Finding and testing your own edge

15 minutesLesson 3 of 6
Femi, the HustlCraft strategy mascot

An edge is not a feeling that a setup looks right. It is something you can actually count. This lesson is about counting it honestly.

What an edge actually is

An edge is a repeatable advantage in a strategy's rules that survives real costs, measured across enough trades to mean something. It is not a single good week, and it is not confidence.

An earlier stage showed a strategy with a genuine edge that still lost money once realistic costs were counted. Finding an edge and testing whether it survives costs are two separate jobs, and both matter.

What the research says about finding one

Less than 1 percent of day traders reliably earn positive abnormal returns net of fees, in this study of retail accounts Barber, Lee, Liu and Odean (2014), Journal of Financial Markets
44, 24, and 15 percent were the one, two and three year survival rates found for day traders in this study of Taiwan Stock Exchange transactions Barber, Lee, Liu, Odean and Zhang (2020), Review of Asset Pricing Studies

These numbers are not a reason to give up before starting. They are a reason to test honestly instead of assuming a good-looking setup is automatically an edge.

Trading carries risk and you can lose money.

A way to test without lying to yourself

  1. Write the setup down as one sentence before testing it, the same sentence that goes into the plan from the first lesson.
  2. Run it on a demo account or the smallest possible live size, for a number of trades large enough to matter, not five or ten.
  3. Add realistic costs into the result, the same spread, commission, slippage and swap covered in an earlier stage, rather than assuming a clean fill every time.
  4. Record every trade in the journal from the last lesson, including the ones that broke the rule.
  5. Only trust the result once it holds up across enough trades and more than one kind of market condition.
Femi, the HustlCraft strategy mascot
Femi's warning

A handful of winning trades on a demo account proves almost nothing. Believing it does is how a good-looking setup gets trusted with real money before it has earned that trust.

What counts as enough

There is no single magic number, but a sample too small to reveal a losing streak is too small to trust. A setup tested across only a handful of trades, in only one market mood, has barely been tested at all.

An edge that only works in one narrow condition is not a broken idea. It is simply a smaller edge than it looked, and the plan should say so.

The picture to keep

Testing an edge is like a market trader tasting a new supplier's rice for months before switching the whole shop over to it, not deciding after one good pot.

PreviousNext: When to use the software, and when not to