Please do NOT Buff Tahm

Lucifer·Jul 15, 2015·104 votes·20,369 views

He is already strong, but he's definitely close to being well balanced. These buffs will drastically put him over the edge. People are still learning to use him properly. He does massive damage while building tanky already, and he's already seeing quite a bit of play top and mid.

PBE Changes: Tahm Kench

Tongue Lash [Q] - Damage increased to 80/125/170/215/260 (up from 80/115/150/185/220)
Abyssal Voyage [R] - Cooldown lowered to 120/110/100 (down from 160/130/100)
Abyssal Voyage (R) passive's bonus damage to basic attacks & damaging abilities increased to 4/6/8% of bonus health from 4/5/6% of bonus health.

131 Comments

SwiftKitten88Jul 15, 201530 votes

plz tell me this is a joke.. is he seriously getting BUFFED?

He needs NERFS not buffs..

LuciferJul 15, 201526 votes

I have a feeling it's a knee-jerk reaction to him not seeing a ton of play right away on release. This isn't because he's weak, it's because he's honestly not that aesthetically pleasing, which turns off a lot of players to playing him. This has nothing to do with how well he is balanced, and if these buffs go through, he's going to go completely over the edge.

Neo CyrusJul 15, 20152 votes

Not to mention his win rate is perfectly normal for a newly released support champion.

260 base damage on his Q... that's more than assassins have.

APrinnyDoodJul 15, 20154 votes

Not to mention his win rate is perfectly normal for a newly released support champion.

Uh, not really man. Every recent support except bard released to a higher winrate than Tahm, and Bard got some pretty major buffs. This low of a winrate doesn't necessarily mean Tahm is weak, but it isn't the norm either.

Neo CyrusJul 15, 20151 votes

Even Thresh had such win rates before he eventually got nerfed. Maybe we're looking at different sources though. I've been looking at Plat+ NA/EU/KR.

APrinnyDoodJul 15, 20152 votes

Thresh didn't have this winrate on release. Tahm currently has around a 38% winrate a week after his release. Thresh released to a 42-43% winrate. Still not great (and not representative of his strength), but higher than Tahm by a significant amount. I definitely agree that Tahm's low winrate doesn't mean he's weak particularly since it is rising, but it still is outside of the normal range for supports.

Neo CyrusJul 15, 20152 votes

Thresh had win rates in the 30s. Maybe you weren't listening, I only look at ranked Plat+ NA/KR/EU. Norms or stats from BR/LAN/LAS/OCE/etc are worthless.

APrinnyDoodJul 15, 20151 votes

Thresh had win rates in the 30s.

No he didn't. He came out to a 42% winrate. Even in Platinum and above, it was not in the 30s.

Do you want to link me to where you're actually seeing these numbers? That might be more useful. I'm looking through the old general discussion threads from his release week, and not a single thread about him is referencing any numbers in the 30s.

Neo CyrusJul 15, 20151 votes

Don't remember or I would have linked it. It might have been OP.gg which is what I check these days but I can't even remember if it was around back then.

APrinnyDoodJul 15, 20152 votes

OP.gg was technically around at the time, but I don't think it had the same focus it does now. I remember that it was partially translated to English at some point in 2013 (so probably after Thresh's Janurary release), and that it was used by Koreans before then, but I didn't pick it up until 2014 so I don't really know any specifics.

I'm sorry mate, but I am almost completely sure you're remembering Thresh's release winrate wrong. There's a possibility that you're right here, and I have to at least acknowledge that possibility, but I'm not seeing anything to convince me of it. I've seen one thread which brought up a 40-45% winrate in gold-plat, but I'm not seeing anything else that suggests anything lower than a 42% winrate. Maybe you're right, but for now, I'm sticking with what I remember and what the numbers I can find suggest.

Mast3rShakeJul 15, 20155 votes

Tahm Kench is played in more places than just support so that isn't a fair comparison.

For CarthageJul 15, 20152 votes

{quoted}

I have a feeling it's a knee-jerk reaction to him not seeing a ton of play right away on release. This isn't because he's weak, it's because he's honestly not that aesthetically pleasing

How superficial, considering he's one of the most interesting champions in the game. He's less of a Super Smash Bros. fighter and more of a character that has actual feeling, motive, and depth. Would you rather play as an imperfect human or this perfect marble statue? Why not the statue? It's pretty, isn't it?

