Profit Factor Explained: What's Good and What's Just Luck

Profit factor is the most quoted statistic on a backtest report and the easiest one to misread. It is a single number that looks like a verdict, which is exactly the problem: two traders can post the same 1.60 and only one of them has anything that repeats.

This is what the number is, what yours should be given how you actually trade, and the three checks that tell you whether it is an edge or one good trade wearing a disguise.

The Formula

Profit factor = Gross profit / Gross loss

Gross profit is the sum of every winning trade. Gross loss is the sum of every losing trade, entered as a positive number. That is the whole calculation, and it is the same definition MetaTrader uses in its Strategy Tester report, so the number on your backtest and the number in your journal are comparable.

Worked on 40 trades:

• 16 winners totalling 9,600 dollars
• 24 losers totalling 6,000 dollars

9,600 / 6,000 = 1.60

Read it as: every dollar this strategy lost bought 1.60 dollars of profit. Above 1.0 made money over the sample, below 1.0 lost it, exactly 1.0 broke even.

Note what is missing from that sentence. Nothing about how many trades won, nothing about how long it took, nothing about how much of the account was at risk to get there.

What Yours Should Be

There is no universal "good" profit factor, and the numbers usually quoted as benchmarks are conventions passed around rather than findings from anywhere. You can do better than a benchmark, because profit factor is not independent of the rest of your statistics. It falls out of them.

Split the formula. Gross profit is the number of wins times the average win. Gross loss is the number of losses times the average loss. So:

Profit factor = (Win rate / Loss rate) x (Average win / Average loss)

Check it against the example above. The win rate is 16/40 = 0.40 and the loss rate is 0.60. The average win is 9,600/16 = 600 dollars, the average loss is 6,000/24 = 250 dollars, so the reward-to-risk is 2.4.

(0.40 / 0.60) x 2.4 = 0.667 x 2.4 = 1.60. Same answer.

That identity turns profit factor from a grade into a prediction. Here is what the arithmetic implies at common combinations:

Three things worth taking from this grid.

Every 1.00 in it is the break-even line. A 40 percent win rate at 1.5R and a 50 percent win rate at 1R are both exactly break-even, which is the same statement as the break-even win rate table of 1 / (1 + R) in our trading expectancy write-up, arrived at from the other direction.

A high profit factor is not evidence of skill on its own. A 60 percent win rate at 3R gives 4.50. So does an unusually kind six weeks. The grid tells you what to expect, not whether you earned it.

Your target profit factor is set by your style, not by ambition. A 30 percent win rate trend follower running 3R targets should expect something near 1.29, and chasing 2.00 means changing the strategy rather than trying harder. Comparing your figure to a stranger's is comparing two different systems.

How It Relates to Expectancy

Profit factor and expectancy answer different questions, and they are linked exactly:

Expectancy = Loss rate x Average loss x (Profit factor - 1)

The example again: 0.60 x 250 x (1.60 - 1) = 0.60 x 250 x 0.60 = 90 dollars per trade.

Confirm it the direct way: (0.40 x 600) - (0.60 x 250) = 240 - 150 = 90 dollars. Same number.

Which one to reach for:

• Profit factor is unitless, so it compares things of different sizes. Your gold trades against your index trades, one strategy against another, a backtest on a demo account against live trading at ten times the size.
• Expectancy is money or R per trade, so it answers whether taking one more trade is worth it, and it converts into an account projection.

The relationship also exposes something the ratio hides. Two traders both at 1.60 have very different businesses if one loses 250 dollars per losing trade and the other loses 25. Profit factor is deliberately blind to size, which is a feature when comparing and a trap when planning.

Where the Number Lies

1. It is a ratio of sums, so one trade can carry it

This is the failure that matters most, because it is invisible in the number itself.

Take the same 40 trades, but suppose one winner was worth 4,000 dollars of the 9,600. That single trade is 42 percent of all the profit the strategy ever made.

Delete it and recompute. Gross profit falls to 5,600 across 15 winners, gross loss is unchanged at 6,000 across 24 losers:

5,600 / 6,000 = 0.93

The strategy went from comfortably profitable to losing money on the removal of one trade. Meanwhile the win rate barely moved, from 16/40 = 40.0 percent to 15/39 = 38.5 percent, because win rate counts trades and profit factor weighs them.

That gap is the whole point. Any statistic built from sums can be dominated by its largest term, and win rate will not warn you, so a healthy-looking pair of numbers can rest entirely on a single Tuesday.

2. A thin denominator makes it unstable

Gross loss is the denominator, so the number of losing trades sets how much the ratio can be trusted.

A sample with 6,000 dollars of gross profit and 4 losses of 500 dollars each shows a profit factor of 3.00. Add one more ordinary 500 dollar loss and it drops to 2.40. Replace that with a single trade where the stop was widened and it cost 2,000 dollars, and gross loss becomes 4,000, so the figure lands at 1.50.

One trade, from 3.00 to 1.50. Anyone reading the first number would have described the strategy in very different words than the third.

So the sample-size question for profit factor is not how many trades you have logged. It is how many losses are in the denominator.

3. Break-even trades cannot move it, and that is a real advantage

A scratch trade at exactly zero adds nothing to gross profit and nothing to gross loss. Which means profit factor is identical whichever convention you use for break-even trades, while win rate swings hard.

46 trades: 16 winners, 24 losers, 6 scratches.

The win rate moves 13 percentage points on a bookkeeping decision. The profit factor does not move at all.

