Stop quoting Champion.GG right now, especially for 125+ game winrates.

Astôlfo·Sep 28, 2016·71 votes·6,069 views

There is literally no meaningful sample size at the moment for the vast majority of champions, and especially not for 125+ games.

As for the Illaoi threads for example who had quoted 68% and 69% earlier yesterday or such and a few hours ago respectively, it now shows 57%. All the other champions with absurd winrates (60-100%) because of equally meaningless sample sizes have also now all adjusted as well a bit.

It's still not settled yet, and it will most likely keep fluctuating for at least the next several days/few weeks. So don't bother making gimmick clickbait/repwhore threads when there's no meaningful data to back it up.

75 Comments

VerxintSep 28, 201619 votes

Ever since stat tracking began people have failed to understand small sample sizes.

Morde jungle was the highest winrate a few patches ago but the playrate was less than half a percent

AstôlfoSep 28, 201620 votes

Nah, we should just ignore sample sizes. Mordekaiser Mid had 100% winrate at 50-125 games several hours ago. Time to gut him, am I right?

kappa.

Amasa DelanoSep 28, 201622 votes

I briefly had a 100% winrate with Nami in ranked. Ergo, I am a god with Nami in ranked.

LauchmelderSep 28, 201616 votes

I had 100% winrate with Thresh for the first 8 games in this season. Thresh OP, Rito pls fix Kappa

The Yetii RiderSep 29, 20164 votes

[{quoted}](name=MadKnife IV,realm=EUW,application-id=3ErqAdtq,discussion-id=fERuBdOQ,comment-id=0000000000000000,timestamp=2016-09-28T21:24:09.566+0000)

I had 100% winrate with Thresh for the first 8 games in this season. Thresh OP, Rito pls fix Kappa

Your name...respect

PennsSep 29, 20161 votes

There is a big difference between around 700-1000 analyzed games and less than 10 games tho.

LauchmelderSep 29, 20161 votes

You missed the joke. Can't blame you though, Kappa doesn't transform into the emote outside of Twitch ;)

PennsSep 29, 20161 votes

I know you were joking, but you are joking about people quoting "low sample size" stats. Thats why I said theres a difference, 70% win rate in 1000 games says way more than 90% in 10 games.

Warlord ZudaneSep 29, 20163 votes

I currently have a 100% winrate with Alistar ranked. I must be Alistar god Ranked?

BlanchimontSep 30, 20161 votes

[{quoted}](name=Amasa Delano,realm=NA,application-id=3ErqAdtq,discussion-id=fERuBdOQ,comment-id=000000000000,timestamp=2016-09-28T20:50:59.705+0000)

I briefly had a 100% winrate with Nami in ranked. Ergo, I am a god with Nami in ranked.

If you have that for like even 20 games, that is already supper impressive!

Amasa DelanoSep 30, 20161 votes

You're close, only off by like a factor of ten.

Malicious MetalSep 28, 201615 votes

http://media.lolusercontent.com/api/embedly/1/image/resize?url=http%3A%2F%2Fwww2.pic-upload.de%2Fimg%2F31777228%2Fmordetriggered.jpg&key=a45e967db0914c7fb472fd4381e6c85b&width=425

GULAG 4 USep 28, 20163 votes

"counterplay for morde is...to not let him get in your face..."

"Counterplay according to the boards... Is to QQ until the said champion is trash"

LeetriSep 28, 20169 votes

Like people need stats to make clickbait/repwhore threads. :P

NiyumiSep 28, 20163 votes

Illaoi was broken before her buffs in the right hands. Now that she lost any weakness she once had, and had her strengths doubled in size, she's near impossible to beat if she's at all competent. They need to take away the % health damage on her W, and she'd be fine. But as it is, she's just absurd.

I don't know what other champions people are talking about. But, as someone who mained Illaoi for a month after her release, I can tell you she was fine back then, and is way too strong and forgiving now.

AstôlfoSep 28, 20168 votes

I don't think you know what you're talking about.

Not sure what weaknesses she supposedly lost. The only changes made were tentacles being visible in brush, changing W from % of AD to % health, and shortening Test debuff from 60 to 12 in return for mildly slightly faster spawns and closer spawns.

The W change made her less effective against squishies and more effective against health-builders. It's not a nerf or buff. It was a shift in focus.

Yes, a month, how incredibly conclusive.

It Hertz When IPSep 28, 20161 votes

I played a game with Kayle once. That makes me a main!

AstôlfoSep 28, 20161 votes

Hey, watch it. I'm a Mastery level Challenged 4 Ivern main, and I'll kick everyone's ass in Ranked with him.

