TL;DR - 343 needs to reset everyone’s CSR and also their hidden MMR because it’s all being affected by bot games and that probably explains some of the problems.
I have been wondering for a while why CSR (visible skill rank) and MMR (true skill hidden rank) feel so mismatched. For example, I am an Onyx 1500-1600 player. I almost always dump on anyone in Diamond, and anyone Onyx 1700+ almost always dumps on me. However, every once in a while I will see an Onyx 1600-1800 player who is absolutely terrible, or a Diamond player do inexplicably poorly like go 1-17. Turns out, a big part of that is possibly that the game uses your performance in the bot playlist to help determine your skill in all the other lists, as evidenced here:
I don’t know if this is the cause of all the wonkiness in the skill based matchmaking and ranked matchmaking systems, but this should have never passed one round of a QA test, let alone get into the live game. It needs to go and it needs to go now, not in season 2 in May.
They need to reset both CSR and MMR and they should do it in the February patch. Playing against robots - especially the awful marine skill robots - does not say anything at all about a player’s ability against other humans.
I am sure at some point 343 plans to reset CSR. But they also need to reset MMR, which I doubt they plan on doing. Trying to bring more attention to this so that perhaps they listen.
The system is trash. Once you’ve played enough games you’re stuck forever once the system decides where you belong.
You can’t hope to go up in Solo Queue once that happens because the matchmaking will purposely make you lose. This whole thing has 0 competitive integrity.
Oh yeah and in social, you’re either bronze or you’re Onyx.
I don’t believe the MMR exclusively tracks damage and kills. I believe it also tracks things like whether or not you pick up power weapons, whether or not you’re playing objectives, the time it takes you from spawnto engage another player, etc… Essentially it tries to create numeric data points to represent your “map and game knowledge” in addition to your actual game ability.
So might you be diamond in regards to aiming and damage dealt? Maybe not. Could you be matching diamond players in terms of knowledge of game type rules, weapon spawns, and map knowledge? Maybe.
I’m not arguing that bot games SHOULD effect MMR, I’m just trying to give you context of what might be going on for you.
For example, assume 2 players enter their first PvP matchmaking game at the same time. Player A has grinded the bot playlist a bunch, player B has only played campaign, or previous Halo MP, or has no Halo experience. Player A will likely drastically outperform Player B in the first (couple) game(s)
Player A gets to go up against Onyx players in their ranking matches and ranks at Diamond 1. Player B likely gets to play Bronze and Silver players in their ranking matches and ranks appropriately.
This is absolutely ridiculous that this would be in the game.
Yea, I read it yesterday. Like i said I’m not arguing that anything shouldn’t be done, I’m just giving you context for why the system is doing what it’s doing. I’m not even arguing that it’s perfect or shouldn’t change .
I think the point about TrueSkill 2.0 not just rating on kills, deaths, and damage alone is worth considering though.
It wouldn’t really be fair for Player A to be matched against a bunch of Player B’s, even more so than Player A being matched against “better” players than themselves imo though. At least Player A would have enough map knowledge to still maybe grab power items and play the objectives and benefit their team some.
I understand MMR is more complex than KDA. That’s not relevant to this topic.
Playing against bots is not comparable to playing against humans of any skill level whatsoever. It does not test the same skill set for a player.
I understand that on the surface that may not be apparent, as it might look like both test basic skills like “can you aim and hit a target?”, but that’s not really true. Bots, in addition to being absolutely terrible, are fundamentally different in their play than humans. You can learn and abuse their behaviors and they don’t really ever adapt to what you are doing. Players, even at low skill levels, adapt and play creatively and unexpectedly; this results in dramatically different gameplay even if on the surface it looks like both require the same thing of “aim gun, hit enemy target”.
Absolutely nothing about bot matches should affect MMR for PVP playlists. If they’re going to track MMR in the bot list, they should just use it to match you against bots of varying skill level so that better players are playing ODST and Spartan bots.
This is how TrueSkill 2 is designed to work. They just disabled MMR for Bot Bootcamp. I don’t see why it matters who you get matched with I’m that playlist.
Long term, I hope the MMR gets changed so one playlist doesn’t impact any other at all. That might take more work, but I think it’s necessary. I mean, if you crush it in Ranked with a team of four well coordinated players it doesn’t mean your good at FFA. If you’re great at objectives it doesn’t mean you can play Swat. It doesn’t make sense that MMR is crossplaylist in my opinion.
I think you should go back and read all the content in my posts and not just skim them. My points about map knowledge, weapon location knowledge, objective rules knowledge, etc are all what I’m saying is impacting your MMR, nothing you keep pointing back to.
And all those things are the same vs players and bots, so the practice does directly correlate. The Odball wil always spawn in the same place, and the game impacts your rating based off things like how long it usually takes a player to get to an objective off spawn.
1st time players won’t necessarily know where to go, and have a longer average time from spawn to objective. Players who have practiced against bots will move directly towards objectives right at spawn.
That’s the type of stuff that’s boosting your MMR.
I read your post. I just don’t agree with you. I do not agree that any of these skill sets applied to bots are the same as applied to human opponents, as per my post above. I do not agree that a player is tested on the same things vs robots as vs humans. Not unless the robots are powered by a creative/learning/adaptive AI, which these are not.
Very simple example: a bot will never try to go hide in a corner and camp if it’s getting dumped on and going 1-10. Bad players might.
I can understand maybe a base MMR level it starts you at in all playlists if you’ve played the game awhile, which is higher than a new player. But it should not impact it this much. Everyone in high level play is going to know all this stuff, so it doesn’t make sense than your Bot Bootcamp would impact MMR at that level.
The stuff you’re talking about might be the difference between a bronze and a silver player, which is to say, it has very minimal effect. And this effect is completely negligible at higher levels.
.