That aside, Tahm does actually need buffs - they're just putting them in the wrong place. Pre-6, Tahm is absolute ass and does fuckall other than soak up damage, which is the problem with his E currently - that while his early game is ass, he's able to survive it no problem thanks to healing a fair amount of harass back up.

I guess he doesn't need so much of a buff as a power shift - move some of his E's in-lane healing to some early Q or W damage so that you can actually have a way to bully him out of lane with poke as opposed to having him soak up everything and watching your mana run dry. His end-game should stay where it is now - he's easily kited but can still do his job of tanking, so he's nice in teamfights but entirely possible to outduel - see Nautilus, Sion, any other tank in existence. No buffs are needed there, so no buffs should be given there.

Nidalee SO NYAAAJul 15, 20151 votes

For one his not humen and he grosses me out. Cho is not pretty ether is Rek but they both don't have a gross design.

For CarthageJul 15, 20151 votes

.......................................................................................................................................................

You alone could give me an aneurysm.

Earl EulrichJul 15, 20154 votes

yea, winrate of players that have more than 10games on him already is >50% - buffing him now will only lead to major nerfs in the future.

RiotZenonTheStoicJul 15, 201553 votes

Where are you getting that number? Our internal data suggests win rate after 10 previous games experience is 45%. To hit 50%, you need an average of 30 games on Tahm. Usually you're at something like a 52-53% winrate after 30 games.

Remember, too, that we aim to be data-informed and not data-driven. We don't make changes to the game because the data says one thing. There isn't a Rito design metagame where we try to nudge certain numbers into certain ranges. You look at numbers like these to guide you in examining play patterns. Armed with these numbers and many more (something else we like to look at is both win rate over game length and win rate by ELO) I examined a frankly unhealthy amount of live games featuring Tahm. Based on these, I have come to the following conclusions:

Tahm is very strong in games where one ally particularly takes off. Say a mid-lane Talon that goes crazy, or a Draven ADC who got a bunch of early kills. In these games, Tahm focuses on ally devour and ally ultimates to both bring the high performing allies where they need to be and to get them out of trouble when the enemy team tries to murder them. In these games we see Tahm hang out near his allies, choosing for the most part NOT to run up at enemies and attack them, and instead devours his high value allies and carries them to safety.
In games where Tahm's team is behind and no lane is doing particularly well, his pattern shifts subtly. He now spends a lot more time interacting directly with the enemy front line, stacking up An Acquired Taste and looking for impactful devours / stuns. In this role he becomes a tank weakener who helps his team actually burn through the enemies that get in their faces.
In games where Tahm does not have a strong ally to protect, his protect pattern changes too: instead of eating and running, he eats and fights. If the enemy spends a lot of resources on an ally who has an average power share, Tahm can safely protect the ally for some seconds and then spit them into danger again. If the enemy spends two rounds of high impact cooldowns on someone who's not exceptionally valuable, Tahm has created a small advantage for his team.

All of the above suggest to me that in the support case, higher base damages will disproportionately buff the losing / even case, which is in bad need of buffing, and will do less to the "we are already super winning" case, because in THAT case Tahm tends to exit combat with his high value ally.

Additionally, in lane, Tahm tends to land Qs mostly when the enemy aggresses on him; due to the range and speed of the spell and due to the fact it stops on first target hit it's not a great opportunistic poke spell. So it feels to me that putting damage into this spell is again help his behind case more than his ahead case.

As for the R CD, we felt that the spell might become oppressive in lane if the jungler makes himself available for a tele-gank every time the R is off CD. This has not manifested in any games I've watched; between the spell's CD, the fact the jungler has two other lanes (and his jungle) to look after, and the fact that bot lane is not always gankable, we rarely see more than one Abyssal Voyage gank before the laning phase ends. We were too cautious with the 160s cooldown on a spell that requires so much ally cooperation and have consequently backed off the very high 160s rank 1 cooldown. Do note that in the end game the CD is unchanged.

KingsgraveJul 15, 20151 votes

{quoted}

Where are you getting that number? Our internal data suggests win rate after 10 previous games experience is 45%. To hit 50%, you need an average of 30 games on Tahm. Usually you're at something like a 52-53% winrate after 30 games.