It Hertz When IPSep 29, 20162 votes

"Challenged"

I lost it. :P

NiyumiSep 28, 20163 votes

Yeah, a month. Except uh, you know, in just this season alone, I have over 2500 games. So, a month for me is probably five times the game of most people. And I still play her, just not religiously like during that month.

And yeah, she lost her weakness of relying on tentacles for damage. With the addition of % health (max health, too) damage, she just smacks people with her W and out trades them regardless of whether or not her tentacles hit them. When she uses E, she doesn't need tentacles to break your spirit because of the same % health damage. And they also increased tentacle damage dramatically.

So yeah, keep on acting like you know what you're talking about. Illaoi's old counters are no longer counters (Darius, Rene), and she just wins lane unless you're constantly getting ganks -- and a good Illaoi can turn every gank into a double kill. Her only downside is her early plateau in terms of power, unless she builds AD.

AstôlfoSep 28, 20163 votes

So... feel free to link the account you actually play her on? Because OP.GG says the last time you played her was 15 days ago, and the last before that was 23 days ago. And Champion Mastery lookup says you only have around 43,000 MP on her, which equates to around 43 wins or less. Soooo....

W damage in early is complete garbage. W damage in late is meant to hit any target equally hard, instead of favoring squishies over tanks like it used to. Tanks still have a metric ton of resistances, and Illaoi really doesn't build armorpen beyond usually Cleaver (and occasionally Maw) if that, at least until late game where her last item might be a LW evolution. She will not kill someone's Spirit even in lategame with just her W alone unless they already had next to no meaningful health.

So yeah, keep on acting like you know what you're talking about. Illaoi's old counters are still counters, and she doesn't autowin lane, and quite to the contrary, will probably lose it against any actually coordinated top/jungle play post-6, or just abusing her weak-ass pre-6. An Illaoi can turn any shit-tier gank by people who have no clue how to gank her into a double kill. Her only strength in early is her E harass.

https://ih0.redbubble.net/image.63252266.3457/sticker,375x360.u2.png

Shadow HazardSep 29, 20162 votes

actually the tentacles were also changed so that they will have increasing damage as a flat value quite significantly basically as a msg build full tank but still decimate enemies. That is the specific change that made her drastically stronger, to the point that people against her would feel she is over the top.

Bots Whilst HighSep 29, 20161 votes

{quoted}

I don't think you know what you're talking about.

Not sure what weaknesses she supposedly lost. The only changes made were tentacles being visible in brush, changing W from % of AD to % health, and shortening Test debuff from 60 to 12 in return for mildly slightly faster spawns and closer spawns.

The W change made her less effective against squishies and more effective against health-builders. It's not a nerf or buff. It was a shift in focus.

Yes, a month, how incredibly conclusive.

Who resides in top lane though? Tons of health builders, (Darius, shen, gnar, garen, etc) which is why the spike in win rate is happening because she dumpsters her health stacking laner and was already a notorious 2v1 monster when the jungle comes to help.

AstôlfoSep 29, 20161 votes

The whole reason her W was changed was in response to the tank meta where everyone else went tank when she's a juggernaut. So she was falling behind with what was essentially flat damage on her W. The damage on her W itself is minimal anyways and is garbage by itself, at least until far into lategame if you build a lot of AD.

Sire HippingtonSep 28, 20163 votes

For many of ther overall low playrate champs, the 125+ value rarly gets enough samplesize even over the full patch on champion.gg

Covert Ops GnarSep 29, 20163 votes

I once saw Urgot at 100% win rate in Diamond+ lol

LuggSep 29, 20162 votes

While the sample size may be small, when a champ in that zone is consistently at a high win rate, that champ is probably too strong. 57% win rate is still pretty high IMHO.

AstôlfoSep 29, 20163 votes

When there's a playerbase of millions of people, where not that many even play Illaoi let alone main her, there's not going to be a significant data size for 125+ ranked games, and definitely not within such a short period of time after initial patch release.

LuggSep 29, 20162 votes

True, but you can see from playing against Illoai just a few games as a melee champ that she's pretty broken. Plus her overall winrate is pretty good right now.

She falls into the category of binary champs ( MasterYi Rengar etc) that totally dumpster a lot of champs and are dumpstered by the rest. Binary is never healthy for the game. It's ok to have counters, but they should never be to the point where a champ has pretty much zero chance against them.

