Request the abolition of MMR

There’s not enough info here to make a judgement.

What were the average MMR’s of the two teams? That is the only way you will know how closely the two teams were matched on paper. I tried to look you up but you have your stats set to private.

Just using the min and max of each CSR your unranked guy only needed to be Onyx 1650 to evenly match the teams.

And that’s before you add any weightings for playing as a squad.

We know that the CSR is very close to the MMR for the Onyx players (they had to come up from D4). But everyone else on the table could have a CSR well below their MMR.

Depends. Which one is your CSR? How many games have you played in S2? What was your ranking at the end of last season?

For example. If you are D6. You could have a CSR of 1499 - which is pretty much Onyx anyway. If you have only just placed your MMR curve (still wide) could well extend into Onyx - in which case you’ll match against them - at least until you prove to the system you are or are not worthy.

It does. But if it’s any consolation. Playing sides ranked higher than yours is the fasted way to “grind” to your final CSR.

It does.

Your RANK is your MMR.

It’s not a problem.

The CSR is just a way of describing your MMR. And by starting on the left hand side of your MMR curve it can be conservative early in the season when it’s still a little unsure of your rank.

After about 15-20 games of consistent play your MMR and CSR will be very close. By 50 games they will be close enough to be the same.

If you want to use your CSR to match then you would have to use the average of your MMR curve (as opposed to it’s left hand edge). For 50% of people this will be higher than their actual skill rank and as the season progresses they will be forced to grind to a lower rank. Yuck.

1 Like

I wasn’t talking about the 1-50 style, just the aspect of MMR incorporated in the ranking system.

Personally, I still like the 1-50 ranking system. Maybe they could made it 1-117 :wink:

This is incorrect, I played 30 games of LSS which you can check. I got the weekly reward of the first week of season 2.

Yes it does. This has been thoroughly tested on new accounts. You can play Bot bootcamp and it affects your MMR. (Unless they changed this for Season 2 which i did not read in the updates)
You can look up videos discussing this.

As I said, it doesn’t take 1000s of games. It doesn’t necessarily take games period. It takes time. You could go to work and come back and lots of the hardcore players are already level 10+. After a week or two, they’re level 20+ and you’re free to start your grind.
Besides, it’s not a big deal to have 1 mismatch game every once in a while. In the end, you’re trying to achieve the highest rank not just play against players at your own skill.

I agree. This game needs a breakdown in game of what the rank represents. It additionally needs more than that.

Legit question - Does anyone actually play this game anymore?

TrueSkill2 is more accurate than TrueSkill, so you wouldn’t rewind that.

And Max kind of nerfed TrueSkill a bit in Halo 3. Which doesn’t make sense. So let’s not wind the clock back there either.

What I don’t like about the current system is focusing on the MMR number 0-1800. It doesn’t work zoomed into that level. The fluctuations around your actual rank frustrate people.

So yes. I’d be happy with 1-50. And I think 1-117 is an excellent idea.

But I do like how the current system is based on the normal distribution curve of the population. One standard deviation for each of the six divisions Bronze to Onyx. Straight away you know how many people are expected to be in each one (eg. 34% for Platinum). And there are 75:25 odds for players in neighbouring divisions competing against each other. It’s beautiful.

I would keep Bronze to Onyx. Get rid of the MMR number. Make Onyx 1 to 6 like the other division. Or if you want more - make each one ten tiers. And for nostalgia change Onyx to Halo 2 icons.

But it still stands that you haven’t played ranked since Apr 28th? So my comment still holds.

What has been “thoroughly” tested is that you can build up a bot account on a brand new account and use that as a seeding MMR for the first time you enter ranked. The system will initially place you against 2000+ players (depending on how many bots you stomped).

But it only works on brand new accounts. For the first handful of games.

It’s not “affecting your MMR”. It’s literally copy and pasting it because it doesn’t have any other data to use.

And it’s a known feature of TrueSkill. It makes sense if you are a Diamond 5 in Team Slayer to assume that you will be around that mark the first time you play Team Snipers.

But this only works for brand new accounts who haven’t played ranked. It does not translate into playing a me playing a few games of Bots with my mates and borking my ranked MMR forever.

This one reddit post has created more “false information” than Flat Earth, Anti-vax, and Trump’s election campaign combined.

OK, but even if, for the sake of discussion, I accept your premise. Why would you pass up a system that does everything you want virtually instantaneously.

Can you imagine the sales pitch?

Option A takes weeks (if not months) and creates lots of mis-matched games.

Option B takes days and creates very few mis-matched games.

Which one do you want to go with?

It’s not every once in a while.

And as I just mentioned. In the end we just want to get to our highest rank. So let’s use TrueSkill2 and get their as fast as we can.

What I said was a rank match.
Even if the matching is judged to be at the same level by MMR, the difference in rank is often severe.
And for some reason, after Season 2, the probability of becoming a low-ranked team has increased.
I’m not convinced if I lose in MMR where the difference in rank is terrible, but if I lose in the same level of rank, I can be convinced.

It’s a video game. There are a lot of people that appreciate the journey. Without the journey, the destination doesn’t seem like a destination, but rather a stagnant location.

My guess is there are many people arguing about ranked when what they actually want is Progression. It’s confusing because previous Halos made the ranked playlist more tied to a progression like system. On top of that, this game doesn’t have any good progression system in it.

Yep. Agree. the lack of an XP progression system has left a huge hole. Combine it with people not understanding (or not wanting to understand) skill ranking and the angst levels are rising.

We need one ASAP.

