Boards›Gameplay

Let's talk about Riot's nerf/buff philosophy

kcStranger··6 votes·1,675 views

Before getting into this, I should make two things clear:

  1. The main purpose of this post is not to call into question any specific nerf/buff decision that Riot has made, but rather to take a more global perspective and question the philosophy that underlies these decisions. Of course, I will evaluate certain nerfs/buffs as examples, but that isn't the main point.
  2. I will refer to data involving champion win rates in ranked solo (from LOLDB/Lolking; mostly I like to focus on data from gold to diamond since lower-level players may not be a good indication of actual strength) as well as how much pro attention a champion gets (can be evaluated using sites like probuilds.net, as well as watching pro team play). My proposition is that the first of these gives some idea how well a champion fares when teams lack coordination and performance is limited by the champion's skillcap, while the latter better indicates the champion's true potential. I believe that being weak in both areas is a sign of being truly underpowered.

My main point is this: I believe that Riot would be well-served to decrease significantly the magnitude of nerfs or buffs applied to a champion in any given patch. Right now, Riot's philosophy seems to be to try to "hit the nail on the head" with every nerf/buff and get the champion to the state that they want it "in one go." However, this frequently leads to overshooting the mark. For instance, the patch-5.3 nerfs to Azir dropped his ranked solo win rate to ~38% and caused him to evaporate from pro play, whereas he was very strong before the nerf. My guess is that this mostly has to do with Riot underestimating the impact of the range nerf to his W (soldier spawn), though I'm not an Azir player so it's hard to comment. Similarly, patch-5.1 nerfs to Syndra dropped her winrate to ~42-44% and caused her to mostly disappear from pro play as well. I am a Syndra player, and I think the biggest deal was the nerf to the hitbox on her E (the stun). Nerfs to Q (Dark Sphere) hurt her early-game bullying, but the difference between hitting a stun and not hitting it is massively impactful throughout the game. In general, I think that it is easy to underestimate the impact of range/hitbox nerfs; while they don't decrease the champion's maximum potential damage/CC, the difference between hitting an ability an not hitting it is game-breaking. Also, I think it is very important to be conservative when nerfing abilities that are core to a champion's identity. For example, it is necessary to be more conservative when nuking the range of Azir's soldiers than when nerfing his tower nuke damage.

A better example of how Riot should nerf champions, I think, is what they did to Lucian. He was given small to moderate nerfs in 3 patches. (If I remember correctly, the first was to his E mana cost, then his R cooldown and W AD ratio, then the early-game damage and timer on his passive.) The last of these nerfs was probably the most severe, but none of them was individually overwhelming, and doing the nerfs a little at a time gave Riot the ability to evaluate the impact of their previous nerfs before committing to the next set. As a result, Lucian is in an OK place now: his win rate is solo queue is sub-50%, but not by too much, and he still sees some fringe pro play (mostly as a backup Graves). So, if you put a bunch of effort into mastering Lucian, you'll certainly be a bit upset, but at least you're not an idiot for continuing to play him.

If Riot adopts a more conservative approach to buffs/nerfs, it is likely that they will have to buff/nerf the same champion multiple times over 2-3 patches to achieve the effect they want. However, they will greatly reduce the number of incidents where a champion becomes broken or unplayable overnight, and avoid angering all the players who have devoted a lot of time and effort into learning a champion. I, for example, play a fair amount but not a ton (~2 games/day on average), and it typically takes me a couple weeks to pick up a champion. That's a lot of wasted time if my champion gets nerfed to the ground.

To apply this to the recent patch, here are a few sets of nerfs I'm concerned about. First, the Kassadin range nerf (reduced from 700 to 450). Although Kassadin was dominant in pro play, his ranked solo winrate was only ~49% (including at Diamond). The ability to Riftwalk to escape/dive is core to his identity, and they just nerfed it by 36%. That's huge. I'm withholding judgment on whether the nerf was fully justified for the moment, but I will claim that a better approach would have been to first nerf to perhaps range 600, then re-evaluate before committing to further nerfs.

