LoL: The Champion Tier List Challenge
Last week I made a bold claim.
I said: “you CAN use Solo Queue data in professional play, you just have to be smart about it!”. In fact, I wrote a whole article explaining how I have landed on this conclusion.
This was met with a range of feedback, a majority of which sat somewhere between “OK, a small amount of solo queue data can be used with a lot of effort” to “Solo queue data is underrated!” — with a handful sitting on the extremes.
What I want to do now is try something a bit riskier. I want to try and “prove” the point. Or at least, add more weight to my argument. So, as I alluded to at the end of last weeks article: I’m going to predict which Champions I think will be strongest in professional play this patch.
In fact, I’m going to do it every patch for the remainder of the Summer season. Then, after every patch concludes I will compare my tier list to the actual results.
There should be no doubt that there will be patches where I get it all wrong. There will also be patches where I look like a prophet. However, all we care about is whether by the end of the season there is a broad trend of “higher tier = better performance”.
Now, for those who subscribe to the mailing list (or check Reddit on a Thursday) I don’t want to fill your inboxes with tier lists and analysis every week, and so everything about this challenge will be kept on iTero.GG, with announcements on Twitter.
Here’s the current live tier list.
If you’re reading this sentence, it’s because it’s my first week making predictions and so there are no results to compare them to yet. This shall be replaced with a link to the results screen in 2 weeks time!
A Deeper Explanation
There’s a few points that are essential to clarify. As an aide, here’s the first Tier list I made for 12.11:
Pro Tier List, Patch 12.11. Assets by Riot Games.
Why a Tier list?
Statisticians amongst you will have wanted a more concrete value to these predictions. Why say Rakan is S tier instead of predicting his win rate?
It’s because I’m not actually predicting Champion win rates.
Above all, the thing that has by far the biggest impact on the outcome are the players. When G2 Esports plays Astralis or Team Liquid plays Immortals, the draft is only a small proportion of what decides the outcome of that game.
If G2 only played Aatrox, Graves & Yuumi whilst Astralis only played Rakan, Volibear and Tristana then the respective win rates is more determined by the players piloting them then the relative power of the Champions.
So, what I am trying to predict is “how much did this Champion impact this teams chances of winning the game”.
Let me make this very clear with an example.
Imagine TL & Immortals played each other 100 times.
Immortals ONLY played Tristana as ADC
Immortals won 30 of their games
In this scenario, Tristana has a 30% win rate. The real question is, how many games would they have won without Tristana. If the answer is, let’s say 15%, then actually Tristana has won them twice as many games!
This is why I don’t predict win rates.
The alternative value (factor increase in percentage of games won against baseline forecast) is too confusing to provide to a majority of users. So, I just give a tier. Each defined as so:
S — Significantly improves your teams chances of winning the game
A — Improves your teams chances of winning the game
B — Has little effect on the outcome
C — Reduces your teams chances of winning the game
DON’T — Significantly reduces your teams chances of winning the game
Why not more Champions?
Technically, I have a prediction for every single Champion in the game. They just come with different levels of confidence. The Champions I have provided are the ones in my top confidence bucket. As we move through the patches, more Champions will pass this threshold and get included.
The confidence is based on a number of things, but mostly it’s how much data do we have in both solo queue and pro play.
If there’s a Champion that’s only been played 100 times in high-elo solo queue this patch — it’s not going in the list.
If there’s a Champion with 10,000 SQ games but has only had 10 stage games since it was released — it’s not going in the list.
X Champion is S-Tier? But I’d just counter with Y!
These tier lists are made in isolation to the Champion. There are certain Champions that just “fit” better with the meta, which means they either counter or pair well with other popular meta picks.
For instance, Yuumi is “DON’T” Tier for 12.11. If, however, there was a team composition where she paired extremely well with the ADC & Jungler (let’s say Twitch/Talon we’re in meta) and countered the enemy composition well (Ashe support anyone?). Then, she rises quickly out of her current doom tier.
Similarly, certain Champions have obvious hard lane counters. If LeBlanc happened to be an S tier Champion, she certainly wouldn’t be if Lissandra is left open.
The tier lists are here to define a Champions baseline. That means, on average, an S-tier Champion should outperform a C-tier one. That doesn’t mean an S is always an S once all the cards are on the table.
With that, the gauntlet is thrown and the challenge begins. The first tier list is already online at itero.gg.
Keep an eye on the site or my Twitter to see how it goes throughout the season. The next time I write about this through the Esports Analyst Club will be to wrap up the results!