Use that. When you are comparing two setups and the win rates disagree, check whether the disagreement is real or whether one of them simply has more scratches in it. Profit factor is the tiebreaker, because there is no convention for it to be sensitive to. The flip side is that a strategy producing constant break-evens looks the same to profit factor as one that never enters, which is why it belongs next to your other numbers rather than replacing them.

4. Gross means gross

If the figures came off a report that measures before costs, the profit factor is a marketing number. Apply an 8 dollar round-turn cost to the running example: gross profit becomes 9,600 minus 128 = 9,472, gross loss becomes 6,000 plus 192 = 6,192, and the ratio falls from 1.60 to 1.53.

Small in this case because the average trade is large. On a scalping strategy where the average win is 40 dollars, the same 8 dollars is 20 percent of it and the effect is not small at all.

5. It has no opinion about the order things happened in

A profit factor of 1.60 describes a bag of trades with the sequence thrown away. The same bag can produce a smooth equity curve or a run of nine losses that ends a funded account before the winners arrive.

If you are trading an evaluation, this is the limit you feel first. The ratio cannot fail you, the sequence can. The arithmetic for that is in how to pass a prop firm challenge and, for the accounts where the loss floor moves under you as you profit, in trailing vs static drawdown.

6. Mixed position sizing makes it a different question

If you risked 250 dollars a trade in March and 1,000 in June, a dollar-based profit factor is mostly telling you which months you sized up in. One large-size winner in a small-size sample distorts the ratio the same way an outlier does.

The fix is the same as it is for expectancy: compute it in R. Sum the positive R multiples, divide by the absolute sum of the negative ones. A full stop-out is 1R by definition, so the sizing drops out and the number describes the strategy again.

The Three Checks

Before you trust a profit factor, including your own, run these. They take about two minutes in any journal that can filter.

1. Delete your best trade and recompute. If the figure falls under 1.0, as it did in the example above, your sample contains one trade and some noise. This is the single most informative thing you can do with a profit factor, and almost nobody does it.

2. Count the losses. If gross loss is built from fewer than a couple of dozen trades, treat the ratio as a rough direction rather than a measurement. Write down what one extra bad loss would do to it, because you will eventually take one.

3. Split the sample in half by date. Compute the first half and the second half separately. A strategy holding at 1.5 and 1.7 is a strategy. One that goes 2.6 then 1.0 is a market condition that ended, and the pooled 1.60 is describing a period that is over. Filtering by year is the fastest version of this test.

A number that passes all three on a decent sample is worth planning around. A number that fails any of them is worth investigating rather than reporting.

Where It Is Actually Useful

The reason to keep profit factor next to expectancy, despite everything above, is that it is unitless. That makes it the right tool for one job in particular: comparing slices of your own trading that are not the same size.

Your EUR/USD trades and your gold trades have different average risk, different volatility and different trade counts, so their net P&L figures are not comparable. Their profit factors are. Same for London versus New York, one setup against another, or the version of your strategy you traded before a rule change against the version after it.

That comparison is where the actionable finding usually is. A total profit factor of 1.4 is a fact about you. Discovering that it is made of 2.1 on one symbol and 0.8 on another is a decision, and dropping the 0.8 raises the total without you learning anything new about the market.

It is also the right lens for a backtest, since replay produces a large sample quickly enough that the denominator gets thick. Our bar replay backtesting guide covers building that sample without letting hindsight inflate it, and if your live fills come from MetaTrader, auto-importing MT4 and MT5 trades gets the executed side in without retyping it.

One more thing profit factor cannot see: whether the trades in it followed your plan. A 1.8 built out of rule breaks is not a strategy you can repeat on purpose. Putting a number on that side is what the discipline score is for.

Where Profit Factor Lives in TradingSFX

It is computed for you from the trades you log, with no setup.

• Dashboard card. Shows profit factor when the dashboard is set to analyse by profit, colour-coded around the 1.0 line, and it switches to cumulative R when you flip to the R view. Available on the free plan.
• Trades tab. The stats strip recomputes it for whatever filter is currently applied, so checking a single symbol or a date range is one click rather than a spreadsheet.
• Per symbol and per strategy. The detailed analysis and strategy breakdown pages compute it per slice, which is the comparison the whole previous section is about. Pro and above.
• Backtest sessions. The replay backtester reports profit factor per session alongside win rate and net R, so a backtested strategy and a live one are measured the same way. Pro and above.
• PDF performance report. Profit factor appears in the printable report for a mentor or a prop firm, next to expectancy and win rate. Pro and above.
• Advanced analytics. Feeds the performance radar next to expectancy, drawdown and Calmar. Premium.

An empty denominator shows as infinity rather than a number, which is the honest display for a sample with no losing trades in it yet.

You can see the whole set running on sample data without an account in the detailed analysis demo.

Bottom Line

Profit factor tells you how many dollars of profit each dollar of loss bought, and nothing else. It does not know your win rate, your position sizing, the order your trades arrived in, or whether you followed your own rules while producing it.

Read it as a ratio you can predict from your own win rate and reward-to-risk, not as a grade. Then earn the right to quote it: delete your best trade, count your losses, split the sample in half. A figure that survives all three is worth building on. A figure that does not is the most confident-looking number on the page and the least informative.

Both numbers need the same thing underneath, which is trades logged with an entry, an exit and a stop. TradingSFX Basic is free forever at 10 trades a month, which is enough to see your own ratio and start pulling it apart.

Published August 12, 2026. All figures in this article are arithmetic worked openly from the stated inputs. No survey data, study results or user statistics are cited, and no competitor pricing is quoted. The gross profit divided by gross loss definition was checked against the MetaTrader 5 Strategy Tester report documentation on 12 August 2026 (metatrader5.com/en/terminal/help/algotrading/testing_report).