Exactly. Specifically talking about this example, it highlights the complexity of all the possible permutations:

• if the Nemesis had been less popular, this bug likely would not have occurred. It was only *because* players actually like the Nemesis that this was an issue.
  • if we'd also put fewer Nemesis (Nemeses? Nemisises?) on the floor, it might not have happened; e.g., if the Nemesis was a crate weapon, this wouldn't have shown up.
• if hardware spec and player skill were more decoupled, this bug likely would not have occurred. It was because high-skill players overwhelmingly tend to also have high-end machines that run at higher FPS that we saw this bug.
• if the gameplay dictated that squads died faster, this bug likely would not have occurred. Because high level matches tend to be decided much later on, when players tend to be fully kitted in the gear they actually want, most players who wanted a Nemesis actually had one. If mor...

the level of accuracy in this example is astonishingly high.

Yeah they do scale. The bonuses are tied to placements, and indirectly, the value of each kill scales with placements as they currently do. With that said, the math is not the same as it is now, and is dramatically more convoluted to explain.

One point of personal concern is the tuning for elimination bonuses, we try to be generous with points in the bonus system, but if climbing with mostly kills is proven to effective strategy, we'll continue to lower elimination bonuses.

Just want to say that, having been away from Reddit for a bit, this thread is awesome. Great to see so much constructive feedback, thoughtful criticisms and concerns, and - of course - the compliments and optimism. I really do feel lucky to work on this game and to have been one of the engineering leads on this particular project. Some nerves, of course, about something so significant. But just wanted to say thanks to everyone in here whose posted for having such a real - and nuanced - discussion about the new system and what we're trying to do.

first time we're applying to an entire mode of the game. It's been sandboxed to certain regions for testing - as we discussed here https://www.ea.com/games/apex-legends/news/matchmaking-2023 - but this is the first time we're using it as a core pillar of how a mode in constructed.

While the MMR system is newer and very different from the Ranked Arenas one - and definitely learns from it, it's also older in some important ways. We started data collection and analysis before Arenas even launched as a mode and have substantially more data (both historically and just from a pure quantity of information) on how people played ...

Well, to be fair, I wasn't a Ranked Designer at Riot

People don't go into pubs with the same mindset as ranked.

You can read more about matchmaking work in this blog we posted a while back: https://www.ea.com/games/apex-legends/news/matchmaking-2023

this seems like all the ranks are going to be officially mixed up now since its going to be matching with mmr instead of ranking.

Ranking is tied to MMR.

Under the hood, there are MMR requirements for each rank that we didn't discuss in the blog (and will be obvious as players reaches their skill peaks) .

eg. Players in Diamond need to perform at a certain level. Unskilled Players attempting to creep up their LP via ratting (without interactions) will struggle as the matchmaking system challenge them with more difficult matches that are appropriate with their rating (which these players will likely lose and drop in LP). In that same situation, the bonus system will detect players trying to rat and withhold unearned bonuses. Their progress upward will dramatically slow down, purely rating is suppose to be an ineffective strategy at climbing.

It's a new MMR system design specifically for Apex's Battle Royale, and we are using a different matchmaking system from Arenas. We took learnings from Arena when designing this new one.

Yes. Same ruleset as previous. You should land somewhat near each other if your MMR are about the same.