A Response to Doublelift on a Constantly Evolving Game

Doublelift posted a great video the other day talking about the rate of change in League. It’s worth watching. Since he referenced me in particular, I thought it was worth offering the Riot perspective here. My intention is absolutely not to try and refute him point by point, because we agree with a lot of what he says. But this way at least you all might have some idea of what our goals are and why we don’t always achieve them.

Costs of Change

First, we totally get that change has a cost. Learning is fun, and as DL says, one of the things that keeps players coming back to League. However, for this to work, you need to feel like there is a sense of forward momentum, not that you’re just running in place all the time because freaking Riot keeps resetting the game. I honestly can’t completely put myself into the shoes of a pro player, because I’ve never had to scrim 12 hours a day because my team and even career depend on it. When a pro feels like they let down their team because of a change we made, that sucks. When any player feels like a lot of the work they put into mastering a champion is reset, that sucks. That’s definitely not the goal, though sometimes it’s an unavoidable consequence of trying to meet the goals. (I’m also happy to acknowledge that sometimes it probably is avoidable.)

Overall, our goal is to make sure that our changes always move the game forward. An (unrealistic) vision I try to impart in our designers is to imagine that there is a perfect League of Legends out there (there's not), and every change we make should move us closer to that vision (even if we sometimes take one or two steps back along the way). We should not make change for the sake of change just to “spice things up.” Now, in all honesty it’s true that we do a little of this. New champions for example are mostly to make sure that players always have something novel to look forward to. But overall we want to move forward.

Goals of Change

So why change the game at all? As DL points out, it is really easy for multiplayer games to become stale if players solve them. This is particularly true today where streaming and video content are so prevalent that when pros or anyone else figure out an efficient solution, it quickly disseminates to everyone. Certain communities embrace this - just ask any Smash player. In that case, the game doesn’t change much at all and it’s up to the community to come up with any kind of evolution to the strategy / meta.

There is nothing wrong with that approach. It’s just a different approach. The approach that Riot has taken is to constantly evolve the game. Riot was founded back when Tryndamere and Ryze got frustrated with an RTS or whatever it was they were playing, in which all the players knew there was one winning strategy and if only the freaking developers would make a few simple balance changes, the game would have so many more options and all those dead units would actually get used. There are League players (and not just pros) who play many, many hours of League every week. If they go up against a frustrating champion or a degenerative strategy every game, they may have to suffer through dozens of games before we fix it, and that’s assuming we fix it for the next patch within two weeks and get it right the first time. Often we don’t quite fix it enough, or potentially over fix it, and that requires even more suffering for players.

Recent Changes

The 2018 season has been a period of very high change. We started with Runes Reforged and followed with Jungle and bottom lane changes that were more disruptive than is usual, and frankly, than we intended.

While overall we are happy with our progress towards the goals of the runes overhaul (and yes I do want to acknowledge that not every player is happy with the overhaul) we knew we were going to have to smooth out large parts of the game after launch. We had necessary followup on the balance front, like Stopwatch, and the design front, like Conqueror, which meant players were constantly having to relearn parts of the system. But runes would be worse off today if we hadn’t made those changes, and we think League will be better in the long run as we polish Runes further.

We also tried bigger even-numbered patches in the first half of the season to allow us to solve larger problems, such as the AP item refactor, that was more change than we have done in the past few midseasons. More recently a lot of this year’s pain has come from midseason stretching for 3 patches with gameplay shifts on each. Ironically, the goal was to smooth out the launch of this midseason's changes, since when many huge changes hit at the same time, it can be hard to tell which problems will clear up over time, and which ones we need to actively fix. For many players, though, the experience wasn't "midseason parts 1, 2, 3" so much as "midseason 1, midseason 2, midseason 3." That cost ended up being higher than anticipated, which is a big part of why we decided not to ship the new fighter items we had slated for patch 8.12.

Looking at the next few months, patches in the run-up to worlds and preseason will be much less disruptive. We're not going to continue with big patch / small patch. In addition, preseason will be a single patch rather than the spread out approach we took with midseason. We're planning to stick with the single-patch approach going forward, too.

Inadvertent Effects of Changes

Also remember that while our changes are intentional, the impact of them isn’t necessarily so. We have a really strong development team who thinks deeply about their goals and the changes they want to bring about, and then we playtest the crap out of the changes we make and only ship the best ones. But I’ll be the first to admit we get it wrong sometimes (and I apologize for when we do).

We wanted to make some changes to ADCs for sure. We didn’t want to remove Trist and Cait and Rekkles from the game. We actually thought it might take larger changes to shift some of the power that ADCs hold over the bottom lane, and thought it was more likely we didn’t go far enough. After a much bigger shift than expected, however, we’ve shifted focus from continuing to open bot up, to making sure we didn’t take too much power away from ADCs. One of the benefits of our patch cadence is that when we do miss the mark, we have a chance to correct it, and hopefully ADCs are heading towards a better place now.

Competitive Changes

The competitive season in general is a big challenge for us. We certainly don’t balance the game only for pros, but as DL points out, pro play is obviously a huge draw and we want to make sure there we can provide a good viewer experience for the major LoL esports events. Unfortunately that often gives us a really narrow window in which to make large changes - pretty much preseason and midseason these days. It’s a tense situation when we have a problem in the game that we want to fix for competitive gaming and (a) risk making the change so late that pros have no opportunity to adapt to the change, or (b) risk not making the change and having a less interesting viewer experience as a result.

Last year, when the Ardent Censer meta emerged, both Riot and pros weren’t entirely sure what it was suppressing. The meta that might have emerged with a weakened / removed Censer could very well had been worse than the Censer meta, and pros at least had a lot of time to build strategies around Censer. This is the risk that can happen with a slower rate of change: something really comes to dominate. We had already patched Censer, but pros were playing on an older patch, again, to reduce the amount of “rules churn” they had to contend with.

We’re in a similar spot today with gold funneling. We’re concerned the viewer experience around gold funneling isn’t great, and while we hoped at first it would be a strategy you’d see sometimes, it looks it may very well come to dominate. Patch 8.14 does have a change meant to curb funneling, but we acknowledge that means we’re forcing pros to adapt to a change well past midseason.

Champ Changes

I could write a whole post just on large champion updates (VGUs). We owe it to players to fix champions who aren’t interesting to play, who lack counterplay, or who just feel outdated. But we know there is a big risk that we end up losing some of the dedicated players of that champion if we move them too far away from their original state. Sometimes we think it’s still worth doing for the sake of the game. Loss aversion is a very real thing, but this isn’t a single player game where it’s okay to let someone have an overpowered character just because it’s fun for them.

I wanted to give a shoutout to Voyboy’s Tweet where he talks about the cost for other players when your champion is a little OP, as well as the challenges of buffing instead of nerfing.

Aatrox himself is a larger topic that illustrates many of the opportunities and challenges of champion updates. I considered going into it here, but this is getting long already, it’s probably worthy of its own discussion.

Thanks for reading. A video response to Doublelift would have been easier to sit through, but we didn’t want to delay our response any longer. Thanks to all of you who take the time through videos, streams, Reddit or whatever to point out where you think the game or our process can improve. It’s not always easy to hear of course, but overall we think it’s one of the reasons League is still around after almost a decade.