Second, the nerf to Jarvan IV's E passive (bonus armor). Before the nerf, E granted Jarvan 10/13/16/19/22 bonus armor; now, it grants none. This has been very incompletely compensated by an increase of 3 to his base armor. The patch notes state "this is definitely a nerf to Jarvan's mid-late game survivability," but doesn't note anything about the fact that it's a huge nerf to his jungle clear! Before the nerf, Jarvan players would typically skill E first to get extra armor for their jungle clear, so the nerf effectively reduces Jarvan's armor by 7 for early jungling. This is actually a huge deal; it almost completely negates the armor that Jarvan earns from runes (typically +9 from 9 armor seals). Again, a smaller nerf would be more appropriate until they have time to evaluate the impact. For instance, if they really care about nerfing Jarvan's mid-late survivability, they could just decrease (or even remove) the scaling on the E passive, by changing it to 10/11/12/13/14 or a flat 10. As for their question "Why would a flag give armor"--well, I'm not claiming that's their actual reason for the nerf, but still: Gameplay should always come before flavor.

11 Comments

Sir ArmaMalum3 votes

First off, thank you, I appreciate your opinion and while I disagree, it's not because I believe your logic as flawed but because I believe that your solution cannot be used in every instance.

Two points:

One) Drastic changes, specifically mechanic changes (i.e. 5.4 Veigar Stun Delay), are not a scalable change. In other words, there is no set formula for trade-off between a delay on Veig stun stun for AP or CD on his other abilities. You have to try to 'hit the nail on the head' because here are bare minimums (like introducing a noticeable, react-able delay) that must be met for changes to have enough impact for compensation elsewhere. For example, if Veig stun delay was introduced as 0.25 sec one week and increased to 0.75 over time there wouldn't be a good way to compensate until, say, 0.3 seconds because any shorter delay isn't easily reacted to and people will say nothing effectively changed. If that makes sense.

Two) Gradual or no, changes will piss people off. Worse, if you give them warning and a gradual introduction to the changes, the people who don't like the changes will be exponentially more pissed because they're continuing the changes nonetheless.

kcStranger3 votes

Yeah, I agree that this doesn't work in every instance. It's especially true for sweeping changes like the new season-5 jungle. It's also true that no matter what you do, someone is going to be pissed off, and some people are likely to be like "WTF they nerfed by champion 3 TIMES." However, I think there would be two benefits to making changes more gradually when possible:

  1. You would actually do a better job balancing champions. You could gradually approach the equilibrium you want rather than continually overshooting in both directions.
  2. You would largely avoid the worst-case scenario: the nerf into oblivion. People are going to complain whenever you nerf their champion, but if they still like that champion and the champion is not unplayable, they still have the option to (and often will) keep playing them. When you nerf a champion as hard as they nerfed Azir and reduce their winrate to 38%, it's just not reasonable to keep playing the champion, and you run a very real risk of alienating the people who devoted countless hours to mastering such a complex champion.
Sir ArmaMalum3 votes

Good points.

You would actually do a better job balancing champions. You could gradually approach the equilibrium you want rather than continually overshooting in both directions.

I agree, but at the same time I'm also going to argue that's what's been happening over time. And the reason it never seems like it is because we're dealing with a literal world of difference when it comes to playerbase skill, and skill level is a major meddler (aha, aha, okay I'll stop) in how well received changes are. For example, a small incremental buff to Riven would be dealt with frothing at the mouth by lower tier players, but higher tier would simply shrug and move on because they will most likely know to keep the Riven shut down and her snowball nature in check. Inversely a nerf to, say, Blitzcrank would be met with praise by lower tiers, but be met with confusion by higher tiers. Very generalized examples sure, but I hope that explained the point.

You would largely avoid the worst-case scenario: the nerf into oblivion. People are going to complain whenever you nerf their champion, but if they still like that champion and the champion is not unplayable, they still have the option to (and often will) keep playing them. When you nerf a champion as hard as they nerfed Azir and reduce their winrate to 38%, it's just not reasonable to keep playing the champion, and you run a very real risk of alienating the people who devoted countless hours to mastering such a complex champion.

I do agree with you that during a season I would much rather balances be met with a side of caution instead of the other way around because under-nerfing/under-buffing is a better mistake than over-nerfing/over-buffing. Pre-season I'm much more 'meh' on that as the competitive scene is in flux anyway.

