Explanation: The formula to the ranking system

Lost R··7 votes·2,229 views

At some point, the community apparently deluded itself into believing some seriously stupid shit about the ranking system. Next time Riot makes any major feature changes like this, they should pin it and make it as clickbaity as possible so people will actually read up on these things and not spiral off into mythology with their delusions. Crates don't factor into what rank you get.

Here's how the system works. It works off of a percentile system, directly comparing you with other people who played that champion in that role. If you play Yasuo mid like everyone else in the universe does, you're playing against everyone else's stats and results, and due to statistics, it becomes much harder to gain a higher rank. If you play Janna top, it's comparatively much easier because Janna top is extremely rare. Yes, the system does know the difference between the same champions in different roles.

With that out of the way, how does the formula work? Now, the EXACT specifics of this, I don't know, but it takes a ton of factors into account. I'm guessing it compiles it all into a lengthy algebraic formula and creates a final score that is compared to other scores, and your ranking is based on where you are in that percentile.

D-: 0-6% (basically AFK) D: 6-12% D+: 13-18% C-: 19-25% C: 26-30% C+: 31-37% B-: 38-44% B: 45-50% (this is the average, and where most players in an average game will wind up) B+: 51-66% A-: 67-72% A: 73-78% A+: 79-83% S-: 84-88% S: 89-94% S+: 95-100%

It's not entirely accurate; the margin of error is probably 1 or 2%, but this is how percentiles work. It's not like the grade system used in schools.

Now, what does the system factor into your ranking? This is where it gets hazy. Yes, it factors in your...

  • Role,
  • Kills
  • Deaths
  • Assists
  • Creep score

But it also factors in...

  • Ward count
  • Damage dealt
  • Damage taken (damage taken being factored as a positive thing. A lot of champions, like tanks, focus on being a damage sponge, and the ones besides are likely doing a good job by the system's logic if they can almost but not die)
  • Time elapsed
  • Turrets destroyed
  • Healing (again, champions like Soraka, Bard, Taric, Alistar, Janna and Sona whose gameplay largely focuses on their ability to heal their allies).

All of this is compiled into an algebraic formula that I don't know the specifics of, and then you get your final score that's translated into your rank. It probably modifies points here and there based on things you did or did not do. Someone who goes 15/0/8 in 20 minutes who didn't ward once or smash a turret probably won't get as high a score as someone who got 13/1/6 with the same champion in the same role and general time scale who did.

Likewise, it becomes harder to score a high rank with certain champions in certain roles if the average player is just that good in that role, doubly so if it's a large pool of players to draw from. And it doesn't matter if you're the support, the top laner, or Yasuo mid. The scoring formula is probably the exact same for everyone, and supports aren't going to have kills counted against them. That's what the community does, not the ranking system.

20 Comments

FACE THE WlND1 votes

Genius!

DeathBurst1 votes

There is one thing that has never been disclosed: you assume that the scoring formula is "probably the same" for everyone, but this has not been confirmed.

And IMHO, it shouldn't be the case. You already talked about kills and supports, so let's take a quick example. Of course, kills shouldn't be counted against a support, but it really should be possible to get S+ with an 0-2-23 Janna. On the other hand, it probably shouldn't be possible to get S with a 0-2-23 Jinx ADC, or Annie mid. So, as I said, the scoring formula shouldn't be the same for every position.

Lost R1 votes

Except I have gotten some form of S rank with Janna with scores worse and more mundane than that just yesterday, or I got an A+, AKA just short of an S rank. Again you're just thinking of the score solely in KDA terms and not factoring in EVERYTHING. Can you imagine that I've destroyed turrets as Janna? I've sustained my team in teamfights by ulting the center of them and using the healing effect and an extraordinary amount of AP to give them a second wind (PUN INTENDED), and I've dealt a reasonable amount of damage besides. I'm not afraid to kill minions or the enemy if it comes down to it, and that has given me a higher ranking than most people will ever see who have those inhibitions.

