I feel like a real problem is that nobody actually knows how accurate league stats are

HateDaddy··18 votes·9,227 views

I’m going to be honest, I really don’t trust any “stat” source that isn’t either:

Directly from the main source (Riot)

Or endorsed directly by the main source.

Furthermore, just from what I’ve observed you see some pretty significant differences between different sites. This alone tells you that these sites either collect and analyze them differently, have different levels of access to stats, have different algorithms for collecting them, or are just outright wrong. No matter how you slice it, that’s a pretty big deal.

My real question is this:

How come you just don’t make stats public yourself riot? There probably are some pretty good sites out there that are pretty close, and that’s all well and good. However, I think I’d want people to have an accurate measurement of what exactly is happening and avoid people referencing some junky site that is horribly put together because the site supports their dogmatic beliefs. And I point this out because it’s not uncommon for people to literally argue in a thread about who’s stats are more correct. That’s a red flag right there.

Seems the reasonable thing to do would just be to spread accurate information as opposed to misinformation.

21 Comments

FOR JUSTICE10 votes

those sites are already drawing information riot is providing though. all of it is drawn from Riots API, which even you yourself if you felt so inclined could also view, provided you have the technical know how.

what makes them display "different" data is that they update their info at different times. some do it every 12 hours, others 24 hours, etc. but even then the data is drawn from the same source, so its never actually "wrong", just outdated by a few hours.

thus they are in fact, perfectly accurate, because riot themselves are the one providing access.

HateDaddy1 votes

[{quoted}](name=FOR JUSTICE,realm=NA,application-id=3ErqAdtq,discussion-id=mlpAifvz,comment-id=0000,timestamp=2019-08-04T00:42:31.098+0000)

those sites are already drawing information riot is providing though. all of it is drawn from Riots API, which even you yourself if you felt so inclined could also view, provided you have the technical know how.

what makes them display "different" data is that they update their info at different times. some do it every 12 hours, others 24 hours, etc.

Right but I’ve seen some pretty major differences well into a patch cycle, which shouldn’t be happening since by the end of a patch there’s not going to be much data that comes in that’ll change a champions stats that significantly between the span of a few hours. So this comes down to how sites present stats, which tells you all you need to know there since you shouldn’t be seeing significant deviation (keep in mind, I’m also referring to looking at stats comparing same demographics).

FOR JUSTICE1 votes

in a perfect world they would of course all be the same but they physically cant actually. sites are only allowed a specific ratelimit for requests, so it most probably ends up extraneous some times because of this.

this is done to stop the api from being overloaded by a bunch of requests at once.

HateDaddy1 votes

Well even then, that seems to suggest to me that the ideal course of action is to severely limit the amount of sites access or do their own stats. What we end up having are sites displaying some pretty significantly different results.

FOR JUSTICE1 votes

i guess thats true. for the most part its just best to use your better judgement by using more then one site.

ex if vayne has a 50% winrate in one site, but 52%, 53%, and 56% in others you can probably guess shes actually around a 54%

HateDaddy1 votes

[{quoted}](name=FOR JUSTICE,realm=NA,application-id=3ErqAdtq,discussion-id=mlpAifvz,comment-id=00000000000000000000,timestamp=2019-08-04T01:29:31.951+0000)

i guess thats true. for the most part its just best to use your better judgement by using more then one site.

ex if vayne has a 50% winrate in one site, but 52%, 53%, and 56% in others you can probably guess shes actually around a 54%

Sure, but even then that’s a crapshoot. It would be better just to have one standardized site (even if it wasn’t riots necessarily). Stats dumping is a pain, I get it. But so is trying to browse 3-4 different sites and listen to everyone tell you their site is right when there’s little evidence to actually support it.

DuskDaUmbreon1 votes

Some, like op.gg, update different parts of their site at different times. Sometimes they update from other patches entirely.

It's the only explanation for showing a 4% winrate difference for the exact same stat for the exact same champion in the exact same region at the exact same rank, just in a different tab.