If you main a champ, you should be able to go against ANY other champ in the game and have a reasonable chance to win the matchup. That just doesn't exist for several champs right now. When your champ has little chance against an opponent of equal skill solely due to the opponents champ just being OP in that regard, it needs to be fixed.

AstôlfoSep 30, 20161 votes

It's more of juggernauts countering most melee matchups in early in general due to usually being lane bullies. Same applies to Darius. Difference being Darius' pressure zone moves because he is the zone, Illaoi's being locked to the pit. Same with other juggernauts as well.

I disagree. It's nowhere close to zero chance. It's just a matchup that's a bit harder than usual for most people because it relies on playing differently as they would compared to other matchups, and most people are not used to having to change their approach.

It's the same concept as Fiora for example, in the sense she's one of the best 1v1 duelists in the game. If you continually try to 1v1 her as you would say, I dunno, an Orianna, it's not going to work out. You need to get your team to help. Same idea here. If you continually try to out-bully/damage them in early, it probably won't work and you really shouldn't be trying it anyways.

It also usually helps a lot when there's a second presence to gank or CC or such in early, which is usually lacking in low ELO.

In addition, there are multiple melee matchups that Illaoi does go subpar or even with (not regarding winrates, but as in the kits and stats/etc themselves). Same with every other juggernaut. It's difficult. It's nowhere near impossible.

LuggSep 30, 20161 votes

Fiora and Darius are great examples of very unhealthy kits. Why does Fiora have the best AoE heal in the entire game??? That is ridiculously stupid. It is better than Soraka's or Janna's ult healts. It should only affect her or be reworked into something else.

Darius needs to either lose his heal or the resets on his ults. Having both is pretty ridiculous. You can dodge the heal, but then you either get right next to him, which he wants, or you are zoned out of the area. Unlimited resets may be the most unhealthy mechanic in the game. Just look at how toxic it is on Kat, Tris, Jinx, etc.

Duelists and high damage bruisers don't really need utility in their kits. Their job is to kill people not be supports for their team.

AstôlfoSep 30, 20161 votes

Janna's heal is better in the majority of cases, Soraka's ultimate being second. Neither of them require any sort of setup for their mass heal, and in Janna's case, it has an instant blow-back for a similarly instant disengage/peel from anything.

Fiora needs to hit all 4 Vitals for there to be any sort of meaningful heal, because if the target dies with less than 4 hit, the Heal is reduced. Also, the fact the heal is AoE means little when her teammates wouldn't be there to receive it unless it was like a 4-5 v 1 in the first place. If Fiora ulted someone in the middle of their entire team and still managed to get all 4 Vitals off/otherwise kill them with significant number of Vitals popped, the enemy team is garbage for not having done squat to stop her.

Darius' Heal only applies if he hits enemy champions with his Q, which is incredibly slow in the first place. It also is not that large of a heal unless he hits multiple. If he's going against multiple in the first place, at least one of them should've applied GW on him or just CC'd him before Darius even got his Q off. The ult reset mechanism is very poor for him actually, in contrast. It encourages killstealing because otherwise he cannot get the resets off, and he won't get more than 2 until level 16 anyways. He also won't be getting resets if someone just bothered to CC him and keep said lower health targets away and kiting him instead of attempting to duel a juggernaut while not dodging anything and eventually dying.

Neither are much of a support.

LuggSep 30, 20161 votes

Janna's ult is better, but due to the knockback, not the heal since it doesn't do as much. Soraka's ult doesn't heal near as much as Fioras. Yeah, Fiora's isn't instant like those, but it's not like it's hard to pull off. Hitting the vitals is very easy since she gets a very significant MS buff during the ult, has a dash if they flash, and to top it off she only needs to hit ONE vital to trigger it. Let's not forget that she also has 8 seconds to hit all four, which is an eternity in League. Also, she can easily do it next to the enemy team since if they focus her, her mid and ADC will wipe the other team.

12% heal for hitting just one person is pretty large IMHO. Especially considering that he can use it at least twice in fights. Keeping low health targets away from his is pretty hard due to Apprehend. Also, if you use your CC on him, his mid and ADC free to rape your team. It's basically a lose/lose proposition most times.

Harvey RabbitSep 29, 20162 votes

{quoted}

While the sample size may be small, when a champ in that zone is consistently at a high win rate, that champ is probably too strong. 57% win rate is still pretty high IMHO.

No, I'm not sure if it's just small sample size or what, but it's definitely jacked. A few weeks ago the site had Urgot listed at 60% win rate for people with 125+ games, but now he is listed at 48% win rate for them. And he hasn't had any changes to his kit since patch 6.5.

