Pwyff

chris tom · @Pwyff

6th Oct 2016 from TwitLonger

Why so long? A year in dynamic queue.


Just some personal thoughts, this will definitely meander.

As you may have heard, we're bringing back solo queue (http://na.leagueoflegends.com/en/news/game-updates/competitive/riot-pls-ranked-pls-2017-season-ranked-changes).

After seeing the conversations play out, there are two recurring questions I've been thinking about. The first, "Why?" (as in, why do this?) I answered on reddit (https://www.reddit.com/r/leagueoflegends/comments/561hpq/ranked_2017_changes/d8fhgn5). A summary of which would be: we finally realized ranked means something very different to a lot of solo/duo players, and while dynamic queue supports a clearer vision for organized, competitive League, it also dilutes the original promise of solo/duo ranked.

A few have argued both queues are equal measurements of skill (dynamic vs. solo/duo), but the point *isn't* that one is better than the other: it's that this year, your rank was impacted by something different than all previous years (organized teamplay as an additional skill). If I take your childhood teddy bear and replaced it with a newer, plushier one, are you happy? What if it does your taxes?

The point is, players grew up with solo/duo ranked for 7 years, and while the ladder names may change, the *achievement,* the cultural recognition, the personal pride, and the individual competition have always remained the same. Dynamic queue swapped out your teddy bear.

The second question I hear a lot is, "Why did it take a year for you to hear what I've been saying the whole time, you fucks."

The short answer is an apology: sorry.

The long answer is: we heard you from the beginning. We acknowledged it in the first post Lyte made in January (http://boards.na.leagueoflegends.com/en/c/developer-corner/XvqbAgrF-letos-talk-about-dynamic-queues-and-ranked-play). We said there were critical flaws in execution we needed to fix, and a philosophical issue (above) we needed to evaluate. I could make a half-joke that it took us 9 months to evaluate, but that's flippant. The issue wasn't that we didn't hear you, it's that we were hearing a lot of things.

Can I generalize? Reddit is a self-selected group of highly passionate players, all of whom are ranked Diamond or above and have played thousands of games on every champion to date. They have a pretty significant internet crush on Imaqtpie and Rush. They used to love lane swaps, but now they hate them. They have many ideas about champion balance and, on occasion, they're right. On other occasions, they say *this* is the nerf that really guts Kayle. This never happens. I suspect reddit indexes higher on hardcore solo players than a general demographic slice of League.

I don't say this to dismiss reddit or the forums (or inven, or any other counterpart), but I think it's important to acknowledge these hyper-competitive communities have certain biases and we have to account for them in the feedback they give.

The reason I highlight this is because, in the first few months of DQ's launch we also saw a 'silent majority' playing more ranked, saying they liked dynamic queue in surveys, being more positive in game, playing in more premades, and so on. You can call me a liar on the data but jesus christ for what reason would we break such a fundamental trust. In the hundreds (maybe more) of heated internal debates we had at the office, we all had to acknowledge this silent majority. To some, things felt wrong but the numbers were there.

So. We had a silent majority against our vocal minority, one that has a generalized bias of being hyper solo-competitive. I'm not excusing the speed at which we listen to critical player feedback, just giving insight.

Whenever we see a large disconnect between what we think of as reality vs. perception (e.g. whether or not Kayle is balanced [she's not]), we attribute it to a communications issue, and so it becomes my issue. While our engineers and designers were spending months fixing queue times and rewriting matchmaking algorithms, we had guys like Cactopus trying to figure out how the fuck we could move the vocal perception to what we saw as reality.

Instead, reality shifted to perception. Well, to reality. Whatever.

Where originally we were confident in making the tradeoff of some perception for more organized competition, player belief in the legitimacy of ranked as a measurement of personal worth has not stopped sliding and now, at this end-of-season juncture, we're unacceptably further away from the mark than before.

So a decision was made.

In the end, we didn't suddenly come to the conclusion that DQ was 'less legitimate' than SQ, we just realized that player perception was forcing it into a reality. This is the same kind of risk, on a much larger scale, when we take on champion updates - how much do we account for biases? How change averse is the audience? Did we get it right? How do we gauge failure? If we flub the landing, can we fix it with iterative improvements? What if it just takes time for players to accept it?

We mistakenly pushed forward with what we felt was a positive change but it turned into a divisive change which turned into a cultural shift and it's *that shift* that has made us pause. League's strength - its identity - is purely distilled from the hearts of the community, and while becoming a team sport through the generations is still a vision, it probably shouldn't come at the cost of its cultural heritage - it should build on it.

p.s. This is also why I enjoy working here. We may not get things right the first time, but by the third Ryze update I'd argue we're a little closer.

p.p.s. I genuinely hope flex and solo/duo queue survive because I think they're both legitimate sides of competitive League of Legends. Even now I get a lot of messages from players who loved dynamic queue and their sadness at the implied illegitimacy of their achievements this year. The team tried to correlate premades with MMR gains and found little to no overall statistical evidence. I'm not saying there are no 'boosted animals' out there but maybe it happened less than everyone believes.

Reply · Report Post