DeathBurst1 votes

That's not my point. Of course a positive factor is a positive, whether you be a support or a carry. I'm just sayin', the factors should be the same for everyone, the weights of the different factors shouldn't. Having kill as a support is good, but it's not important. Having kills as a carry is a necessity. On the opposite side, having wards as a carry is nice, but maybe it's less important than for the support. And it has nothing to do with KDA specifically...

Lost R1 votes

They probably are; as I said, you're being ranked based on the same champion in the same role and no one else, but I didn't design the system, so I can't say with absolute certainty that all weights are the same for every champion and role across the board. It would be baffling if they were.

DeathBurst1 votes

Well, not really baffling, no. Since you are only compared to other people on the same champ in the same pos, if all Janna sup are scoring very few kills, you don't need to score a lot of kills to get a high score. That is, on paper.

So a system using the same formula for everyone is probably "good enough" in a vast majority of cases, and (what's important) much simpler to implement. But I still think that it could be better, and I do have the feeling that the current system values too highly kills for supports. Just my personal feeling, though...

gorenzel1 votes

There's a D rank!?

Lost R1 votes

Yes, there is. I don't know if there's a D- rank, but it would stand to reason that there is. The one time I would have been able to see it, when someone DDOSed me right as the game started, I wasn't able to see the post-mortem, and Riot doesn't allow you to see rankings in the match history (it would probably breed more toxicity in the community than there already is).

Mysticman891 votes

Not that it matters at all to your point (which as far as I understand it, is accurate), but riot has confirmed that S+ is top 2% in a podcast, so slightly harder than top 5% as you guessed (but better than the challenger level play [i.e. top 0.01%] that some disillusioned folks sometimes think).

Lost R1 votes

Holy crap. I always suspected it to be just the top 5%, but 2%?!

Mysticman891 votes

Yeah you have to be pretty good to get an S+.

Still, if you go by all ranked players, then the top 2% of all players consists of high plat and above players, which isn't that exclusive considering how many people play. And considering that only a subset of those players will actively play any given champion in a given role, top 2% of a particular champion in a particular role doesn't seem that insane if you have a particularly good game. Which is how it should be really.

KakitaMike1 votes

I guess that helps me make sense of the game where i was shyvana top, went 6/4/1, had 304 cs, took 3 towers, and we lost, but I somehow pulled an S+ rank out of it.

Because I didn't even feel like i earned an S that game, let alone all the people annoyed by it in end game chat.

gatchi1 votes

Thanks for writing this out!! Knowing this really gives the system more of an impact for me.

Cloak and Sword1 votes

Support death don't lower your chances of scoring higher? I am rank six with Gnar and I really support with him more than I top. But Gnar is a fighter/tank. That is is actual role, but if I play as a support will my deaths still count?

Lost R1 votes

All scores across all champions and roles are weighed the same. So yes, support deaths still count against you.

Cloak and Sword1 votes

"The scoring formula is probably the exact same for everyone, and supports aren't going to have kills counted against them."

All scores are weighed evenly except the support role? Where the deaths do not count against you right?

Lost R1 votes

DEATHSCOUNTAGAINSTYOUEVERYWHERE

Darkslayer851 votes

I play Gnar jungle.

RUMBLER J1 votes

and what kind of algebraic formula does this signify?

BloodyTompon1 votes

One thing you missed (or I didn't see). It compares you to everybody else playing that champ in that role within roughly your MMR. This is to allow people in say Bronze V who aren't as good at doing some of these things like CS and warding to still achieve S ranks.

[{quoted}](name=Lost R,realm=NA,application-id=3ErqAdtq,discussion-id=16y6KbmQ,comment-id=,timestamp=2016-04-13T17:32:15.506+0000) I'm guessing it compiles it all into a lengthy algebraic formula and creates a final score that is compared to other scores, and your ranking is based on where you are in that percentile.

It's most likely the sum of normal distributions.

More from Gameplay

More by Lost R