Remember, too, that we aim to be data-informed and not data-driven. We don't make changes to the game because the data says one thing. There isn't a Rito design metagame where we try to nudge certain numbers into certain ranges. You look at numbers like these to guide you in examining play patterns. Armed with these numbers and many more (something else we like to look at is both win rate over game length and win rate by ELO) I examined a frankly unhealthy amount of live games featuring Tahm. Based on these, I have come to the following conclusions:

Tahm is very strong in games where one ally particularly takes off. Say a mid-lane Talon that goes crazy, or a Draven ADC who got a bunch of early kills. In these games, Tahm focuses on ally devour and ally ultimates to both bring the high performing allies where they need to be and to get them out of trouble when the enemy team tries to murder them. In these games we see Tahm hang out near his allies, choosing for the most part NOT to run up at enemies and attack them, and instead devours his high value allies and carries them to safety.

In games where Tahm's team is behind and no lane is doing particularly well, his pattern shifts subtly. He now spends a lot more time interacting directly with the enemy front line, stacking up An Acquired Taste and looking for impactful devours / stuns. In this role he becomes a tank weakener who helps his team actually burn through the enemies that get in their faces.

In games where Tahm does not have a strong ally to protect, his protect pattern changes too: instead of eating and running, he eats and fights. If the enemy spends a lot of resources on an ally who has an average power share, Tahm can safely protect the ally for some seconds and then spit them into danger again. If the enemy spends two rounds of high impact cooldowns on someone who's not exceptionally valuable, Tahm has created a small advantage for his team.

All of the above suggest to me that in the support case, higher base damages will disproportionately buff the losing / even case, which is in bad need of buffing, and will do less to the "we are already super winning" case, because in THAT case Tahm tends to exit combat with his high value ally.

Additionally, in lane, Tahm tends to land Qs mostly when the enemy aggresses on him; due to the range and speed of the spell and due to the fact it stops on first target hit it's not a great opportunistic poke spell. So it feels to me that putting damage into this spell is again help his behind case more than his ahead case.

As for the R CD, we felt that the spell might become oppressive in lane if the jungler makes himself available for a tele-gank every time the R is off CD. This has not manifested in any games I've watched; between the spell's CD, the fact the jungler has two other lanes (and his jungle) to look after, and the fact that bot lane is not always gankable, we rarely see more than one Abyssal Voyage gank before the laning phase ends. We were too cautious with the 160s cooldown on a spell that requires so much ally cooperation and have consequently backed off the very high 160s rank 1 cooldown. Do note that in the end game the CD is unchanged.

Any thoughts on Top Tahm?

He is literally the most annoying thing to face in lane but also can't really do anything other than "you can never dive me." He feels like Maokai in this regard, but Maokai is just better than him most of the time.

The buffs don't seem to really help his Top lane at all, at least not directly.

RiotZenonTheStoicJul 15, 20158 votes

Top Tahm is suffering like LongPanda. He is sitting at a 36% winrate with little sign of improvement. I'm actually very concerned we trapped players into playing him top by promising viability in the champ spotlight, and the R % dmg buff is definitely aimed at top lane, where you can afford to build a lot more health more quickly, where you hit level 11 and level 16 much more quickly, and where you're often in a position in the team comp where it's not your job mostly to keep the ADC alive, so you can attack a whole bunch more.

KingsgraveJul 15, 20152 votes

{quoted}

Top Tahm is suffering like LongPanda. He is sitting at a 36% winrate with little sign of improvement. I'm actually very concerned we trapped players into playing him top by promising viability in the champ spotlight, and the R % dmg buff is definitely aimed at top lane, where you can afford to build a lot more health more quickly, where you hit level 11 and level 16 much more quickly, and where you're often in a position in the team comp where it's not your job mostly to keep the ADC alive, so you can attack a whole bunch more.

His problem Top is that he has no reliable waveclear. It seems pretty decent if you max W first, but you are making Q even better to max first. Also, the way W works just makes it so you can only reliably hit 2-3 minions at once with it. Once you hit 6 your trading is a lot better, even somewhat decent, but by then most Tahms have been destroyed completely.

The only real way I could see Top Tahm get better is if you moved the R passive to his Innate.

He doesn't have enough damage to justify the lack of waveclear against the majority of tops, and since he is AP he has no way to deal with this problem. If he was AD he could buy Ravenous (like Wukong has to do because of his E not being good enough) but alas he usually rushes RoA into Full tank. Iceborne might be decent but slows don't stack anymore.

