Profit factor compares the sum of winning trade outcomes with the absolute sum of losing outcomes. Payoff ratio compares the average winner with the average loss magnitude. In a crypto trade report, both require the same trade grouping, currency and cost convention. Their difference comes from how many winners and losers contributed to those amounts; neither ratio describes open exposure or the order of losses.
Payoff ratio below one and profit factor above one can be consistent: frequent small winners can outweigh fewer larger losses. Reconstruct what the report counted.
Totals and averages answer different questions
For this article, let G be the sum of positive outcomes, L the absolute sum of negative outcomes, W the number of positive outcomes and D the number of negative outcomes. All outcomes use one reporting currency. Zero outcomes belong to neither group.
| Metric | Calculation | Question answered |
|---|---|---|
| Profit factor | G / L | How many units of positive outcome accompanied each unit of loss? |
| Average winner | G / W | How large was the mean positive outcome? |
| Average loss magnitude | L / D | How large was the mean negative outcome in absolute terms? |
| Payoff ratio | (G / W) / (L / D) | How large was the average winner relative to the average loser? |
TradingView reports average winner divided by average loser separately from gross profit divided by gross loss. TradingView's strategy-information example
NinjaTrader documents these ratios separately, with display-unit and classification conventions. Inspect the underlying amounts and counts; here, loss magnitudes are positive denominators. NinjaTrader's trade-performance definitions
“Gross profit” and “gross loss” in the ratio describe the separate positive and negative pools. They do not, by themselves, tell you whether the underlying trade outcomes include fees. Label that choice explicitly.
One ledger, two ways to count the exits
This original hypothetical ledger contains six completed crypto trading campaigns. A campaign runs from opening a position until it returns to flat. Every amount is in USDT and already includes the costs assigned to it. There are no open positions or unallocated costs. These are invented accounting values, not observed trades or KLINEO results.
Campaign A has six partial exits, each contributing +10 USDT after allocated costs. Every other campaign has one exit. Changing the report's grouping does not change any fill or cash flow.
| Campaign | Net result of complete campaign | Exit-leg results in the same records |
|---|---|---|
| A | +60 USDT | Six legs of +10 USDT |
| B | +30 USDT | One leg of +30 USDT |
| C | +10 USDT | One leg of +10 USDT |
| D | −20 USDT | One leg of −20 USDT |
| E | −30 USDT | One leg of −30 USDT |
| F | −10 USDT | One leg of −10 USDT |
Both views have positive outcomes totaling 100 USDT and loss magnitudes totaling 60 USDT. Both therefore show profit factor 100 / 60 = 1.6667, rounded, and a total net result of +40 USDT. Their averages differ:
| Reporting unit | Winners | Losers | Average winner | Average loss magnitude | Payoff ratio |
|---|---|---|---|---|---|
| Complete campaign | 3 | 3 | 33.3333 USDT | 20 USDT | 1.6667 |
| Individual exit leg | 8 | 3 | 12.50 USDT | 20 USDT | 0.625 |
Splitting the +60 result into six positive observations reduced the mean winner and increased the winner count. The payoff ratio falls below one through a reporting change; no cash result changed.
That distinction matters when comparing a strategy with frequent partial take profits against one that exits each position once. Request both the fill-to-campaign mapping and the report's grouping rule. The completed-trade counting discussion develops the broader expectancy question without treating every exit as an independent decision.
Reconcile the ratios through the counts
When W and D are both positive, rearranging the definitions gives:
profit factor = payoff ratio × W / D
The campaign view gives 1.6667 × 3 / 3 ≈ 1.6667. The exit-leg view gives 0.625 × 8 / 3 = 1.6667, rounded. If a report does not reconcile, inspect filters, units, costs and grouping before blaming rounding.
Using winner and loser counts avoids a common breakeven error. With zero-outcome trades present, the losing fraction is not necessarily one minus the winning fraction when both use all trades as their denominator. The count identity still works because zero outcomes enter neither pool.
The observed count relationship is descriptive. It does not make eleven exit legs eleven independent experiments. For the uncertainty in a binary win fraction, the win-rate confidence-interval guide explains the assumptions behind the interval. That interval does not supply uncertainty bounds for the payoff amounts or profit factor.
Profit factor is not always immune to grouping
The earlier profit factor stays unchanged because the split pieces all have the same positive sign. Mixed-sign grouping can change the positive and negative pools themselves.
Suppose another hypothetical campaign has a +50 USDT exit and a −40 USDT exit. The exit view has profit factor 50 / 40 = 1.25. Group them as one +10 USDT campaign and there is no losing campaign: the denominator is zero. The net result remains +10 USDT. Do not generalize the first table's invariance to every regrouping.
Reclassify trades after applying costs
A separate invented example has three gross outcomes: +6, +4 and −5 USDT. Its positive pool is 10, its loss pool is 5, its profit factor is 2, and its payoff ratio is 5 / 5 = 1.
Assume a chosen total charge of 3 USDT per completed trade. Net outcomes become +3, +1 and −8. Profit factor becomes 4 / 8 = 0.5; payoff ratio becomes 2 / 8 = 0.25. Total net result is −4 USDT. These charges are illustrative, not a venue fee schedule.
If a charge instead pushes a small winner below zero, move that outcome into the losing group before recomputing every component. Do not combine gross winner counts with net sums. The transaction-cost reconciliation guide helps establish which fees and execution effects are already included, preventing a second deduction of the same cost.
Keep the undefined cases visible
With no losing outcomes, profit factor has a zero denominator. Report “no observed losses” with the sample count rather than presenting infinity as proof of safety. Payoff ratio is also undefined because an average loser is unavailable.
With losses but no winners, profit factor is zero, while payoff ratio is undefined because there is no average winner. If every outcome is zero, neither ratio is defined. A missing value should retain its reason rather than becoming an arbitrary large number or a favorable score.
Use unrounded ledger amounts for classification. A tiny loss displayed as 0.00 is still negative in the underlying data; rounding before grouping can silently remove it.
A worksheet for comparing two reports
| Check | Record before comparing | What a mismatch means |
|---|---|---|
| Observation unit | Campaign, position, round trip or exit leg | Conditional averages may refer to different events |
| Included outcomes | Identical start/end rules and closed records | The samples cover different activity |
| Currency and scale | USDT, another currency, percentage or risk units | The weights applied to outcomes differ |
| Costs | Included charges and allocation rule | Signs, pools and counts may change |
| Components | G, L, W and D before rounding | The ratio cannot be reproduced from its inputs |
| Remaining risk | Open positions and marked equity | Closed outcomes omit exposure still held |
A ratio computed from dollar PnL need not match one computed from each trade's percentage return when position sizes differ. Multiplying every outcome by the same positive constant leaves both ratios unchanged; changing only selected trade sizes can change them. State which weighting answers your question.
Neither ratio records loss order, holding time or capital committed. Two ledgers with identical outcomes can have different drawdown paths. Pair the completed-trade report with marked equity; the Ulcer Index guide measures squared drawdown observations and explains why their sampling frequency matters. Reconcile the realized and unrealized PnL alongside those closed records so an unresolved position remains visible. A high historical ratio is not a future loss limit or a universal pass threshold.
Method and limits
This AI-assisted article uses primary platform documentation reviewed on 23 October 2026 and original hypothetical ledgers. The examples isolate aggregation, conditional averages and reporting conventions. No exchange account, customer result or KLINEO performance was analyzed. Use the worksheet to make reports reproducible before interpreting their investment significance.