expectancy and the 400 trades it takes to see it

[Shadow Trader]

Expectancy is the average result per trade, usually in R. Win 45% of the time, win 1.5R when right, lose 1R when wrong. That is 0.45 x 1.5 minus 0.55 x 1 = 0.675 - 0.55 = +0.125R per trade. Positive, but thin.

Here is where it gets uncomfortable. The example is illustrative, but the maths of the noise is real. Each trade ends at either +1.5R or -1R, which gives a standard deviation of about 1.24R per trade. The standard error of the average after n trades is roughly 1.24 divided by the square root of n.

After 100 trades the standard error is about 0.124R. The estimated edge is 0.125R. So the estimate sits only about one standard error above zero, which is nowhere near enough to rule out no edge at all.

To get the standard error down to half the edge (about 0.0625R), the count needs to be around 400 trades. Then the average sits two standard errors above zero, a common rough bar for "probably not luck".

Thin edges need big samples. For a busy scalper 400 trades might be a few weeks. For a slower setup it can be most of a year.

How many trades do you wait for before deciding a setup has an edge, and where does that number come from?