I'd say if his innate passive stacks worked on Monsters he could be a decent Jungler, but right now he is too unhealthy in clears. It would certainly do him better to be a Jungle/Support rather than a Top/Support.

Jungle Tahm seems really good tho if his innate worked on Monsters. The lack of Ultimate really hurts him in Lane, but it really helps him in the Jungle. Q can be pretty useful for Ganks (on par with Nunu) and he could probably coordinate with a laner to swamp their way into the middle of a lane for a gank. It's odd you guys wanted him to be Top instead of Jungle, he seems perfect for it.

AzlagorJul 16, 20152 votes

i find his wveclear not that bad, W+sunfire clear out waves faster than any mini-gnar could ever dream off. I also find him to be incredible strong 1v1 vs any champ that can't simply kite him down, atleast post 6, but most of the time even pre 6.

His main problem top aswell as jungle is the lack of escape. he can't go for counterjungleing if he doesn't have alot backup from his team, cause he has no escapes to get away once cought, which also is a problem if he gets counterjungled by strong early duelists like lee. HIs Ganks also are very dependend on your laner, as you have no mobillity, and you only cc is a skillshot slow. In top, he can't really push out the lane, as he's insanly vulnerable to ganks. No aoe cc, no hard cc unless he comits to a fight, no mobillity. This also makes his splitpush despite the global ult rather risky, especially for soloQ.

I think one of the major reasons for his bad performance top is that people max Q(acording to champion.gg) It has a horrible scaling with ranks, only 35 damage per rank and 5% slow. 50% slow is more than enough in most situations, and the cd stays the same, while manacost even go up. If you max W instead, you gain 50 more damage per rank on a longrange aoe nuke, which gives actually some good waveclear, 300damage rank 5 is pretty high for an aoe nuke, especially since it profits from the ult passive aswell. it's also less cd, more speed for ally eat, more duration for ally eat and more % damage while the manacosts stay the same. Maxing W first is just so much more effective.

Nameless VoiceJul 15, 201519 votes

This is a really detailed and well-explained look at the thought process behind these changes.

Thank you for the excellent post. I would love to see Riot posts like this more often.

RiotZenonTheStoicJul 15, 201518 votes

Thanks! We try!

Leaf of the lakeJul 15, 201514 votes

you need an average of 30 games on Tahm. Usually you're at something like a 52-53% winrate after 30 games.

Doesn't this suggest that he has a high skill floor, and that after spending that time to learn the champion, he may very well be balanced? If a champion is difficult to play, but balanced when a player finally learns how to use them, buffing the champ to make things easier for the new learners will just result in it becoming OP for those that have already gotten past that phase.

RiotZenonTheStoicJul 15, 201517 votes

I think you mean skill ceiling. A skill floor is where you start your first game.

He has both a low ceiling and a low floor right now, but that is absolutely as expected given how new he is. You're at 50.2% on Tahm after 30 games. On Rek'Sai that number is 54% (!!), and she starts at 40% (Tahm players currently start at 32%, but that'll come up very quickly as general "how to be a Tongue man" knowledge spreads).

Tahm is not mechanically difficult to play. He has one skill shot that is free to hit at medium and close range, one targeted ability, and one "just press the button whenever" ability. Low mechanical complexity was one of my goals on this kit! Strategically, the situation looks a little different, but in solo queue environments this should be by far the less important one.

Just to be clear: a champion being at around 52% after 30 games is considered balanced. Most people never get to 30 games. Or you could play just Malphite and have a statistical 52% win rate on your first game...

KingsgraveJul 15, 20152 votes

{quoted}

I think you mean skill ceiling. A skill floor is where you start your first game.

He has both a low ceiling and a low floor right now, but that is absolutely as expected given how new he is. You're at 50.2% on Tahm after 30 games. On Rek'Sai that number is 54% (!!), and she starts at 40% (Tahm players currently start at 32%, but that'll come up very quickly as general "how to be a Tongue man" knowledge spreads).

Tahm is not mechanically difficult to play. He has one skill shot that is free to hit at medium and close range, one targeted ability, and one "just press the button whenever" ability. Low mechanical complexity was one of my goals on this kit! Strategically, the situation looks a little different, but in solo queue environments this should be by far the less important one.