EvansR1 votes

Another thing to mention, at least for op.gg, is the champion stats are based on Plat+ games and taken from the korean server because "there isn't enough data from other regions". Other sites might take all ranks and regions into account, haven't looked at the small text on other sites.

preternatural4 votes

hashinshin said that, but no one likes hashinshin, so it's invalid. all the stats sites are correct. welcome to the boards.

IainG101 votes

I hate Hashinshin with a passion, but that doesn't mean he can't be right sometimes. I mean, a broken clock is right twice a day, right....

WoonStruck3 votes

LOL YOU THINK RIOT BELIEVES IN ACTUAL TRANSPARENCY.

How many years have you been following this game now?

C'mon, I know you know better than that.

GigglesO2 votes

Stats don't even matter at this point. Its how shitty it feels to play against auto burst and oppressive shit.

If you don't support that statment, thats how we got all the defense in the game removed. Tanks are too opressive Visage on adc's is too opressive Randuins on bruisers is too opressive atmogs is too opressive warmogs is too opressive fon gives too much mr

Well now its time to remove/reduce damage.

Eedat1 votes

I normally compare lolalytics, u.gg, and leagueoffraphs to get a decent idea. Champion.gg used to be a decent one but they made some changes a few years ago and now they're always really off and their graphs and wrong to the point of literally being impossible. Op.gg only uses KR stats so I stay away from that too

Ragnaveil1 votes

U.gg is the only site with accurate statistics. I mean why else would they be the only endorsed site rn?

Eedat3 votes

Because they paid money?

Modl Ryden l1 votes

Riot does publish official stats through their API as well as the client, where you can compare your performance on X champ to those in other ranks as well as players in the LCS or other LoL affiliated eSports scene.

Aatrox Airlines1 votes

Riot stated somewhere (cant remember where rn) that u.gg has the stats that are the closest match to their own

HateDaddy2 votes

[{quoted}](name=Aatrox Airlines,realm=EUNE,application-id=3ErqAdtq,discussion-id=mlpAifvz,comment-id=0006,timestamp=2019-08-04T09:26:19.617+0000)

Riot stated somewhere (cant remember where rn) that u.gg has the stats that are the closest match to their own

Literally everyone says this about the sites they reference. It’s not that I don’t trust you, but I hear this all the time. I’d like to see some proof of this.

Daddy Ants1 votes

Riot themselves have said u.gg is closest in terms of accuracy.

haaaaaaalp1 votes

Riot already releases all of the stats through the api. The problem is that everyone is looking for a different subset of stats. One person wants to see stats for korea diamond in the past 48 hours, another wants to see global diamond for the entire patch, another wants to see every plat player over the past week. The only way for riot to "publish" all of the data so that it meets what everyone wants to see is by releasing all of the raw data into the api and letting people sort the data the way they feel is appropriate.

The variations you see between each website is simply due to them reporting the data in a different way. All of the major websites are accurate as far as the data they display. They just all choose a different subset of data to present and there is no consensus on what the correct subset should be.

SòlKıng Rıven1 votes

Yeah those sites use the Riot API to retrieve stats from what I understand. (correct me if im wrong.) The main issue with statistics in my opinion is that there are many different ways to interpret them. Riot for example buffs and nerfs champs based on winrates. But winrates are different for each elo and what might be op in one elo is troll for another. For example ive heard that blitzcrank is pick/ban in bronze. In high plat its a troll pick since he has no lane pressure. Another issue is just because a champion has a good win rate doesn't mean the champion is broken. It could literally mean that the people who are playing it are on average more skilled then the average of people playing other champs. The opposite is true as well, I know a few champs that have low winrates that are absolutely disgusting to play against and if you would ask me while I wasn't looking at a chart I would say they have over a 52% win rate, in reality they might only have a 49% win rate. Theres so many other variables as well, like new champs for example usually have terrible stats at the beginning of the season so its hard to tell if they're good or bad, because everyone is just learning to play them.

More from Gameplay

More by HateDaddy