However, to play devil's advocate, how would this apply to champion updates that involve the entire kit and/or new champions? Naturally these guys can't just be bundled into pre-seasons, and they won't be perfect coming out. PBE or no, they will shake up the game a lot simply by existing. Should they intentionally under-tune the new/updated champion or gradually introduce them in a way such as disabling from ranked for X days?

kcStranger1 votes

I think we agree for the most part now. As far as your last question, I think that their present philosophy is currently to aim to slightly overshoot on new champions, so that people are excited to play them when they first come out. Honestly, I'm ok with this. It makes sense from a business perspective. Moreover, the argument I'm making simply doesn't apply to reworks and new champions. There, I think, they should just do the best the can, along the lines of what they're already doing. Just be more careful with subsequent tuning.

Mandang01 votes

I agree, and in an ideal world, this would be the way to go.

I think the practical concern, however, is whether Riot has the manpower to keep track of that many small balance changes over that many patches for so many champions. If something there gives, it's likely the timeframe, which means we're left with certain permaban champs for years on end -cough-Kass-cough-

If it's the choice between this or nerfing Zed's attack speed, I'll take this.

kcStranger1 votes

I agree, the number of people who are devoted to champion balance is a problem, and it's probably difficult for them to financially justify much more than they already have. I don't think implementing this would be that big of a deal, though. For instance, take Kassadin. They clearly think that a nerf of his Riftwalk to range-450 is appropriate, and it probably is correct to try out a nerf of this magnitude on the PBE. Instead of implementing the entire nerf this patch, though, they could first nerf to 550-600, then re-evaluate. If the champion is still too strong, they could complete their already-planned nerf to 450 a patch or two later. It wouldn't take more than an hour of work for one person to look over some statistics and verify whether the champion needs an additional nerf, and they're already basically doing this work anyway by checking up on the impacts of over-nerfs. (For instance, note the fact that they've partially reverted the nerf to Syndra's E twice now.)

Mandang03 votes

Yeah, but the flipside of this argument is Kassadin's balance over the last 2 years. It's not without any balance changes, but the ones Riot picked (and sometimes later backed down from) didn't accomplish much, and they came very infrequently.

The situation only gets worse with each champ Riot releases OP, while refusing to increase the ban count above what it was when the champ pool was half this size. I'll say relying on bans to cover up your balance shortcomings is lazy, but at least it's something. The fact that the same set of champs can remain in the most-banned for so long speaks volumes (IMO) to how slow Riot's balance usually is - something that seems much easier to fix with impactful changes than with a larger number of Zed attack speed nerfs.

Lastly, because I think your point stands on its own, but I want to talk to your specific examples: These 2 were long overdue for nerfs, and could stand to take a larger nerf than most. There's a mass of Kassadin threads calling Riot out for hitting his "core identity", which Riot's said they wanted to avoid doing in the past. I think the reason they're doing it now is because Kassadin has proven a persistent enough problem that they actually need to just hit the mobility. Jarvan's problem is a bit more subtle, but we saw it when J4/Vi/Xin was the trinity of jungles back in league of warmogs - they have enough always-relevant steroids in their kits that they can build full tank and still be a damage threat. I guess I'd be a bit surprised to see Xin nerfed, but I expect a Vi nerf's in the pipeline.

kcStranger1 votes

You're not wrong. You may also be right about the specific cases of Kassadin and J4. (Again, I'm not attempting to argue that the nerfs are definitely overkill, just that they're overkill all at once.) I don't really like the idea of increasing the number of bans because of how random bans can sometimes be in solo queue (due to the lack of team coordination/agreement about bans), but I agree that they're necessary and serve as a band-aid for balance issues.

The only thing I might encourage you to do is look at Kassadin's winrate numbers in ranked solo. He was already sub-50% from Bronze to Diamond before the nerf. Now, his dominance at the pro level (and very good showing in Challenger) indicates that this is probably a skillcap/coordination issue, so I'm not claiming he isn't perhaps a little OP, but it's worth considering the fact that changes that make him "balanced" at the pro level will likely leave him significantly underpowered at most other levels.

