POC
🔍 TOOL · Is my edge real?

Real or Noise?

A profitable track record feels like proof — but luck alone produces plenty of them. This tool compares your results with thousands of no-edge systems, then does the part that really matters: it allows for how many variations you tried to find this one. That's where a lot of promising-looking edges gently come apart — and that's completely normal, not a failing.

How it works
1 You describe your recordHow many trades you took, how often you won, and how big a typical win is next to a typical loss.
2 We simulate 2,500 systems with no edgeSame win and loss sizes as yours, but no skill in them whatsoever — pure chance. Then we count how often chance alone matches you.
3 We allow for how many you triedSearch hard enough and chance throws up a winner on its own. The more variations you tried, the higher the bar your result has to clear.
One unit to know first: “R”. R is simply whatever you risk on one trade. So a loss is −1R, and a 1.7R win makes back 1.7 times what you put at risk. Measuring in R rather than pounds means the answer doesn’t depend on the size of your account. Every figure on this page is in R.

What this can’t tell you

Worth reading before acting on any verdict above — including a good one.

  • “Likely real” is not permission to trade. It means luck is an unlikely explanation for these particular numbers. It says nothing about whether the edge will keep working, or whether it suits your account, your costs or your temperament.
  • It only knows the four numbers you typed. Working from a win rate and an average payoff, it cannot see losing streaks, changes in position size, or one enormous trade quietly carrying the whole record.
  • It assumes every trade is independent and from the same system. If you changed your approach partway through, or your positions overlap and move together, the true uncertainty is wider than what is shown here.
  • It cannot spot an edge that has already died. A record that was genuine for two years and flat for the last six months can still pass.
  • It cannot check your data. A backtest with look-ahead, survivorship bias or optimistic fills passes exactly as easily as an honest one.
  • The variation count is yours to be honest about. Nobody can verify it, and under-counting it is the easiest way to make a mined result look real.
The thinking behind it
09
Luck looks exactly like skill
READ →
Next

Suppose the edge is real. The next question is whether it survives what trading actually costs — because friction is charged on every trade, and it never has a losing day.

The cost of trading →