How we grade

What you can rely on

• Every strategy shows its record on past prices, then the days since, paper tracked.
• When part of a record was rebuilt after the strategy was chosen, that span is shaded and labelled.
• When we withdraw a strategy, it is listed with the reason. See withdrawn strategies
• A test result is a measurement. It is not a promise.

Step 1 · Written down first

Before we look at a result, the idea is written down: which stocks it may hold, when it buys and sells, and what would count as success or failure.

Step 2 · Tested on separate stretches of past prices

The same fixed rules are scored on several separate stretches of past prices. We chose our strategies after seeing these results. A final window is held back from design. For several of our own strategies it has already informed which version we kept, so we do not treat it as a fresh test.

Step 3 · It must beat its own luck

A lucky strategy can look like a good one. So we shuffle the strategy's own timing 200 times and ask whether the real result stands out from its own luck. Some older house tests used fewer shuffles.

Step 4 · Paper tracked

A strategy that passes is then paper tracked day by day, with no real trades. This is where many tests on past prices fall apart, and we show what happens anyway.

What the labels mean

What we do not claim

• No guarantees. Testing on past prices does not see the future.
• A paper tracked record is not your money: real trades, costs and taxes differ.