Just to be clear: a champion being at around 52% after 30 games is considered balanced. Most people never get to 30 games. Or you could play just Malphite and have a statistical 52% win rate on your first game...

Malphite is far more straightforward than any of your other examples. It's hard to play him incorrectly.

RiotZenonTheStoicJul 15, 201544 votes

I mean, you could theoretically lose track of where on your keyboard your R key is.

noodle0117Jul 16, 20153 votes

I like you. I like you very much.

Static DuckJul 16, 20156 votes

I've done that...

A Wild Trinket Ward appeared.

Earl EulrichJul 15, 20152 votes

you´d think, but i would really compare malph to oriana - it´s mostly about being patient/experienced enough to land a good ult in a teamfight. And many players are either not patient enough or too hesitant...

TotalJerkJul 15, 20158 votes

I feel like not every champion is of equal relative difficulty to everyone. I feel like I did pretty well on my first few Thresh, Orianna, and Syndra games, but I struggled the first time I tried to play Ryze, Malphite and Sivir. What makes a champion "difficult" likely varies wildly from player to player.

RiotZenonTheStoicJul 15, 201510 votes

That is absolutely true! But I do believe it reduces to an overall statistically relevant tendency.

KingsgraveJul 15, 20154 votes

{quoted}

That is absolutely true! But I do believe it reduces to an overall statistically relevant tendency.

Don't really know how much this is true.

Malphite isn't truly difficult, he is one of the easiest to play objectively. However, it doesn't mean every single player will have an easy time learning them.

SquarefighterJul 16, 20153 votes

But on the other hand, if you analyze 1000 players on their first game of malph vs their first game of say, zed, the average win rate for malph will most likely be higher. Even though it's not true for all players, it's true for the majority.

Sire HippingtonJul 16, 20155 votes

Mechanical, yes, decission wise...i see so many horrible Malphits that for some reason think they are clean-up assasins that lurk around fights to finish of a low health support with their ult...and those guys are often enough diamond players... So apparently, he isn't as easy to use as it looks(however, i also feel like correctly useing a fork already is an imense achievement for those players...)

Huang ShaotianJul 15, 20156 votes

I love posts like this where riot actually gives us real information and numbers as well as the thought process. I had not even thought about some of the things you have posted here. The patch notes these days go a long way to to explaining riots goals and thoughts but it never can replace discussions like these ones.

Thank you for your contribution zenonthestoic.

Earl EulrichJul 15, 20153 votes

He has both a low ceiling and a low floor right now, but that is absolutely as expected given how new he is. You're at 50.2% on Tahm after 30 games. On Rek'Sai that number is 54% (!!), and she starts at 40% (Tahm players currently start at 32%, but that'll come up very quickly as general "how to be a Tongue man" knowledge spreads).

Tahm is not mechanically difficult to play. He has one skill shot that is free to hit at medium and close range, one targeted ability, and one "just press the button whenever" ability. Low mechanical complexity was one of my goals on this kit! Strategically, the situation looks a little different, but in solo queue environments this should be by far the less important one.

I think the major problem of Tahm top might indeed be more of a strategic one for SoloQ...he just really needs rather specific team-comps to unload his potential - and early jungle help/specific matchups to really shine in lane. Low wave-clear is definitly hard to play around in current league, as it really creates a lot of map-pressure against you if you canßt clear huge waves fast). So what you could try to specifically buff toplane Tahm would be to add bonus-dmg on minions to his w (maybe just for spitting cannon-minions and monsters). And as he´s such a teamplayer - donßt you think data of 5v5s is actually more relevant to conclude upon his state of balance than SoloQ-data?

DeclinedJul 16, 20152 votes

Hey hey hey, he has TWO skillshots, not one, and you can use this 2nd skillshot to add a a stack of his passive on long range, making him 'mechanically' more difficult, as you suddenly have to consider using his W preemptively or to add stack faster, to use the Tongue for a stun instead, or a long range finisher.

Just saying.

Sire HippingtonJul 16, 20156 votes

Strategically, the situation looks a little different, but in solo queue environments this should be by far the less important one.

