@Riot: Tahm Kench officially broken

Lugg·Sep 3, 2015·5 votes·13,710 views

Tahm jungle has now reached a 91% (no, that isn't a typo) win rate in Platinum and above games with players who have 125+ games on him.

Tahm top and support are at 59% for the same demographic. This guy may be the most in need of a hotfix in League history. Time to break out the Thor nerf hammer.

EDIT: Someone found out the actual sample size, it is pretty small as could be expected. It will be interesting to see after a few weeks since his win rate is climbing pretty fast across the board.

35 Comments

RiotMeddlerSep 3, 201518 votes

Not sure what data you're looking at, sounds like a small sample size issue though. Looking back at high skill Tahm Kench games where Tahm took Smite during the month of August for example I'm seeing around a 45% win rate for jungle Tahm. People that have played a lot of Tahm will push that a bit higher, won't be more than roughly 5% though as an estimate (I don't have that exact breakdown to hand at the moment). Similarly top and support win rates are both in pretty expected ranges given Tahm's overall performance (mid 40s on average for high skill players).

Krys StarSep 3, 20151 votes

I just had a game yesterday where Tahm Kench had Iron Solari, Dead Man's Plate, Cinderhulk and Spirit Visage was doing enormous amounts of damage to our team. He literally 1v3'd us and took all of us down by himself when we were ahead by maybe 17 vs 9.

http://matchhistory.na.leagueoflegends.com/en/#match-details/NA1/1938955993/32288739?tab=overview

He was nearly unkillable thanks to being tanky, with high health and his gray health and his Tongue Lash was doing absurd amount of damage as well as Devourer and Abyssal Voyage's passive. Is this the intended direction you guys are going with him? A tank that can dish out a lot of damage by himself? I'm just curious, because that match left me appalled with how many buffs he's been receiving consecutively when he isn't even in a terrible spot, just more team coordinated before then he is now.

Woook3rSep 3, 20151 votes

[{quoted}](name=Frozen Beauty,realm=NA,application-id=3ErqAdtq,discussion-id=qM24jndG,comment-id=00010000,timestamp=2015-09-03T21:34:39.676+0000)

I just had a game yesterday where Tahm Kench had Iron Solari, Dead Man's Plate, Cinderhulk and Spirit Visage was doing enormous amounts of damage to our team. He literally 1v3'd us and took all of us down by himself when we were ahead by maybe 17 vs 9.

http://matchhistory.na.leagueoflegends.com/en/#match-details/NA1/1938955993/32288739?tab=overview

He was nearly unkillable thanks to being tanky, with high health and his gray health and his Tongue Lash was doing absurd amount of damage as well as Devourer and Abyssal Voyage's passive. Is this the intended direction you guys are going with him? A tank that can dish out a lot of damage by himself? I'm just curious, because that match left me appalled with how many buffs he's been receiving consecutively when he isn't even in a terrible spot, just more team coordinated before then he is now.

shyvanna seems to be in the same place with titanic hydra or whatever its called.

LuggSep 3, 20151 votes

http://champion.gg/champion/TahmKench/Jungle

Is where I got the info. Yeah, the sample size may be small, but that is still an incredibly high win rate for any sample size. You guys should at least have this on your radar.

On the support side, TK is pretty miserable to lane against. His devour on the ADC makes them almost impossible to kill. Would you guys consider increasing his CD if he uses it on an ally? I had one game where he literally saved his Corki from death 11 times.

DecritSep 3, 20153 votes

Looking at how the curve evolves... yup, i think it is too a sample size issue. I really doubt that there is such a difference between playing 120 games or playing 150, and for playing 150 already competitively then you have to be a well dedicated player, thus reducing even more the pool.

furthermore, that is only in jungle.

LuggSep 3, 20151 votes

59% top and support is pretty high as well.

DecritSep 3, 20152 votes

Not with such a restriction: is obvious that you perform better than the norm if you are highly trained on a champion. ( because you are matched up with people that might not have a champion trained like yours, you don' t always pick the main after all ).