Mandang02 votes

{quoted} but it's worth considering the fact that changes that make him "balanced" at the pro level will likely leave him significantly underpowered at most other levels.

I think this is a separate problem that's always going to arise when you design champions who are only countered by things you can find in one play mode but not in another (say, coordination that you can find in 5v5 but not solo queue). When push comes to shove, I expect Riot will continue to choose to balance for pros/5s rather than solo queue, which I might disagree with, but it's a bad choice to make either way. Much better is if Riot can fix a champion's design (or core identity) to not present such a situation. I think Kass's extreme mobility will always be a problem for that reason.

For the rest of your post, I think you need to also consider ban/pick rate when trying to use winrate as a measure of strength. I think Heimer currently has one of the highest winrates, but I wouldn't say he's a balance concern for that reason because of his low pick rate. Likewise, in the past we've seen Orianna have a low winrate in solo queue, yet be heavily contested in pro play & used well in the higher tiers. Personally, if I had to pick a metric to balance from, I'd pick banrate. Even if something's not necessarily being banned because it's too powerful, it's something your players don't want to play with, which is something a game should address.

kcStranger1 votes

Fair enough. Another notable thing about Heimer's winrate is that it's significantly lower in Diamond than in Gold, for example (although still fairly high in both), indicating that properly counterplaying Heimer is possible (albeit difficult) even without the extreme coordination seen in pro team play. It's possible that ban rate at high levels (Plat and above I'd say?) may be a useful statistic to consider, although of course it will probably skew towards champions with strong laning matchups instead of, e.g., some random Support, even if the Support is stronger overall.

As for balancing for solo queue vs. pros, I think part of the issue is that if you balance mainly for pros, the imbalance required for one champion to start dominating is very small. Two champions may be very nearly the same strength, but if they have a similar playstyle, the pros will lean towards whichever one is marginally better and ignore the other one (e.g., Graves vs. Lucian). Achieving the level of balance necessary to have more than 30-40 champions viable at the pro level is almost impossible. So, the best you can do is enter a buff/nerf cycle such that different champs are stronger at different times. This approach succeeds in keeping the pro scene (i.e., the game for viewers) fresh, but can make things very unsatisfying for non-pro players, as I've argued above. I wish they would skew a little more towards balancing for non-pros than they currently do, but I won't argue that a small amount of strength cycling isn't necessary. Just, avoid the nerfs-to-death please.

kcStranger1 votes

I sincerely doubt this post will get much more traffic at this point, but I did want to add one follow-up concerning the specific cases of Kassadin and Jarvan IV. I just watched Riot's patch rundown, and I do think their arguments for nerfing Kassadin and Jarvan--in particular, the lack of clear avenues for counterplay--make a lot of sense. Still, I hope they will provide buffs in other places if the nerfs prove more significant than expected.

I still stand by the core point of this article, though it's possible that a few recent major goofups--mainly Azir and Syndra--have led me to underestimate how good of a job Riot is already doing.

Update after seeing the actual stats on these champs: The Jarvan nerfs are fine. I was more right than I even expected about Kassadin, though.

More from Gameplay

3 Refunds per season, upvote if you want 3 refunds per season.
Gameplay764 comments
7,167
votes
3 Refunds per season, upvote if you want 3 refunds per season.
Gameplay156 comments
3,510
votes
UPVOTE FOR URF
Gameplay606 comments
2,440
votes
Yes plz.
Gameplay456 comments
1,721
votes
3 Refunds per season, up vote if you want 3 refunds per season.
Gameplay114 comments
1,716
votes
When is ANYONE at riot going to address the awful new player experience?
Gameplay634 comments
1,383
votes

More by kcStranger

OK Nidalee Haters, Let Me Try One Last Time
Gameplay194 comments
91
votes
4 out of 5 Bjergsens think Syndra could use a mana buff
Gameplay41 comments
40
votes
*This* is why you have a support
Gameplay76 comments
35
votes
Meditation/Expanded Mind Changes and Mid-Lane Mages
Gameplay28 comments
24
votes
Syndra public opinion poll
Gameplay63 comments
19
votes
Dear Rito, pls give us more mana (utility masteries discussion)
Gameplay5 comments
13
votes