So they’re updating the broken ranking system - yay!
Are they going to stop social matches affecting ranked MMR? No? OK…
But at least bot matches won’t affect ranked MMR, right? What? They didn’t fix that either?
Maybe they’re improving game stability, crashes, and cheater detection? Huh? Not any of that either?
Oh… They just made sure everyone gets bumped down a CSR tier. Got it. Yup, they sure fixed Halo Infinite’s ranked mode problems. I know the main reason I stopped playing ranked was because I really thought I should be diamond instead of low onyx and I was mad the game ranked me too highly.
It’s not a simple case of just taking a bunch of CSR points off everyone. But that’s the bit that is going to upset the masses - so they got out onto the front foot early.
I think it’s safe to assume that the various metrics and weightings will be tweaked to help constrain the population back into a more realistic (normal) distribution of skill. That’s the point of them. But I can see why they don’t publicise the actual values.
Will that include bot matches into MMR? Maybe. But even if they don’t the seeding MMR only affects new accounts and is short lived (placement) so no real harm is done.
That’s not how it currently works. You can try it for yourself. Go play a few bot games going 0-10 with 1% accuracy ratings and you can guarantee yourself at least one easier ranked match afterwards. The system auto-corrects for that very quickly (you can’t sustain smurf rank levels without repeatedly sandbagging in bot games), but the game is constantly evaluating and adjusting your overall MMR no matter what playlist you are in, rather it’s ranked, BTB, Fiesta, or bot boot camp.
Their system as it is currently is irredeemably stupid, and until they separate ranked and social MMR to be different ratings (and ideally they would just have per-playlist MMR), each game mode will harm the other. And so long as bot matches affect ranked MMR at all, the CSR rank means nothing.
If this were a developer with a history of being at least semi-competent, I would assume fixing these things are part of the tweaks they mentioned. But this is 343, the absolute bottom of the incompetence barrel. I don’t have any confidence they see this as a problem to be fixed.
Do you have a source for your assertion about how bot sandbagging only affects new accounts?
I didn’t know we had a source for -any- actual definitive information about how MMR works in Infinite, much less niche questions like how bot matches affect MMR. I thought the only reason we learned that was through experimentation, but maybe I missed something from 343.
Using the knowledge we have from it paired with real world examples we have experienced helps show a decent amount of what parts of it are certainly still relevant to some degree.
Hey, I don’t know why people get off on it. It just seems like another potential source of error that produces the games we all get where we see huge CSR gains/losses that seem to have nothing to do with the quality of the other team (because the other team has intentionally moved their MMR from the CSR).
You may disagree with those assumptions, but the data does not. Most players will find strong correlation between modes.
Even bot bootcamp. There will be rare cases where a players ability may fluctuate wildly, outliers often exist, but it doesn’t change the body of supporting evidence.
Some people just like to pub stomp. I’m sure there are people out there that prefer to go 15-0 against players that can barely hold the controller. It means nothing to me, I’d rather go 12-11 against a really hard team.
If the enemy can’t fight back then lots of kills are meaningless, but hey some people seem to love ruining other people’s day.
I mean… There have been a lot of examples on reddit and this forum showing exactly the opposite.
Basically anyone who decided their main way to play the game is “I’m going to play bot boot camp to quickly rank up my battle pass” cannot play ranked modes because the game thinks they are onyx.
Anyone with a platinum or higher rating can easily learn how to go 20-0 every single game against robots. Bot boot camp does not test your ability to compete against human players, at all.
I’d be intrigued to see an account that only plays bot bootcamp (at a moderate level) to then go into ranked and be placed against high Onyx competition.
cf. this thread
https://old.reddit.com/r/halo/comments/sjcg8u/its_official_matchmaking_is_broken_social_games/
This shows the opposite effect really clearly. After 100+ bot slaying games, the system places him at a 2000 MMR level. His CSR was put as Diamond 1, the highest possible after placements, despite the fact that he lost every placement game. Even a few games after placement matches, his MMR is not decreasing quickly or at all (nor his CSR but who really cares), and he is playing high Onyx players. He played from 4 weeks ago until 2 weeks ago very regularly. Then didn’t play for 2 weeks. Then played 1 game the other day. Maybe our noble bot slayer has given up
Is pretty funny, but in bot slayer you are put with other people who perform well. So if you can kill bots quicker than other high level MMR players it would suggest maybe you are capable of a higher MMR. As I said outliers may apply. If you’ve got someone who only plays Bot Bootcamp and they are bashing up bots quicker than other high level players, I’m assuming there is some correlation there.
Obviously system overinflated it of he lost all 10 placements but based on his performance against higher Onyx players, the guy didn’t do bad at all so Diamond 1 is definitely fair. Would average out over time.
Thanks for sharing, that is interesting. The system has definitely got it wrong there.
The match history showed it at the time of the reddit post. Don’t know if it’s still accessible. Most the onyx players he was matched against were 1800-2200.
For what it’s worth, I’d say they could have all been onyx 1501 rated players and every point I am making still stands.
Their system is broken, at least in this one aspect.