I feel like here lies a mistake. For Tham, decission makeing is very important, especially on a macro level, aswell as coordination with his team. In soloQ, especially in lower elos, decissionmakeing and macro play are the weak spots of nearly every player, and teamwork overall comes short. Actually, mechanical difficulty isn't the main part of skill ceiling. The winrate difference between new Lee Sins and those who have 125+ games is ~5% Yasuo is ~9.7% Garren has ~7% Trundle has 12% I'd Argue that trundel and Garren are really straight forward in terms of mechanics, while yasuo and Lee are really deamanding, yet Trundle has the biggest difference between new and experienced players, while lee has the lowes of those examples.(data from champion.gg)

So going by that, his soloQ performance, especially on inexperienced players, is going to be rather bad compared to his performance in coordinated teams.

Also, he's a singletarget focused tank, so his teamfight presence isn't to high, he mostly exels in pickcomps For soloQ however, it's mostly either early sonwball, siege or deathball, as those need a lower level of coordination and mostly visioncontroll compared to pick or splitpush comps.

So i expect tham to perform better in competetive play than in soloQ, and given that experienced Tham already have a better winrate than Ezreal, atleast based on the data provided by champion.gg, he seems to do rather well right now given the circumstances(tham 56%win at 15-50games, ezreal 52.6% for 50+ games, even worse as adc)

I also remember that after Yasuo, who had a poor performance right after release, got buffed in the next patch, just to get continously nerfed in multiple patches after, riot stated to have learned from their mistakes and go with hasty buffs in the next patch if a new champ doesn't perform well after release.

So i think that it is to early to go for buffs, epecially buffs as bug as those fot tham on pbe.

And even if it shows after some more weeks that Tham indeed is a bit to weak, i don't think buffing his Q and especially R-passive damage is the right way to go. He already is pretty opresive 1v1, especially vs other melees. His singletarget damage compared to his tankiness is pretty damn high, and his singletarget cc also is quite impressive. Buffing the damage even more would make him a really poor play-against experience for other melees, and probably ends with him mostly going top for splitpush.

The main problem for him once behind is that he can't get to the right targets in a teamfight, he has no other choice than fighting the frontline if the enemies group, so his cc and damagepotential is focused on the 'wrong' targets. It would be better to increase his unique strengths. Improve the W targeting, you still often eat the wrong target instead of your intended target, or increase his AA range by 25 per ult rank, which would help him alot at applying 3 stacks to a target in a teamfight, as he often gets CCed or simply bodyblocked while his target gets out of range so he can't get the 3 stacks for cc. Reduce the CD his W recast, so you can eat allies and more or less imidiatly burp them to a saver location. Or increase his ms duerin W on enemie slightly, so he is even better for geting enemies out of position. Or give him Tenacity duering the E shield, so he isn't totally stuck within cc, unable to really use any of his kit in a big fight Or give him some kind of escape mechanism. The main reason he does bad if behind and also isn't as great toplane is probably that he is incredible vulnerable. He has problems fighting effectilvely vs multiple, coordinated enemies, and he has no way out of a bad situation. If he ever gets cought slightly out off position, he's done, no way out at all. Another place where he could get buffed is his ult. Righ now, it simply is way to telegraphed for a ganking tool, with also rather limited range early on. For me, right now it works more like a Reksai ult, helps a bit with controling sidelanes and still offering presence at dragon, or simply a tool to get faster back to farm after shopping, but actuall, ganking? Nope. With a faster travel time, his mappresence(and that of 1 ally) could get improved greatly. Maybe it was nerver ment to be a actuall ganking tool but more a global mobillity, and thats fine too, but still a more interesting place to explore buffs than his singletarget damage...

Maybe i have a different perception of his identity, but insane dueling power and absurd bully vs melees don't seem to be his indentity, but by buffing his damage numbers, especially the %hp ratio, this patterns will profit the most...

btw, thanks for the detailed posts, i really apreciate it, getting some more insight in the though process behind changes is something that i and probably many others from the comunity were missing lately.

And if your internal data differs from data provided by champion.gg and co, could you share a bit of it? I agree that balanceing shouldn't be data driven, but haveing some data to back up your suppositions sure comes in handy.

