Huh I guess that would explain my situation as before I started taking advantage of the bot games for challenges my games were all pretty evenly matched
4 Likes
I would prefer to have MMR limited to the playlist you are playing in. Now I’m wondering if my performance in Big Team Battle is causing harder matches in ranked. It’s easier to get a huge KDA in BTB – so with what happened to the the OP in mind, it makes sense large KDAs in BTB would boost my hidden MMR, even though that playlist is nowhere near as competitive as ranked.
4 Likes
It should be locked to each playlist. Like, I could be great at Tac Slayer and get 5 KDA and then have a horrible time in other modes because I’m “such a great player”
Yet another issue that they need to fix.
Also, the teams shouldn’t have an overall rating, they should pair the teams with equal players.
Example
Team 1
- Diamond I
- Diamond II
- Gold III
- Gold I
Team 2
- Diamond I
- Diamond II
- Gold III
- Gold I
This would make the games a lot more fair. And we simply need lobby VC, the “social” games aren’t social, you aren’t socializing with the other players, it’s boring. It’s NOT Halo
1 Like
Sure, except the ranking system is basically meaningless because it is just as broken as the matchmaking.
Baby steps
2 Likes
I would prefer per playlist too. It’s just that from what I’ve read having a global MMR is baked into TrueSkill 2. And as someone who isn’t a coder I don’t know how hard it’d be to have MMR in playlists and not have it impact the global level, if that makes sense. It doesn’t seem like it’d be hard to just disable it for one playlist though. I was just suggesting an easy short term idea, but long term I absolutely want what you’re suggesting.
3 Likes
! UPDATE !
I tried editing my post again to notify anyone who cared to read, but I’m either ignorant or unable do to a ‘edit limit’? Sorry, I’m new at this…
Anyway, just wanted to let everyone know I have submitted a ticket late last night have yet to hear any feedback regarding the issue. I’ll be sure to follow up with any new information as it unfolds.
Thanks again everyone, greatly appreciate the support to get the word out!
6 Likes
Yikes! This isn’t good! Normally I’d just, NOT recommend Bot Matches but definitely explains where the occasional stomp I get put in, comes from.
2 Likes
Yeah, technically wouldn’t you be able to lose 100 games on purpose to bots and get an artificial mmr?
1 Like
Has anyone ever lost a bot match? I’ve never seen it. Every last one of them ends in a Steaktacular.
3 Likes
I think this “glitch” is actually pretty funny.
From a programming point of view you can see how a MMR exists for bot slayer. You want as much code consistency across all your modes. It’s really more a case of “they didn’t turn it off” then specifically putting it in.
We also know that if you are placing in a new playlist for the first time - TrueSkill2 extrapolates a starting MMR from the closest type of playlist you already have.
In this case the only MMR the system had was the bot MMR.
You could then see why the system had so much trouble finding the real rank in just 10 games. Ten is plenty for "ball parking: a player but you actually need closer to 50 to be truly accurate. Plus, because they had played so many games at that level - their skill curve would have been quite narrow (a rank locking effect) and hard to shift - a high Onyx player is not going to drop to mid-platinum in 10 games.
Can you imagine someone in 343’s Q&A putting their hand up and saying “but what if a player does nothing but bot matches for a month and then tries to go straight into ranked play?”
So this is an interesting glitch. No more. No less.
We should value it for the insight it gives into placement matches. It doesn’t, in anyway, reflect on the utility of TrueSkill2. It’s affect on anyone else’s rank would be virtually zero.
I’m sure once 343 is aware of it they will just turn it off in case someone else tries to glitch their starting MMR up (or down).
4 Likes
As others have said, which I entirely agree with here… is your hidden MMR needs to be tied to individual playlists only. No stacking, especially ranked. If they keep SBMM for casual.
This is part of the reason you can match evenly rated opponents (I.E 1500 vs 1500 onyx) and gain 1 csr… because they have a very low hidden MMR. but you can beat a diamond 4 and go up 15… lol… Ranked should only go off of CSR when determining how much you lose or gain.
2 Likes
343 needs to fix this. It’s one thing to get smacked for a match or two but this is a huge oversight on their part.
