

See what it checks for your case

You are about to pay for a prop-firm challenge

One bad streak can end the account even when the strategy works.

See what it checks for your case

You invest with a manager, a signal or a fund

The percentage you are shown may come from deposits, a few good months or a backtest.

See what it checks for your case

Another case: Before you copy a signal, see the risk it does not show

Another case: If you invest on your own with DEGIRO, Trading 212, IBKR or XTB, check your history

The problem

A good-looking backtest is not evidence.

Almost any strategy looks good on paper. Rigor is there so a pretty curve does not fool you before you risk your money or pay for a robot. These are the three reasons most of them do not hold up outside the tester.

01

Overfitting

Try a hundred settings and keep the best. Chance alone already draws a beautiful curve.

02

Costs

Commission, spread and slippage eat thin edges. Many backtests count them as zero.

03

Data

Duplicate bars, frozen prices or gaps inflate the result without anyone noticing.

How it catches it in the sample

A Sharpe of 1.8 in the tester. Class C in Rigor.

The sample report is an MT5 backtest built from synthetic data. Rigor lowers its class for three things the tester does not show:

See the sample report
  • It came out of 120 settings tried; once they are counted, its Sharpe no longer reaches the threshold.
  • At twice the reference cost it ends in a loss.
  • Its live account, 180 trades, falls outside what its own backtest would lead you to expect.
What we measure

Six dimensions. One verdict from A to D.

Rigor checks your file on six points. Each one comes out as Pass, Weak, Fail or Not measured, with two plain sentences on what it means for you. The methods are public and written down in the methodology.

Statistical significance

Whether the Sharpe ratio is distinguishable from zero given length, skew and kurtosis.

Number of settings tried

How much survives after discounting the number of trials you declare.

Costs

What happens at 1x, 2x and 3x the trading cost, and the break-even cost.

Out of sample

Whether the out-of-sample window you declare holds up.

Data quality and trading pattern

35 red flags: duplicate data, spikes, frozen marks and more.

Benchmark

A comparison against the benchmark you supply, if you supply one.

  • MeasuredWe computed it from your file.
  • DeclaredYou or your platform stated it; we cannot check it.
  • Not measuredData was missing to measure it, and we tell you which.
Process

How it works

  1. Create your account with your email: your first full report is free.
  2. Upload your platform's file as it is: a backtest, an account history or a return series.
  3. You see the A to D class, the charts and what each dimension means, in plain language.
  4. Each further audit is paid from your account, and your reports stay saved there.

Upload the file your platform already saves. If you are not sure which one to export, there is a short guide for each. See the export guides

Real work, not hype

No bots, no signals, no promises.

Rigor sells no strategies and no results: it measures the file you upload and shows you how it measures it. You can check everything in this section before you pay.

See a whole report before you pay

The sample is a full report, with its PDF, built from synthetic data: you see exactly what you get.

See the sample

We sell no bots and no signals

We place no orders and ask for no broker keys, and no report promises results: an automatic filter stops any text that does.

Your file stays yours

It is never published. If you do not unlock the report it is deleted after 30 days, and you can delete your account and your reports whenever you like.

Privacy policy

We execute no trades, hold no funds or keys, recommend no strategies and predict no results. We audit the file you upload.

Who is behind it: Sergio Valle Zarate, México.

Questions before you upload? Write to us on WhatsApp; a person answers. Write on WhatsApp

Pricing

Pricing

Preview
Freeyour first full report; then 3 previews a month

A to D class, what each dimension means, charts, red flags and hashes.

  • Class A to D and a plain-language summary
  • What each dimension means for you
  • Equity, drawdown and monthly return charts
  • Red flags and your files' fingerprints
Start free
  • Card payment from the report itself, processed by Stripe: you see it in full at once, without waiting for a code.
  • Pay by bank transfer or another method we agree on WhatsApp: once the payment is confirmed you receive an access code and enter it in the form or in the report. Ask for a code
Start here

Your audit, from a single file.

Upload the report exactly as your platform saves it. The class, the charts and the explanation of each dimension are free.

  • Your file is never published.
  • Your first full report, free when you create your account; then 3 free previews a month. No card.
  • Deleted automatically if you do not unlock the report.
Questions

Frequently asked questions

Which file do I upload?

Your platform report as it is: MetaTrader 5 or 4 (HTML), TradingView (CSV or XLSX), NinjaTrader, QuantConnect, backtesting.py or vectorbt. To review another trader's account, the CSV history exported by Myfxbook, FX Blue or an MQL5 signal. From your broker, exchange or journal, its trade history as CSV or Excel: it recognises the export format of Interactive Brokers, Tradovate, TopstepX, thinkorswim, TradeStation, Charles Schwab, Fidelity, E*TRADE, Webull, Robinhood, tastytrade, eToro, XTB, DEGIRO, Trading 212, Revolut, Zerodha, cTrader, Sierra Chart, Rithmic, Binance, Kraken, Coinbase and KuCoin, and in any other the columns are recognised by their names. An equity curve in CSV works too, or a PDF statement with its trade table, whose columns you check before it measures.

Does it work for stocks, crypto, futures or a fund?

Yes. Rigor does not depend on the market: it measures the history you upload. For a stock or crypto portfolio, or for a fund or a manager, upload its equity curve or return series (daily, weekly or monthly) as CSV or Excel; for a fund, the monthly returns table from its factsheet works too. TradingView, NinjaTrader, QuantConnect, backtesting.py and vectorbt reports work for any asset.

What do I get?

For free, the preview: A to D class, charts, red flags and what each dimension means. The full report adds every figure, stress tests, risk and capital, the challenge simulator, the live account against the backtest if you upload it, questions for the vendor and the PDF. See the full sample before you pay.

Why upload the MT5 optimisation XML?

Because it counts the configurations you tried. With that real number, the deflated Sharpe discounts the luck of picking the best of many.

What do Measured, Declared and Not measured mean?

Measured was computed from your file; Declared is what you stated and could not be checked; Not measured could not be computed from what you uploaded.

Does this predict results or the outcome of a prop-firm challenge?

No. It measures the statistical evidence in the file you upload. The challenge simulator and the resampled risk are estimates from your own history, with their assumptions written down, not predictions.

Do you check my trades with the broker?

No. We audit the data you supply; we connect to no broker and ask for no keys. That is why the badge says the data was not checked with a broker.

What happens to my file?

It is kept so your report can be regenerated. If unpaid it is deleted after 30 days and only the class and the hashes remain. It is never published: the verification page shows the class, the dimensions and the hashes, and only if you publish it.

What if I forget my password?

In My account you make a recovery key and keep it. If you forget the password, your e-mail and that key let you set a new one yourself, without waiting for an e-mail. If you turned on two-step sign-in, we also ask for the code from your app. We keep only the key's fingerprint, never the key itself.

How well protected is my account?

You can turn on two-step sign-in with an authenticator app (Google Authenticator, 1Password or another): then signing in or recovering the account takes two of these three: your password, the code from the app or your recovery key. In My account you see where your account is open and sign out each session, and you see your most recent sign-ins and security changes (up to 90 days), including wrong-password tries. You can also sign in with a passkey: your phone's or computer's fingerprint, face or PIN, without typing the password. In My account, "Your account's protection" shows what is on and how to turn on the rest.

How is the badge used?

Publish the verification from your report and copy the badge code to your site, Telegram or forum. The badge describes a statistical audit; it is not a promise of results and must not be presented as one.

Before you put money on a strategy, take a close look.

Upload the report and get the class, the charts and their explanation free.

Get guides and news