Fortnite Stats Organiser

Lookup needs stats to be public on the account — Fortnite turns that on under Settings, Account and Privacy, “Show on Career Leaderboard”. Everything below also works if you type the numbers in from your own career screen. Names are looked up in batches, one account at a time, because the stats service refuses requests that arrive together.

The accounts

Lifetime totals are the ones on your career screen. Fill in the current season too and the page ranks the two separately; leave it blank and only the overall table appears.

This season, against the rest of the career

The season figures are the big ones. The career line beneath each is the rest of the career — everything except this season — because a lifetime total includes the season inside it, and comparing a season against a number it is already part of understates the difference.

Ranked — current season

Ranked — overall

The skill score is the average of two standardised figures — elims per match and win rate — measured against the others in this list, so it says who is ahead of this group rather than putting a number on anybody in the abstract. Zero is the middle of the group. Why those two and not K/D is the panel below.

Which gaps are real

Ranking on “win rate plus K/D” counts winning twice

It is the obvious composite, and it double-counts, because K/D already contains the win rate. Deaths are matches minus wins, so

K/D = (elims per match) ÷ (1 − win rate)

exactly — not approximately, and checked across 200,000 random careers rather than asserted. K/D is therefore fully determined by elims per match and win rate. Adding win rate to it weighs winning twice and fighting once, which is a choice nobody makes deliberately. So the ranking uses the two figures that genuinely vary on their own, and shows K/D beside them as the combination of the pair that it is.

It changes the answer. Four players, ranked both ways:

Fighter tops the K/D table on twice everyone else's elims per match, which carries the ratio even though their win rate is the second worst here. Note that this is not K/D flattering a low win rate — the denominator is matches minus wins, so winning more raises K/D, and Fighter leads in spite of that rather than because of it. The ranking puts them third because it weighs closing a game out equally with fighting, and on that half they are nearly last. Which table is right is a question about what you value, not about arithmetic, so both are here.

One consequence of standardising within the group: with exactly two accounts the two scores can only be equal and opposite, so two players with mirrored strengths tie at zero. That is reported as a tie rather than broken arbitrarily, because the numbers do not contain a winner.

Season history

Nowhere publishes a season-by-season history, so this builds one the only way anybody can: it keeps a dated snapshot of the lifetime totals each time you save, and the gap between two snapshots is exactly what happened in between. Saved in this browser only.

A death is a match you did not win

Battle royale reports matches, wins and eliminations, and never reports deaths — because it does not have to. You die once in every match you do not win. So deaths = matches − wins, and the two figures everyone quotes are K/D = elims ÷ (matches − wins) and win rate = wins ÷ matches.

That has a consequence worth sitting with: a win does not merely add to your win rate, it removes a death. The two headline numbers are not independent measurements of two different skills. Winning more raises your K/D even if you never land another elimination, which is why seeing both move together tells you far less than it appears to.

A season is a difference, not a reading

The commonest mistake in reading these numbers is subtracting two K/Ds. K/D is a ratio, and the difference of two ratios is not the ratio of the differences. Take a career sitting on 1.20 after 4,400 matches, then a season of 220 matches played genuinely well:

The career figure moved by 0.0775. The season was played at 3.00 — nearly 39 times the movement it caused. Nothing is wrong with either number. A lifetime K/D is a death-weighted average, and this season is only 4.3% of the deaths in that career, so it moves the headline by about that share of the distance. Read the movement and you will conclude nothing happened.

Where a change actually came from

Since K/D is elims over deaths, a change in it has exactly two sources: the elims moved, or the deaths did. Splitting a change between two causes normally depends on which you credit first, so this credits each with the average of both orders — and the two parts still add up to the whole change exactly. Here are two seasons of the same length:

The second season has fewer eliminations than the first — 430 against 450 — and a higher K/D. That is not a contradiction, and the split says exactly where it came from:

One part negative, one positive, and the whole gain came from winning rather than from shooting. A single K/D figure cannot tell those apart, because it is one number holding two stories — and the story it usually gets told as is the one about aim.

You can improve at every mode and get worse overall

This is the one that looks like an error and is not. Solo, duos and squads have different K/Ds, so your career number depends on the mix as well as the play. Here is a player who got better at both modes and moved most of their play from the softer one to the harder:

Every mode up, the total down by more than a point. This is Simpson's paradox, and it is not a trick of contrived numbers — it happens whenever the mix shifts towards a mode where the figure is lower. Before: 200 solo against 1000 squads. After: 1000 against 200. The practical consequence is blunt — a single career K/D cannot be compared between two players who play different mixes, and most people play different mixes.

The same number is not the same evidence

Two accounts, identical K/D to four decimal places, wildly different amounts known about them:

Same headline, and the shorter career's figure is 5.5 times less certain. Treating eliminations as arriving at some rate per death, a K/D over D deaths carries a standard error of about √(K/D ÷ D), so precision improves with the square root of how much you have played. Two accounts are called apart above only when the gap exceeds 2 standard errors of the difference — which for these two would need 0.17, and they differ by 0.00.

One honest caveat about that interval: it is a floor on the uncertainty, not the whole of it. Elims per match are streaky — ten in one game, none in the next — and that overdispersion widens the real interval beyond what this model allows. So where the page says a gap is noise, it is noise; where it says a gap is real, read that as the optimistic case.