Blog

How to read a trading track record: losses first

A win rate means nothing without the losing count, the sample size and proof that no entry was removed. Here is how to read a track record, and a drawdown, properly.

A win rate on its own tells you nothing. It only means something with the losing count beside it, the number of closed signals behind it, and proof that no entry was removed.

What a track record has to show

A track record is a list of closed outcomes. Check five things:

  1. Losses are listed. Not summarised, not averaged away. Each losing signal appears as its own entry.
  2. The losing count sits next to the win rate. A win rate shown alone hides how many trades produced it.
  3. Only closed signals count. An open signal has no result yet. Counting it as a win because it is currently in profit inflates the record.
  4. The sample is stated. A handful of closed signals can show any win rate by chance. Ask how many closed.
  5. Entries cannot be edited or deleted afterwards. If a record can be rewritten, it is a marketing page.

Why win rate misleads

Win rate counts how often a signal hit its target. It ignores how far the target was from the stop. A robot can win most of the time and still lose money if each loss is larger than each win. A robot can lose most of the time and still pay if its wins are much larger.

So read win rate with the size of the average win and the average loss, which come from each signal's entry, stop and target. See how to read a trading signal.

How to read a drawdown

A drawdown is the fall from a peak to the next low. It answers a different question from win rate: how bad did the worst stretch get, and how long did it last?

Recovery is asymmetric. A loss of half needs a gain of the entire remaining half to get back, which is a doubling. The deeper the drawdown, the more disproportionate the climb back. A record that shows no drawdown has either not run long enough or is not showing everything.

Also ask how long the drawdown lasted. A shallow one that lasts a year is harder to live with than a deep one that recovers in weeks.

Where invalidation fits

Each Compoyo signal states its invalidation: the condition that proves the reasoning wrong. When a signal closes at a loss, the record can show whether the invalidation was respected. That is more useful than the loss itself. See what is trade invalidation.

The live record

The live record is the permanent published history of closed signals, losses included. A win rate there never appears without the losing count beside it. Signals come from our robots and from Prediction Edge, which uses shifts in expectation as an input to a tradeable asset; see shifts and gaps.

Compoyo never executes a trade and never touches anyone's money. A signal is information, and you act on it, or not, in your own brokerage account. To test a signal first without risk, use a demo account: how to practise a signal on a demo account.

What would prove this wrong

If records with hidden losses reliably matched fully disclosed ones over large samples, the case for demanding full disclosure would weaken. That cannot be tested without seeing the hidden entries, which is the point.

Trading involves risk of loss. Compoyo provides market information and signals only. It does not trade your account or give investment advice. Past performance does not indicate future results.