An 80 percent win rate can still lose the account

A prop firm dashboard shows six or seven numbers and most traders read exactly one of them. If you want profit factor, win rate and drawdown explained properly, start somewhere other than the definitions: each metric exists to answer a separate question, and reading the wrong one is how a losing account looks healthy for weeks.

Win rate answers how often you were right. Profit factor answers how the money split between right and wrong across the whole record. Drawdown answers how bad it got on the way. Expectancy answers what one average trade is worth. Trade count answers whether any of the other four mean anything yet.

Every figure below comes from an invented worked example built to show mechanics. It is not our performance, not any firm's performance, and not a projection. Prop firm rules described here are commonly structured that way, vary by account type and platform, change over time, and must be confirmed on the firm's own rules page before you trade.

Take the closed trades that finished positive, divide by all closed trades, and you have the simplest line on the panel. It answers one question, how often, and it is silent on how much being right paid or how much being wrong cost.

Invented worked example, for mechanics only. Account A closes 100 trades. Eighty win an average of 50 dollars, so gross profit is 4,000 dollars. Twenty lose an average of 300 dollars, so gross loss is 6,000 dollars. Win rate is 80 percent and the account is down 2,000 dollars. Account B closes 100 trades too. Forty win an average of 250 dollars, gross profit 10,000. Sixty lose an average of 100 dollars, gross loss 6,000. Win rate is 40 percent and the account is up 4,000 dollars.

Account A is not unlucky, it is engineered. A high win rate is the easiest thing on this list to manufacture: widen the stop, or drop it, and most trades come back eventually. The invoice arrives as a small number of very large losers, which is exactly the shape that ends an evaluation. The size of your average loser is a sizing decision before it is a strategy outcome, so if you have not fixed [how much you risk per trade](/blog/position-sizing-prop-firm-challenge) as a percentage of the account, decided before the entry rather than during it, win rate is measuring your patience, not your edge.

Profit factor is gross profit divided by gross loss over the same set of closed trades. In the invented example, account A is 4,000 over 6,000, a profit factor of 0.67. Account B is 10,000 over 6,000, a profit factor of 1.67. One ratio, and the ranking flips.

How to read it: below 1.00 the record lost money, 1.00 is breakeven before costs, and 1.30 means the record produced 1.30 dollars of gross profit for every 1.00 dollar of gross loss. It is more honest than win rate because every dollar won and every dollar lost lands in the same ratio, no matter how the trades were spread.

It has two blind spots. First, profit factor is order blind: reshuffle the same trades and the ratio is identical while the drawdown is not, so it can never tell you whether an account survived. Second, one outlier can carry it. If a single trade supplies half the gross profit, the number describes that trade rather than the method, so ask what the ratio looks like with the largest win removed. Very clean ratios on short records are also the classic reason [tester results and live results diverge](/blog/backtest-vs-live-ea-results).

Nobody argues about the formula here. Drawdown means the largest fall from a peak to a later low, and the disagreement is always about which two things you measured.

Balance drawdown counts closed trades only. Equity drawdown counts closed trades plus whatever is floating on open positions. The gap between them is where accounts die. Invented example: a 100,000 dollar account reaches a peak balance of 105,000 and holds one open position floating at minus 3,000. Balance drawdown is zero. Equity sits at 102,000, so equity drawdown from the peak is roughly 2.9 percent. Your platform report may show the first number while a firm measuring intraday equity sees the second, live.

Three details decide whether the firm's number matches yours: whether the limit is measured on balance or on equity, whether the peak is fixed at the starting balance or moves up with new highs, and whether the daily limit resets at a server hour that is not your local midnight. These are commonly structured differently between static and trailing accounts, they vary by platform, and they change. Confirm the exact wording on your firm's own rules page before you size a single trade. If the two limits still blur together, start with [how daily loss and max drawdown work differently](/blog/daily-loss-vs-max-drawdown), because one of them ends your day and the other ends your account.

Expectancy is the average result of one trade, and it is the only figure that holds win rate, average win and average loss at the same time. Written out: expectancy equals win rate times average win, minus loss rate times average loss.

Same invented accounts. Account A: 0.80 times 50 dollars is 40, minus 0.20 times 300 dollars, which is 60. Expectancy is minus 20 dollars per trade. Account B: 0.40 times 250 dollars is 100, minus 0.60 times 100 dollars, which is 60. Expectancy is plus 40 dollars per trade. Account A is not waiting for luck to turn. It is paying 20 dollars for the privilege of pressing the button, one hundred times in a row.

Expectancy in R travels better. Express every result as a multiple of the risk accepted at entry, so a trade that made twice its stop is plus 2R and a full stop out is minus 1R. Expectancy then reads as R per trade, independent of account size, which lets you compare a 10,000 dollar evaluation against a 100,000 dollar funded account without rescaling anything, and is why [the account size you buy](/blog/ftmo-account-sizes-which-to-buy) moves the dollars far more than it moves the test. Expectancy per trade multiplied by trades per month is the sober version of the question everyone wants answered, and it belongs in [a written risk plan](/blog/prop-firm-risk-management-plan) rather than in your head.

Every metric here is an estimate, and trade count is the confidence interval nobody prints on the dashboard. Twelve trades tell you close to nothing. A profit factor of 3.0 across 18 trades is a sentence about those 18 trades, not a property of a system, and a single different outcome can move it far enough to change your conclusion.

There is no magic threshold and anyone who hands you one is selling something. The direction is what matters: the fewer the trades, the more of the result is noise, and the more concentrated the profit sits in a handful of positions, the less the averages describe anything repeatable. So ask two things of any record. How many closed trades, and across what calendar period and what conditions. Forty trades inside one trending week and forty trades across four months are not the same evidence.

This is also the quiet mechanism behind accounts that pass and then fail. A short evaluation can be cleared on a sample that never met a hostile regime, and the funded account meets one in month two. The metrics did not degrade, the sample simply got bigger, which is one of the recurring patterns in [why funded traders lose the account](/blog/why-funded-traders-lose-the-account).

Someone sends a screenshot. A vendor, a signal channel, a stranger in a group chat. You do not need suspicion, you need six questions in this order.

Two answers end the conversation early. If the trade count is small, the metrics are opinions wearing decimals. If the person cannot separate gross profit from gross loss, they are showing you a curve, not a record. The same checklist makes vendor pages readable when you are [comparing automated tools for prop accounts](/blog/best-ai-trading-bot-prop-firms-2026), because a sales page quotes the metric that flatters the product and leaves the rest for you to request.

During an evaluation, drawdown outranks everything. Nobody pays you for expectancy in that phase, you are being asked to stay inside two limits long enough to reach a target, so max drawdown and daily loss are pass or fail while profit factor is merely encouraging. Once the account is funded the ranking shifts: expectancy per trade and trade count start to carry the weight, because the account now only has to survive long enough for a positive expectancy to turn into a withdrawal. That shift is what [the timeline to a first payout](/blog/first-payout-timeline) is really measuring.

One note on tooling. Measuring and enforcing are different jobs, which is why a trader can read the dashboard carefully and break a limit anyway. Enforcement is mechanical: PraxAI GUARD applies the limits you define, in code, closing a position when your limit is reached and blocking an entry that would breach your ceiling. It is not a forecast and it is not intelligence, it is your own rule executed without your permission at the moment you would rather ignore it. Disclosure, since this article is about auditing other people's numbers: this blog is published by PraxAI, and we sell software, not challenges.