To be honest, going from 46-49 % median winrate to 60% while highly trained might not be enough..? i am not sure on that honestly, but for sure is not too much.

LuggSep 4, 20151 votes

60% win rate is too high for ANY champion. Anything over 55% needs nerfed for sure. You have to realize that every champion has mains as well and they all can't have win rates that high without being broken.

DecritSep 4, 20151 votes

Ehm, nope.

That applies only to general win rate, taking consideration of who plays well and who plays badly from a different range of players, but these numbers are from a specific subset of players.

If you isolate the most expert and dedicated of those players, obviously you get higher win % compared to all those who play it.

LuggSep 4, 20151 votes

True, but those players that are that good are probably at a high elo as well. This puts them against other high elo players who probably also specialize. So if you can get that high of a win rate against similar skill level players, your champion has to be OP.

DecritSep 4, 20151 votes

not said at all, because they might be matched against someone that is not playing a champion as good as them. Plus there might even be plats that have an higher win % with a champ compared to a dia, but are less skilled with any other champion ( hard , but probable ) so they provide more champion wins than elo balance.

5000000000000000Sep 3, 20154 votes

Dude, the sample size is tiny. People with 125+ games played as Tahm jungler is 0.63%. Total games analyzed is 2496. If we multiply those, then we get 16 games. Now we multiply by 91%. That's 14 games. So 14 wins out of 16 games. This can happen based on chance.

LuggSep 4, 20151 votes

Thanks for pointing out the .63%, I didn't see that. I was trying to figure out what the actual sample size was.

You do have to take into consideration that he is still a pretty new champion and people are still learning him. If the win percentage had been 60-70%, I would have totally written it off. But 91% is pretty insane. I am looking forward to seeing where it's at in a couple weeks.

Also, since the last patch, his win rate has gone up 14 spots in the rankings and he has a high win rate for top and support as well. I think it's a classic case of Riot overbuffing a champion too soon before seeing the results of the previous buffs.

5000000000000000Sep 4, 20152 votes

People who played 125 games as a champion isn't "still learning." They are people who mastered that champion's mechanics and probably know a lot of tricks that most players don't.

Also, people have win streaks randomly. 14-win streak can happen, and even if that didn't happen, that might be two 7-win streaks, etc. I'm sure if we wait a little longer for whoever is playing Tahm Kench jungle to play more Tahm Kench jungle, the win rate will go down, simply because of chance. See Law of Large Numbers.

LuggSep 4, 20151 votes

The Law of Large Numbers doesn't apply at all in a situation with this many human variables. That only works on coin flips and roulette wheels.

As for still learning, I was referring to the rest of the gaming populace that plays TK. The average TK player right now only has 15 games on him. So as more people get him down, the higher his win rate will go up. The mains are a good indication of what direction that will take.

It will be interesting to see where everything settles out over the next few weeks.

ThalesSep 3, 20151 votes

The sample size is ludicrously small. There's 2.4k games played as Tahm Kench jungle, and games by people with 125+ Tahm Kench games make up less than one percent of those. While I'm unable to find something that's totally consistent with the numbers there (2469*0.0063 is about 16, which can't produce a 90% win rate), it seems like there were 15-23 games played as jungle Tahm, of which all but two were won. Certainly, this is very impressive. But it's probably the product of one guy.

Riot doesn't nerf based on one guy. At that sample size, it's not going to be very meaningful. I once won 17 games in a row as Orianna. I'm not a very good Orianna, and I can't say for sure how that happened. Is it so unreasonable for some guy who likes Tahm Kench to hit a similarly lucky run in one patch? For every champion, in every patch, there's probably going to be some high skilled player who does astonishingly well with that champion, and it's just that in this case, there's so few people playing Tahm Kench jungle that there's no one to obscure this guy's run.

Also consider that you chose jungle, when there's three different places you could have looked for the number you were trying to find. Be very careful of starting with a postulate — Tahm Kench is OP — and then going to search for evidence to back it up. Much better science to just try to investigate how good he is in general.