XonraJul 15, 20153 votes

[{quoted}](name=Leaf of the lake,realm=NA,application-id=3ErqAdtq,discussion-id=Hd6ZAtqq,comment-id=001000000003,timestamp=2015-07-15T21:12:29.280+0000)

Doesn't this suggest that he has a high skill floor, and that after spending that time to learn the champion, he may very well be balanced? If a champion is difficult to play, but balanced when a player finally learns how to use them, buffing the champ to make things easier for the new learners will just result in it becoming OP for those that have already gotten past that phase.

Sort of like how people thought Yasuo was harder when he was first new, then riot overbuffed him, then come to find out he was already really strong to begin with.

CupcakeTrapJul 15, 20156 votes

{quoted}

Where are you getting that number? Our internal data suggests win rate after 10 previous games experience is 45%. To hit 50%, you need an average of 30 games on Tahm. Usually you're at something like a 52-53% winrate after 30 games.

Is there any chance that sort of data could be made publicly available? I've been wondering about it for a while. It would be neat to see a graph like that for each Champion—a rather literal depiction of the learning curve.

I also think it would be interesting to look at Team Builder's Champ-specific MMR data. See how a Champion's specific MMR generally compares to that Summoner's average across all Champions.

RiotZenonTheStoicJul 15, 201513 votes

The data question is a very difficult one and frankly above my paygrade. It's very, VERY tricky to make this kind of data available. It's very easy to come to very wrong conclusions without sufficient context and understanding of the wider implications. We have entire teams full of scary smart people here at Riot helping us mortal designers interpret and understand these numbers, and we still get it wrong a lot. I'd be very concerned that if we dumped all this data on you guys there'd be strong expectations of, I don't know, "lol look at Heimerdinger's mastery curve, why the hell have Rito not nerfed him into the ground yet?" (And btw Rito, what's up with that)

CupcakeTrapJul 15, 20153 votes

{quoted}

The data question is a very difficult one and frankly above my paygrade. It's very, VERY tricky to make this kind of data available. It's very easy to come to very wrong conclusions without sufficient context and understanding of the wider implications. We have entire teams full of scary smart people here at Riot helping us mortal designers interpret and understand these numbers, and we still get it wrong a lot. I'd be very concerned that if we dumped all this data on you guys there'd be strong expectations of, I don't know, "lol look at Heimerdinger's mastery curve, why the hell have Rito not nerfed him into the ground yet?" (And btw Rito, what's up with that)

That's pretty fair. But what about a processed version? Maybe said smart people could assist with preparing some simplified graphs? I would read/share the heck out of an article on this.

ploki122Jul 15, 20156 votes

It's very, VERY tricky to make this kind of data available. It's very easy to come to very wrong conclusions without sufficient context and understanding of the wider implications.

Well... isn't that true for every single piece of data that's distributed though? Allowing people to look at others' ranked history could lead to wrong conclusions like how a certain person is bad on X/Y/Z champions, or builds someone wrong. Allowing sites to access peoples' division during matchmaking can draw the conclusion that the game is obviously skewed. Champion mastery can lead to the conclusion that someone is good on a given champion when he's in fact 0-125 on him up until now.

You don't grant access to data with the intention of giving out information; You do it with the intention of giving out data. If that data was out, "scary smart" community members could actually make more articulate posts having some Riot-sanctioned data, instead of having to use a 3rd party site whose accuracy is arguable. For instance, if I want to talk about Poppy, I will rely on that site to show how Poppy's win rate is much better than people think it is, and that her main problem is how her playstyle is unintuitive. You can also look at the game length data to make people realize that her early game is actually decent and it's the mid game that's troublesome. However, that only covers Platinum+, which means that you don't have anything to talk about her performance in pubstomp environments. You also can't look at win rates in ARAM, or play rate in ARAM (aka what champion is rerolled the most/least). You can't check 3v3 games' data to try and better understand the meta.

Now, this is a bit rant-y, and obviously not directed at the right person (you're a designer, not the leader of the team, nor a guy that works directly on that data, and this topic would've to be directed at the leader of the data team), but I simply find it annoying how often that "releasing data is dangerous because of how it can be percieved" argument is thrown around, and sometimes even by the people who are actually using that data and should know what they're talking about.

noodle0117Jul 16, 20151 votes

it's above his paygrade.

ploki122Jul 16, 20153 votes

I actually acknowledged that in the last paragraph... It's simply that this topic seldom comes to the surface.

kDrakariJul 15, 20152 votes

champion.gg has graphs like that. Tahm Kench's page is here. The graph labeled "Win Rate % by games played" is what you're looking for. I don't really know how accurate it is; there are a lot of champions which actually have LOWER win rates for players with a lot of games on them. But that graph shows an average 37% win rate in the first 5 games (which make up 85% of players) and it goes up to 56% in the 15-50 games range. Apparently nobody has more than 50 games played, which is why the graph goes back to the average after that.

champion.gg only takes into account Ranked games of Platinum or above.

