The Esports Analyst Club

LoL: Player Scouting and Statistics

The flaws and remedies of using statistics to scout new players

Published on: 12th March, 2022

There are no silver bullet statistics in esports. There isn’t a number that one could point to and say “see, I have proven the point”. This is particularly true for League of Legends and even more so when trying to determine the value of a player.

Let’s take an example: imagine you were playing mid-lane in the LEC, you shove your wave and roam. You force the opponent top laner to flash and you head back to base. Shortly after, your jungler returns to finish the job and secures the kill and a few turret plates.

Most will know that what you did was valuable. Yet, look at game summaries and you won’t find a single statistic that will capture the impact of a roam like this.

By most of the conventional means of measurement you have made a losing play. You gain no experience, no gold, no tower plates. Not a kill, not even an assist. In fact, in the time it took to get up there, you lost resources in your own lane and now find yourself behind the curve.

This issue on the face of it may not seem important, but how many times have you heard the analyst desk say something like:

“He’s a middle of the pack player; only 5th for GD@10 in the league”

This sets the narrative.

Without realising it, the opinion of viewers and professionals alike is being shaped by these numbers and these numbers miss so much of the wider picture!

So, what then is the solution?

The first and most obvious: the eye test. If you know the game well, you can simply observe and let intuition guide you. This is what most analysts are paid to do. They watch the games, they make notes and they form their own opinions based on what they are witnessing first-hand.

If you are deciding which of the two new players you want to fill your LEC teams Support role, watch as many games of each as you can and that will provide considerably more value than looking at any amount of statistics.

However, this assumes you’re down to two.

Let’s do a thought exercise. What if you’ve just set-up a team to play in your regions local league and there’s 100 top-laners asking to join. How would you pick one, or at least narrow the list down?

Solo Queue analysis can of course offer a lot of value. As the ranked teams you play with will for the most part be different every game, once you’ve played enough games your impact will be seen through your win rate and thus your rank.

You could have a KDA of 0.9, if you are in Challenger then you are doing something right (prime example here). It doesn’t matter that you can’t capture the “something right” into a statistic, simply winning proves the point!

So, let’s delete all the sub-Diamond elo-players from the list and say we’re down to 20 or so players.

Well, now what? It’s not feasible for a small team to employ someone to watch a mountain of VODs for each player. Even if they had the time, how could they possibly keep track of everything they’re seeing. We need the statistics back!

To do this, we have to be smarter. To do this, we need to come up with a new set of statistics, including but not exclusively the common values like Gold Difference or KDA.

For instance, often the limitation of a high-ranked player is that they are a one-trick. They only know how to Shaco jungle and everything else they do is reportable. This isn’t any good for competitive play as it would just get banned every game.

So, what we want is to create a custom statistic that measures the size of a players Champion pool. We could look at how many Champions they’ve played in their last 50 games. Or, how many they’ve had more than 10 games with and better than a 50% WR. Or, any other way of measuring this you could think of!

We rank our remaining players by this new metric, have a little dig and decide to shave off the bottom 5 for being too much of an OTP for the team.

Is this a fool-proof system? Of course not! Every metric has flaws. However, often these flaws can be improved upon, little by little. That’s really all it comes down to:

  • Translate your understanding of the game into a basic set of metrics (i.e. Ranked Rating)
  • Use these to rank the players
  • Identify their weaknesses (i.e. High-Elo OTPs)
  • Determine a slightly better metric that includes the weakness (i.e. Champion Pool Size)
  • Iterate

Thanks for Reading!

You got to the end of the article! My name is Jack J and I’m a professional Data Scientist applying AI to competitive gaming & esports. I’m the founder of AI & Esports start-up iTero.GG and the analytics site You can follow me on Twitter, join the iTero Discord or drop me an e-mail at See you at the next one.

More Articles

The Esports Analyst Club

Want to get these articles emailed to you once a week?

No spam, no paid promotions, no bullshit - just one article every week.

Sign-up Here