Backtesting, free — and honest.
Strategy backtesting is free on Basis, with no account and no run limit. More importantly it is built to tell you when a result is not real, which is the part that usually costs money and is usually still missing.
What you get
- Next-bar-open execution, so a signal is never filled at the price that produced it
- Fees and slippage applied per trade, because a strategy that only works at zero cost does not work
- Walk-forward and out-of-sample testing across chronological splits
- Deflated Sharpe: how good a result would look from pure noise given how many settings you tried
- Strategy signals usable as screener conditions, so a strategy can be tested across a universe rather than one symbol
Why this is usually paid for
Backtesting is nearly always a paid feature, and the free versions that exist tend to be the dangerous kind: they report a return and a win rate and nothing that would tell you the result is an artefact.
The expensive part of an honest backtest is not the compute. It is being willing to show a number that discourages the user, which is a product decision rather than a cost one.
How to use it
- 1
Test on a split, not on everything
Choose settings on the first part of the history and read the result from the part you did not touch. A result quoted from the same data that chose the settings is not a result.
- 2
Count how many settings you tried
Trying thirty parameter combinations and reporting the best is thirty chances for noise to look like an edge. The deflated Sharpe tells you how good the best of thirty would look if none of them worked.
- 3
Put the costs in first
Add fees and slippage before you look at anything else. A strategy that trades often will lose most of its apparent edge to them, and finding that out at the end wastes the whole exercise.
- 4
Compare against holding
On a market that spent the period rising, beating zero is not the benchmark. Compare against simply holding the same instrument over the same window.
Questions
Is backtesting free?+
Yes. Backtesting on Basis is free with no account, no trial and no cap on how many tests you run, because it executes in your browser against market data the chart has already loaded rather than on a server that would have to be paid for. There is no paid tier for it to be limited against.
Does it avoid lookahead bias?+
Yes. Signals are evaluated on closed bars and filled at the next bar’s open, so a strategy can never trade on information the bar had not finished producing. Indicators used in a backtest are the same causal implementations the chart draws, and every one of them is checked to confirm that a later bar never changes an earlier value.
Can I backtest my own strategy?+
Yes. Build the logic in the Formula Lab or the strategy sandbox, and it becomes testable the same way the built-in strategies are — including across a whole screened universe rather than one symbol.
What is deflated Sharpe and why does it matter?+
If you try enough variations, one of them will look good by chance. Deflated Sharpe estimates how high a Sharpe ratio the best of N attempts would reach on pure noise, so you can compare your result against the bar that random chance already clears. A Sharpe of 1.2 from one attempt and from three hundred attempts are very different findings.
Open it and see
The chart opens with this already turned on. Nothing to configure, nothing to sign up for.
Open the chartAlso free
Updated 2026-08-23 · Indicator reference · Documentation · Terminal