What the numbers on a scouting page actually mean
Every scouting page shows you the same handful of numbers: CS per minute, gold per minute, KDA, kill participation, and a win rate. Each one measures something real, and none of them means anything until you know what role the player was in and what rank the lobby was.
That second half is the part nobody tells you, and it is why two people can look at the same profile and reach opposite conclusions.
The five numbers
| Number | What it measures | What it is not |
|---|---|---|
| CS per minute | How many minions and camps they take per minute of game time | A measure of skill for a support, who is not supposed to take any |
| Gold per minute | Total income per minute, from farm, kills, objectives and passive gold | Interchangeable with CS, and it is much harder to fake |
| KDA | Kills plus assists, divided by deaths | Evidence of impact. It rewards not dying, which is not the same thing |
| Kill participation | The share of their team's kills they were present for | Meaningful in a very short or very long game |
| Win rate | Games won over games played | Anything at all under about 30 games |
Why a number without a role is noise
Roles do different jobs, so the same number is a different achievement depending on who produced it.
The baselines we use inside LaneDiff make the point. A support at roughly 270 gold per minute is doing the same job as a laner at roughly 400, because a support's income comes from a support item and assists rather than from a wave they are not allowed to take. Judge the support against the laner's number and every support in the game looks terrible. Judge the laner against the support's number and every laner looks like a smurf.
CS per minute goes further: for supports it is not a weak signal, it is not a signal. Our scoring skips it entirely for them, because the number measures whether they made a mistake rather than whether they played well.
Why a number without a rank is also noise
Six CS per minute is a poor Diamond laner and a very good Silver one. The metric did not change, the field did.
This is the reason a scouting read should be relative. An absolute bar, "call it elite above 8.5 CS per minute", flags most of high elo and misses the low elo smurf who wins without farming at all, because they are killing their lane opponent instead of last-hitting. Comparing to the lobby catches both.
So the two comparisons that turn a number into information are: against the same role, and against the rank actually in the game. Everything else is a number on a screen.
What "good" looks like
You will notice this post has not printed a table of target numbers, and that is deliberate. A single number for "good CS per minute" is wrong for four of the five roles and wrong for every rank except the one it was written for, and it goes stale every time the game changes how gold works.
What is worth internalising instead:
- Gold per minute is the honest one. It is hard to inflate without actually doing something, and it works across every role once you know the role.
- CS per minute is a laning metric, most useful in the first fifteen minutes and increasingly meaningless after that.
- Kill participation reads intent. Low kill participation on a jungler or a support says more than any of their other numbers.
- KDA is the most over-read number in the game. A 9/1/14 line in a game their team won at minute 20 tells you they were not the problem. It does not tell you they were the reason.
Three numbers people over-read
Champion mastery points. They measure hours spent, not current skill, and they never go down. A player with two million points on a champion may not have touched it in a year.
Summoner level. Genuinely useful for one thing, which is spotting a new account, and useless for everything else. It measures games played, not games played well.
A win rate on a small sample. Ten games at 70% is what ten games look like. The number starts saying something around thirty, and starts saying something loudly around fifty.
Where we fit
These are the numbers LaneDiff reads for all ten players once a game starts, and the whole point of the product is the comparison rather than the numbers: every player is measured against the baseline for the role they actually played, scaled to the rank of the lobby they are in, so a support is never judged against a mid laner's income.
If you want to see what those comparisons are used for, the most common question they answer is whether the person who just beat you was smurfing.