It's very unclear what "2014-2015" means in the stats about win rate by % game played, but indicates to me that it is probably worthless data (vis a vis champions' current strength) regardless of sample size.

LuggSep 29, 20161 votes

It's pretty amazing how small changes to champs or the game and suddenly make some champs OP.

It usually has to do with their main(s) counter getting nerfed. This is an unintentional buff for them.

PennsSep 29, 20161 votes

For mains with over 125 games played 57% is still pretty ok imo.

LuggSep 29, 20161 votes

I agree, but when the rate bounces between 57 and 68, there's probably a problem. A healthy champ for a main would average between 55-60. That means you can definitely climb, but you aren't going to dominate most games.

AstôlfoSep 29, 20161 votes

But you're not accounting for the small sample size attributing to said bouncing in the first place.

By that logic, Heimerdinger top is probably a problem because it bounced from 66.95% winrate to 57.19% winrate at 50-125 games.

LuggSep 29, 20161 votes

Agreed. At one point Aatrox had like an 80% win rate one patch for 125+ games. Then it dropped to like 54% the next one.

A lot of the time it's due to one or more of that champs main counters being nerfed, which is a buff for them.

ADD: Heimer does need some rework though. They need to lower turret damage and move it to his grenades. That way he is a more skill shot based champ and would be way less toxic.

AstôlfoSep 30, 20161 votes

I honestly feel like nerf/buff-ing counters should be considered as a valid option more, instead of being forced into the stereotypical mindset of adjusting a champion themselves, or occasionally, their items.

LuggSep 30, 20161 votes

I again agree with you. There have been several times that a champion had a very high win rate solely due to their counters being under powered at the time.

ChaoticPineconeSep 28, 20161 votes

Its not an entirely meaningless statistic.

Champion.gg at the moment

2,814,540 champions analyzed on Patch 6.19
Illaoi has a 4.53% play rate
7.4% of those players have 125+ gamest.

So what we find is that there there are about 9,500 games played on Patch 6.19 where Illaoi was Top and had 125+ games played. This is NOT a small sample: though this example compares apples and oranges, in a random survey 9,500/3,000,000 gives you a 95% confidence interval with a 1% margin of error.

If this were a population survey, that sample size would be perfect. Unfortunately, League as too many variables to really analyze - such as the millions of different combinations of the 9 other champions and the different skills of different players. But if youre going to bring up any statistic for League, it is subject to these variables.

tl;dr - Its 9,500 games over the last week. This isnt like Jungle Mordes with 125+ games played, which is less than 300 games over the last week. It might not be the best statistic, but falls far short from meaningless.

Harvey RabbitSep 29, 20161 votes

{quoted}

Its not an entirely meaningless statistic.

Champion.gg at the moment

2,814,540 champions analyzed on Patch 6.19

Illaoi has a 4.53% play rate

7.4% of those players have 125+ gamest.

So what we find is that there there are about 9,500 games played on Patch 6.19 where Illaoi was Top and had 125+ games played. This is NOT a small sample: though this example compares apples and oranges, in a random survey 9,500/3,000,000 gives you a 95% confidence interval with a 1% margin of error.

If this were a population survey, that sample size would be perfect. Unfortunately, League as too many variables to really analyze - such as the millions of different combinations of the 9 other champions and the different skills of different players. But if youre going to bring up any statistic for League, it is subject to these variables.

tl;dr - Its 9,500 games over the last week. This isnt like Jungle Mordes with 125+ games played, which is less than 300 games over the last week. It might not be the best statistic, but falls far short from meaningless.

Are you sure? Then what does "2014-2015 Ranked Experience" mean in reference to Win Rate by Percent Games Played?

ChaoticPineconeSep 29, 20162 votes

Its difficult to say "_____ IS WHAT IT MEANS," without accounting for all the possible variables. For example, a 60% winrate in Challenger means something different than a 60% winrate in Plat. A 60% winrate with 100 games played on the champion and 200 games overall is different than 100 games played on the champ but 5000 games overall. But heres a few possibilities that could explain it:

For champions like Mid Morde with a high winrate, low percent of games played, and a spread of Ranked Experience that starts with a very very low winrate and spikes to a high winrate, it might mean it is from dedicated Morde mains in the midlane who know the matchup. But lets be fair - not many people face a good Morde Midlane, so it might be a lack of experience from the opponent. Or it can also be from Morde Top players who queue as Top/Mid and selectively picking into good matchups when they get mid and happen to get into a good matchup.