I think if we weighted it to wins, medals, and objective scores - then it could scratch the itch and we wouldn’t need to tie the skill rank to it.

You could of course still reward dual achievements. eg. have military medals or rank for getting to a certain skill rank AND a certain level of XP.

Maybe have an XP bonus multiplier for your current seasons skill rank; 1.0 for Bronze, 1.1 for Silver, 1.2 for Gold, 1.3 for Platinum, 1.4 for Diamond, 1.5 for Onyx.

And I think have different paths would help sell the journey to more people.

1 Like

Sadly, I think 343’s definition of “ASAP” is not the same as mine.

1 Like

Agreed

Personally i would just like to see the same system as MCC. Otherwise you would create a system where people are still without progression option, because they are not skilled enough.
On top of that:

  • you already have a rank for you skill (the colour system tied to CSR)
  • people will then complain they can’t play casual on social, because that would lower the amount they could progress (and it removes the aspect what makes it social: not having to care about winning/losing or how good you perform)
  • people will then complain that others are just playing to sweaty and don’t play casual.

So make the progression system purely based on XP. Then it already is a bit combined with skill (doing better gives you more XP obviously), although it would be mostely loyalty (how much you play). Wich is completely fine. Next to that progression system that gives you a militaire rank (like in MCC) you also have/keep the ranked system wich give you a colour-rank based on your skill.

Only for ranked matches perhaps. Social matches don’t have anything to do with that, so it would be unfair to have such a multiplier for socials. You would then only get people to rank up in ranked and then play in social playlists where they still have a low MMR on to rank up pretty quick (or use a friend with a smurf on social to get easy matches). And i don’t think we should incentivise people to smurf even more then they already do, because that hurts players who are actually playing to their ability.

I would also make ranked matches have more XP then social no matter what rank you have. So perhaps have the multipliers something like this:
0.5 bot bootcamp
1.0 social
1.5 ranked bronze
1.6 ranked silver
1.7 ranked gold
1.8 ranked platinum
1.9 ranked diamond
2.0 ranked onyx

1 Like

This all explains a lot. I was wondering why I kept hemorrhaging so much CSR after every loss even with solid performances. Now I’m starting to reconsider the times I complained about facing off against Diamonds as a Gold / low Plat because now that I think about it, those seem to be the only matches I can actually afford to lose lately.

Looking into the future you project… Just a moment… Ah! Got it.

“You are literally FORCED to play Ranked - the worst mode Halo has, to progress on the battle pass AT ALL!”

Tell me you don’t see that coming with a straight face.

Rank is CSR not MMR.

CSR literally equals your rank in the ranked playlist, and has no bearing on anything else in the game, other than that particular playlist. Hence its name Competitive Skill Rating.

MMR is an assumption of your player skill based on all of your games across all playlists. From FFA to Big Team, all of it.

CSR does not apply outside of the ranked playlist. However your MMR does.

This is why some people will quit games/ play poor on purpose in Quick Play/ None ranked playlists, to throw their MMR. So that when they play in ranked they are being matched against easier opponents.

In ranked, team composition is determined by your CSR. The AI will try its best to make sure those searching for a game are matched against those with equal CSR. You can see a teams approximated CSR at the post game screen. Team CSR is one of the variables used to determine how much CSR is given or taken based on the outcome of a match - Win/ Loss, KDA.

This is why someone with a relatively high CSR (High Diamond to Onyx) will spend up to 3/4 minutes searching for a match, as the average CSR is sitting around Plat 2-3 according to the latest information on halo tracker. Lower ranks generally find matches quicker.

It’s mostly a moot discussion.

There is global shared data.

But each playlist has it’s own offsets. I imagine this is enough data for it to act as it’s own MMR curve (eg. mean, standard deviation, and decay control).

And for each playlist your CSR and MMR eventually come together to be the same value. But it would be the MMR value that is used in matchmaking etc (it’s the one that is mathematically a curve)- so that’s the one I would consider your true rank.

The different playlists have their own functional MMR.

This is how it comes across in the TrueSkill2 discussion paper and relevant GDC talks by Menke et al.

It’s the way H5 operated. You could quite clearly have a different skill ranks in different ranked playlists. And most importantly you could undergo placement in a one playlist and not have it unduly affect your other playlists.

And now we can peak at your MMR for FFA games. And it appears separate from you ranked value.

Which doesn’t get you far.

First of all I don’t think the influence is anywhere near as strong as some people think. I certainly haven’t seen a huge difference in my own play (keeping track of opponents average MMR over the last few weeks).

And I think it was Menke in a twitter post that once described the cross playlist influence as “weak” (but doesn’t mean 343 can’t have tweaked it in his absence).

You would get a drop the part of your global MMR data that reflects “form”. So you may (or may not) get some softer opposition. Which could earn you some extra CSR - but not MMR. And when you eventually lost a game it would all be taken away.

But at least you can jump onto Waypoint and complain that your day’s “grind” was taken away in a couple of losses.

Going into Data Hive for my last ranked game without someone placing;

  • My team; 1192, 1119, 994, and 1299 (average 1,151)
  • Vs enemy; 1583, 1085, 1051, and 512 (average 1058)

Our team MMR averages were listed as 1354 vs 1352.

So, matchmaking isn’t strictly using the CSR.

It’s to do with lots of maths on curves. This the MMR part of your data. But again, the CSR is describing the left hand edge of your curve.

The problem is people are manipulating the game through smurfs and sandbaggers and grinding wins way more than they lose. They may only get a few CSR points a win (driven mainly by personal performance) so they need to win way more than they lose (where they can lose -15 if their MMR is lower than their CSR).