Interestingly, Tahm Kench in top lane starts at 32% for the first 5 games, then falls to 14% by the time you've played 15 games.

CupcakeTrapJul 15, 20151 votes

{quoted}

champion.gg has graphs like that. Tahm Kench's page is here. The graph labeled "Win Rate % by games played" is what you're looking for. I don't really know how accurate it is; there are a lot of champions which actually have LOWER win rates for players with a lot of games on them. But that graph shows an average 37% win rate in the first 5 games (which make up 85% of players) and it goes up to 56% in the 15-50 games range. Apparently nobody has more than 50 games played, which is why the graph goes back to the average after that.

champion.gg only takes into account Ranked games of Platinum or above.

Interestingly, Tahm Kench in top lane starts at 32% for the first 5 games, then falls to 14% by the time you've played 15 games.

This pleases me greatly. Thank you!

I don't think it's weird that some Champions have lower winrates at higher numbers of games played. Remember that winrate is sort of a "derivative" statistic. It shows a rate of change, because MMR increases when you win. Someone with a constant winrate of 55% over 1000 games is not "stagnating", but steadily rising through the ranks.

Another way of phrasing that: it's not just number of wins over number of matches, but what MMR those wins are happening at. You could have two people with the same 50% winrate, one at Silver tier and one at Diamond tier.

This is part of why one stat I'd really love to see would be Team Builder Champ-specific MMRs, as they relate to the Summoners' overall MMR. For example, if the average Summoner's MMR with Pingu is 25% higher than their average MMR with all other Champions, that suggests that Pingu is strong. (That would need to be corrected for number of matches played.) I'd argue that the real definition of "OP" is "increases your MMR faster than other Champions", with the implication that this is the result of the Champion being super-strong rather than you being well-suited for that Champion.

kDrakariJul 16, 20152 votes

Oh hey, I didn't even notice I was replying to CupcakeTrap until I got the reply notice. Glad to help!

Chaotic SageJul 15, 20152 votes

I haven't actually played Tahm myself yet, but from what I've seen in the games he has been in (not a huge amount as I really like supporting so I tend to fill that role) he generally does exactly what you're saying here if he has at least got a few games under his belt prior. I would now like to thank you profusely for your response in it's depth of explanation and clarity. It actually makes me want to give him a try soon to see how the buffs effect the game play patterns you've explained that I didn't have the words for before. So, once again, thank you!

Earl EulrichJul 15, 20153 votes

http://champion.gg/champion/TahmKench suggests that winrate rises significantly for the 15+ games players. So yea, 10 is a tiny bit exaggerrated, but it just shows that all the champ needs is a little experience and certainly not a buff (except maybe the ult cd...for thee lackluster lvl 6 range it´s really high).

All of the above suggest to me that in the support case, higher base damages will disproportionately buff the losing / even case, which is in bad need of buffing, and will do less to the "we are already super winning" case, because in THAT case Tahm tends to exit combat with his high value ally.

two things here:

he´s a melee without mobility, so if he/his team fall back it will always snowball rather hard against him, as he´s easy to single out and he doesnßt offer much to defend sieges (neither good initiation nor waveclear nor sustain)
you tunnel too much on him as a support, he´s already really suppressive to play against as a melee - if these pbe changes go through he´ll be an abomination toplane in many matchups. The scaling might be alright for a support that doesnßt snowball, but with access to some more money he´ll be just broken

so i really would want you to wait some more time till everyone that plays him is a little more experienced (and doesn´t gut his teams dps in a fight by eating the savly positioned adc). If you then still feel he´s in need of buffs i´m fine with reducing the early ult cd (which by the way noc could use aswell) + buffing his q base - but please don´t buff the hp->dmg scaling, or you´ll create a hardcounter for any other melee in the game.

But I really want to thank you for sharing some of your data and analysis, it really enlightens me why you´d go for the changes that felt incredible random and unwarranted.