If you have never played a playlist your mmr should be a starting point for a match or two but nothing more. It sounds like 343 wants to pull your win % and k/d down which is bs. The playlist mmrs should be seperate and not co mingled.
I believe it is. There is a core part of the skills data that is shared (for effiicieny) - but each playlist has it’s own offset. Not confirmed but I would expect that to be an offset for both average MMR and standard deviation (confidence).
You can effectively have different MMRs across your playlists.
If it is part of the reason, I am sure it is minor.
The main reason you only gain a small bit of MMR is that you are either favourites (which most stacked squads would be) or pretty evenly matched.
Plus they have nerfed how much you can gain per game in Onyx and weighted towards losses. This is to stop people climbing too high on a run of expected wins. Everyone is already getting upset with dropping back to their expected MMR with a loss… imagine if the climb was faster and the fall even steeper.
You have to keep in mind that the system becomes inefficient and loses accuracy the further out to the tail you get of the population skill / curve. It really needs tournament type structures to sort the MMRs into a valid order. But with squad stacking, time zones, and distant servers - the fine tuning of your MMR is guess work at best.
It’s a minor oversight at best.
How many players are going to spend a month or two gaming their MMR through bots?
TrueSkill2 will drag them back pretty quickly - although as the OP found out, probably not in the run of 10 placement games.
No, it makes perfect sense. Overall a player will have a similar MMR across all playlists - so making this assumption is very safe. And remember it’s only a starting point - the placement games (with a volatile MMR) should polish towards their actual MMR very quickly. Certainly more accurately than starting with an average MMR.
Calm the farm. There is no flat-earth type conspiracy here. It was just an oversight in the mechanics. You could probably count the players affected on one hand.
Well, looks like I’m going to be avoiding bot matches for a while.
1 Like
But why?
I suspect the only time it really matters is as per the OP’s situation; where your bot MMR is used as a seeding MMR for placement in ranked.
Otherwise it’s influence through the global part of the MMR would be minimal.
Bot bootcamp affecting your background MMR is just about the most 343 thing I’ve ever heard. I’m actually laughing. I legit can’t anymore…
4 Likes
Just to be clear, this didn’t happen with the intention to fabricate my MMR. Regardless… I think you’re severely undermining this and the ramifications it holds for the integrity of skill based matchmaking.
What if I were to tell you that this very issue has and continues to plague the manipulation of ranking placements, allowing players to falsify their respective positions on the bracket?
I’ll do you one better… here’s proof:
https://halotracker.com/halo-infinite/profile/xbl/AsauItz/overview?experience=ranked&playlist=edfef3ac-9cbe-4fa2-b949-8f29deafd483
This individual “Asualtz” lowers his MMR in SWAT and manages to match players as low as Gold Ranking while being a 2200+ Onyx. He is the prime example of just how ridiculously easy it is to alter the MMR despite ranking.
My point is… this problem goes far beyond my situation alone. I stumbled onto this unknowingly, but others have and will continue to seek this out willfully and that poses an incredible breach of competitive integrity. Especially when this is the franchise that essentially created Esports and the legacy that follows.
343i has been handed the mantle of responsibility to keep that legacy intact. And not to go off topic here, but this is only one of the multitude of issues hindering Halo Infinite’s potential concerning the performance of multiplayer… cheating, dysenc, forced cross-play, custom games, and now this.
I believe 343i is capable and willing to resolve these issues and I’m under no illusion that it’ll be accomplished quickly. Nor do I imagine it will be without it’s flaws, but as long as they’re actively working while (and this is VERY important) being completely transparent I will continue to support Halo.
343i has just within these past few days shown that they’re taking customer feedback seriously. With the further changes soon to be seen in the store and the attempted fix of BTB; it’s clear there are those in development who have our backs.
P.S.
Sorry for the long winded ramble.
2 Likes
I literally cannot comprehend how they messed this up so badly.
3 Likes
If this is true it’s shocking.
How can performance against bots count the same as PvP in terms of matching you against others?
Who thought this would be appropriate?
3 Likes