What is profit factor? The metric that beats win rate

Win rate is the first thing everyone looks at. It's also one of the least useful numbers on a trader's profile — taken alone. A strategy that wins 80% of the time can still lose money. A strategy that wins 40% of the time can be exceptional. The number that explains both is profit factor.

The definition

Profit factor is simple: gross wins divided by gross losses.

Add up every winning trade. Add up every losing trade. Divide the first by the second.

• PF above 1.0 = the strategy made more than it lost
• PF of 1.0 = breakeven (before costs)
• PF below 1.0 = the strategy lost money overall

A profit factor of 1.5 means for every $1 lost, the strategy returned $1.50. A PF of 2.0 means $2 back per $1 lost. Real, durable edges rarely exceed PF 3.0 on live data — when you see higher, look for a small trade sample or cherry-picked dates.

Why win rate misleads

Imagine two strategies, each with 100 trades:

• Strategy A: 80 wins of $100 each, 20 losses of $600 each. Win rate 80%, total result: $8,000 won, $12,000 lost. PF 0.67 — a loser.
• Strategy B: 40 wins of $300 each, 60 losses of $100 each. Win rate 40%, total result: $12,000 won, $6,000 lost. PF 2.0 — a strong edge.

Strategy A feels comfortable to trade and bleeds money. Strategy B feels uncomfortable — more losses than wins — and consistently profits. Win rate alone told you the opposite of the truth.

The three-number picture

Profit factor works best alongside two other numbers:

• Trade count (n): PF on 20 trades is noise. On 200 trades it starts to mean something. Always ask: how many trades built this number?
• Drawdown: A PF of 2.0 with a 40% max drawdown is a different proposition to PF 2.0 with 12% max drawdown. Profit factor tells you the edge; drawdown tells you the cost of riding it.
• Backtest vs live: Backtested PF is always optimistic. The question is how much it degrades on live data. A 20–30% live degradation is expected and acceptable. A 60% drop is a sign the edge was curve-fitted to history.

How we publish it at Glasshouse

Every strategy in our book shows both backtest PF and live PF, side by side. Some strategies run ahead of their backtest on live data (small samples, likely to revert). Some run behind (degradation from real costs and market impact). We show the gap rather than hiding it — because the gap is the information.

As of this writing, our book spans a median backtest PF of 1.34. Combined live PF across closed trades sits above 1.0. Neither number is a promise — both are the honest current state of work in progress. You can verify the live figures independently via Binance's lead-trader page.

What to look for before you copy anyone

When you're evaluating a trader or desk to copy, run these checks on profit factor:

• PF below 1.2 on live data is thin — transaction costs alone can flip it negative in harder conditions
• Ask for the trade count behind the PF — anything under 50 live trades is provisional
• Ask how much PF dropped from backtest to live — a gap of more than 40% is a warning sign
• A high PF (2.0+) on few trades is less meaningful than a moderate PF (1.4) on 200+ trades

The honest summary

Profit factor is the single ratio that tells you whether a strategy actually has an edge — not whether it felt like it did. It takes win rate, loss size, and win size, and collapses them into one number. Use it alongside trade count, drawdown, and the backtest-to-live gap, and you have most of what you need to judge a trading desk before you copy a cent.