Free courseLesson 02 of 10

Why most trading ideas fail

Three reasons nearly every idea falls apart, and how to spot each one.

Costs: the quiet killer
Overfitting: learning the past by heart
Chance: when luck looks like skill

Costs

Every trade costs money twice: the commission you pay your broker, and slippage, which is the small gap between the price you wanted and the price you got.

A rule that makes a little on each trade before costs can lose on each trade after them. That's why every test in this course includes costs from the first run.

Overfitting

Tune a rule long enough and it fits the past perfectly. That's the problem: it has learned the past by heart, not how the market behaves.

It's like memorizing last year's exam answers. You'd score 100% on last year's exam and fail the next one.

Chance

Flip a thousand coins ten times each, and a few of them come up heads ten times in a row. None of those coins is special.

Test a thousand trading ideas and a few look brilliant by luck alone. If you only keep the prettiest one, you've probably kept a lucky coin.

Three questions for every idea

Does it survive costs?

Does it survive data it has never seen?

Does it beat a random version of itself?

The rest of the course is these three questions, asked properly.

Do it with your AI assistant
Explain these three things to me as if I'm new to trading, with one everyday example each: trading costs and slippage, overfitting, and luck when testing many ideas. Then ask me three short questions to check I understood. Ask them one at a time and wait for my answer.
You're done when
  • You can explain costs, overfitting and chance in your own words
  • You answered your assistant's three questions