For champions like Top Kayle with a high winrate, most games played at the 5-15 Ranked Experience and a highest winrate at that level, it could be that Kayle is being picked in favorable matchups. For people who main Kayle and are willing to pick her into any matchup, that winrate drops.

For Illaoi her winrate goes from 45% to 58% linearly based on the Ranked Experience. Her winrate goes from 53% to 48% based on the game length, again pretty linearly. This might be because as your experience on her goes up, you know how to put down more pressure in lane and snowball a game out of control before 30 minutes. This also might explain why her E max has a higher winrate than her Q max - Q max allows you to oneshot a wave, but E allows you more dualing power/poke/lane pressure.

Harvey RabbitSep 30, 20161 votes

Great post, I think you are totally on point. But what I meant is literally: what does it mean that for that particular statistic champion.gg specifies "2014-2015 ranked experience"? If they are using the amount of games with that champion in 2014-2015 to select the pool to draw the data from, then that could be a vanishingly small sample size in many cases where the particular champion or role has been very unpopular.

Most people see the "Win Rate % by Games Played" statistic and assume it means that players who have played that number of games with the champion are having that much success, or lack thereof, but since there is a parenthetical "(2014-2015 Ranked Experience)" attached to the statistic it indicates to me that it is not so straightforward. Nothing in the site's faq explains what "(2014-2015 Ranked Experience)" means, but if it means simply what it sounds like then it is not the statistic it seems to be and not necessarily a useful statistic at all.

ChaoticPineconeSep 30, 20162 votes

Ya know, I havent looked into it. I also was a little confused about the "ranked experience" - its weird to have 2014-2015 as your source of experience, right? I'm glad others picked up on it.

Now, the explanation that I'm hoping for is that its a failure by the moderator to update his site - when the season changed, they didnt update the category. Then the statistic would be more useful and relevant, especially if they meant "The season that occurs between 2015-2016."

It would also explain why many champions have a dropoff in winrate at 125+ games played. Youd expect them to be the most practiced, right? Its weird that some champions winrates are highest at 15-50 games played but significantly lower at 125+. But if it were only took the data from this season, it makes sense - the ones at 15-50 games are still climbing, while the ones at 125+ have climbed to the highest they can get.

This conjecture isnt exactly out of nowhere, but theres nothing proving it. My personal thoughts - the site is run automatically. It gathers the data from Riot's API, and there are programs in place to sort the data into the categories we see. In the F.A.Q.s section, the creator/moderator links to his reddit, /u/joeldo. It seems like he had very high activity over a year ago, but has 1 post (dated 9 months ago) since then.

I figure that he was done with his project and moved onto new things. The tag still reads 2014-2015 ranked experience, since thats when he was still active, but the actual numbers are from this season.

Or maybe its wishful thinking.

Harvey RabbitSep 30, 20161 votes

{quoted}

Ya know, I havent looked into it. I also was a little confused about the "ranked experience" - its weird to have 2014-2015 as your source of experience, right? I'm glad others picked up on it.

Now, the explanation that I'm hoping for is that its a failure by the moderator to update his site - when the season changed, they didnt update the category. Then the statistic would be more useful and relevant, especially if they meant "The season that occurs between 2015-2016."

It would also explain why many champions have a dropoff in winrate at 125+ games played. Youd expect them to be the most practiced, right? Its weird that some champions winrates are highest at 15-50 games played but significantly lower at 125+. But if it were only took the data from this season, it makes sense - the ones at 15-50 games are still climbing, while the ones at 125+ have climbed to the highest they can get.

This conjecture isnt exactly out of nowhere, but theres nothing proving it. My personal thoughts - the site is run automatically. It gathers the data from Riot's API, and there are programs in place to sort the data into the categories we see. In the F.A.Q.s section, the creator/moderator links to his reddit, /u/joeldo. It seems like he had very high activity over a year ago, but has 1 post (dated 9 months ago) since then.

I figure that he was done with his project and moved onto new things. The tag still reads 2014-2015 ranked experience, since thats when he was still active, but the actual numbers are from this season.

Or maybe its wishful thinking.

Yeah, that's totally plausible, maybe even the most likely explanation. But at the same time the volatiility of the win % based on ranked experience makes me suspicious, like the op, that there is something amiss with that statistic.

Craft ZeppelinSep 29, 20161 votes

even with a large sample size the website is a literal troll producing factory Like ADC first item hurricane rush has 56% winrate RIGHT?

GayFluffyUnicornSep 29, 20162 votes

That's what adcs do. Atleast the crit building ones. Get zeal, bf sword, finish runaans or rush runaans immediately. Crit is so much stronger than prior seasons.