One last thing. You said "Yeah, the sample size may be small, but that is still an incredibly high win rate for any sample size." I think you'll agree that it's not an unreasonably high win rate for a sample size of one game. Nor of two, or three, and so on up to some point. We'd also agree that a 90% win rate is incredibly high for a sample size of, say, 10,000 games. The question of where to place the dividing line is one that's the focus of a lot of stats classes.

TotalJerkSep 3, 20151 votes

Who the fuck has already played 125 ranked games with Tahm Kench jungle?

LuggSep 4, 20151 votes

Evidently some people are doing it.

ReshiKillimSep 3, 20151 votes

I'm seeing a 90% win rate among the 0.55% of players that play jungle Tahm. But only on the tiny portion of players with over 125 games on him.

His overall winrate is 10% higher than Tahm jungle's winrate for players that have 50-125 matches on Tahm.

And that playerbase that is over 125 games as Tahm jungle is 0.63% of that 0.55% players that play jungle Tahm. So it's an extremely small sample size, I'd say at most 1-200 players total. Probably actually around 10-20 players.

buniesSep 3, 20151 votes

{quoted}

Not sure which data you're looking at, sounds like a small sample size issue though. Looking back at high skill Tahm Kench games where Tahm took Smite during the month of August for example I'm seeing around a 45% win rate for jungle Tahm. People that have played a lot of Tahm will push that a bit higher, won't be more than roughly 5% though as an estimate (I don't have that exact breakdown to hand at the moment). Similarly top and support win rates are both in pretty expected ranges given Tahm's overall performance (mid 40s on average for high skill players).

he is probably looking at the statistic on one single player who probably only pick tahm in very advantageous situations and has duo que ect.

Pomax powa 423Oct 9, 20151 votes

Not sure what you're looking at, sounds like some people have over 90% win ratio in challenger with tahm kench, kda over 7 while underperforming with any other champ in the elo they are placed.

Pomax powa 423Oct 9, 20151 votes

Oh and I just made a game with a tahm kench is overfarming his jungle as a tank since he has way too much damage + aoe on his devour. He basically made more damage than the vayne who was 6/0 and got feed in lane pretty hard. And he was full tank. Tell me why this is acceptable as a champ

luckmallorySep 3, 20157 votes

You are gonna need to provide some actual proof of those statistics

MisterHealYoGirlSep 3, 20151 votes

I can't find the 91% number, but this website says that his winrate for players with 125+ games with him is about 59%. (Second graph in the rightmost column.) http://champion.gg/champion/TahmKench/Top

SKOBODOSep 3, 20155 votes

You're looking at the top data. The jungle data for 125+ games played is 90.8%. 50-125 games is 39.81%. It seems to me that there's a lack of data.

Live2LetDieSep 3, 20151 votes

Unbench the Kench!!!!

powerbatsSep 3, 20151 votes

If you want to use data to make a claim then you need to ensure you have a ample data size to go off of, such a small sample size is meaningless. If his numbers are that high after thousands of games across all skill levels and roles then yes it's an issue.

LuggSep 3, 20151 votes

Well it's every TK main from Plat to Challenger. That should be a decent size group.

The FastSep 3, 20152 votes

except its not every TK main, its ever TK game. There is a pie chart that describes the amount of players in each experience range, and the 125+ game margin is way too small to mean ANYTHING.

powerbatsSep 3, 20151 votes

{quoted}

Well it's every TK main from Plat to Challenger. That should be a decent size group.

Well that'd be true except for the player sample size is still exceedingly small compared t6o what you need to have for a true baseline. If you say there's 1k players playing him yet ignore the other 200k you're ignoring a huge sample size database to fit your numbers claim. You need both sample size of players and games played. He also hasn't been out long enough compared to say someone like Darius or Ekko as examples in both playerbase size as well as games played.

General MuricaSep 3, 20151 votes

Tahm is fkn broken,

Sniper 0Sep 4, 20151 votes

The number of players who have played 125+ games with tahm is going to be proportionately small compared to players who have played few games. You're also completely ignoring the white line...

Summonerrr 1Sep 4, 20151 votes

I think tahm needs to exist tanky enough so he can throw toxic teemo players into the enemy team (preferrably a singed who can flip him into nexus) if they are harassing teammates.

Fight poision with poision.

Righteous AF

LuggSep 4, 20151 votes

Something like this?

Pomax powa 423Sep 25, 20151 votes

Tahm Kench champion countermechanics explained in details ... Found a part where they said he was kinda balanced. Had some fun to refutate.

“But Zenon, how is that fair? I just spent all my CC on this guy, we were about to kill him, and then suddenly General Gulp comes in and whisks him away!” Well yes, imaginary FAQ person, but let me tell you about a few things I added in to help balance things a little. First of all, as I mentioned, while Tahm Kench has this pretty swell devour ability, he lacks a lot of the traditional tools **Except a stun of 1.5s every 6s without cdr ** that tanks usually carry around. He can’t lock down enemies at will **Except a stun of 1.5s every 6s without cdr ** , for instance, and he has absolutely no utility AOE tools that your average punching bag brings to a fight_** Except a stun of 1.5s every 6s without cdr _. Secondly, while eating an ally to keep them safe is super clutch, the ability does come with some baked in drawbacks. Its relatively short range means Tahm, who most of the time will be on a support budget and thus not the tankiest man in the game_ Or maybe both because of his shield and crazy peel (it is like the broken version of braum) Oh and epic basic tankyness_, will have to stand close to your team_Because you were mean to use him not to peel ??? , or at least close to members of his team who you want to kill Because He is mean to have big kill potential with big damage even when he is support. So uh… you can just kill him if you focus him Or can you ? Killing the most tanky champ of the game 2x because of his shield. Thats how riot makes countermechanics**_. Lastly, if Tahm Kench does eat his ally, the ability’s minimum duration means you’ll still have a few precious seconds to leverage your temporary numbers advantage! **Or maybe get destroyed because you can't kill the target you focused **

“Okay. Tell me about enemy devour.” As a raw mechanic - taking enemies out of a fight and severely reducing their vision for a few seconds - enemy Devour is incredibly powerful, which forced us to consider the circumstances through which it would be okay to use, even for a few seconds_** Consider if he is okay is one thing, balancing is another ... . We knew that Tahm Kench couldn’t just go around eating enemy champs at will - that would be ridiculous - **Then why did you make him do that ?!**so we started figuring out ways to backload the ability, or establish x criteria that you need to meet before y can happen. For comparison, Malphite can ult into a teamfight at will because Unstoppable Force is a frontloaded ability - it has immediate impact and once it’s spent, he has to wait a long while before he can use it again.Because it is nearly the only thing good about malphite ? And you make it way longer to use than Tahm Kench devourer _ Because we had decided he would NOT bring the usual tank toolset of locking down an enemyYou decided that ? That's why he can remove a champion for 6s ?, we couldn’t allow Tahm Kench the same flexibility, so ended up implementing passive stacks into his kit2 auto attacks, that's just enough. So, while Tahm Kench can eat enemy champions and take them out of a fight for a few seconds, he has to hit them with a few_ (2) _** basic attacks (or licks) first. This takes time and proximity**2 seconds if you have 1 attack speed per second. And you slow him to make sure he don't get too far. And you stun him if he want to get away without being killed**, and telegraphs his intentions to the enemy team, giving them plenty of time to peel Peel ? Did you mean realize they will die for the River King’s target. _

I find everything about him wrong. Oh yeah and never play 3v3 again, it's even clearer that he got way too much damage ( most of the games in nearly all my games) , tankyness ( more than shen who is mean to only be tanky) and lockdown ( Cmon 3 AA , no wait 2 because he licked you before. It's a bit too broken).