I play ranked matches both in 4-person parties and solo.
In the case of a 4-person party, I often get a big increase in CSR if I win if I am higher in the party, and almost no decrease if I lose.
Of course, the rank difference at the time of matching is probably taken into account, but it seems to me that the evaluation is more lenient in the case of a 4-person party than in the case of a solo player.
If this is the case, I think this needs to be corrected.
The game is making bets on whether or not it thinks you would have won based on your team composition. If it bets you would have lost and you do you don’t drop as much. Also having a good KD prevents dropping much.
It sounds like you are a higher rank than your friends and if so this is definitely the case as I bet you have a great KD when you play with them as its trying to match based on the average of your ranks where solo will be with people of your rank (or at least closer to it) thus your KD probably isn’t as good thus bigger drops and less gains.
So no its not really based on being in a group or solo but more so your KD in those games that determines gains and drops.
1 Like
“Making bets” is a weird phrasing.
I’m kind of envisaging the match maker standing there holding wads of cash as the crowd screams at the combatants.
Your post game CSR gains or losses are not related to personal performance in that game.
And the metric used by TrueSkill is KPM. Which does influence your MMR (it’s just a weighting).
And it’s a fundamentally different beast to K/D or K-D.
“Making bets” such a good way to put it. The whole thing feels like gambling. Except you never know what you’re going to win or lose. I once lost more rank in one match than I did by winning five in a row. This system is baffling, unsatisfying and over-engineered. One of the many reasons why no one is playing this game.
Okay but unless you get a couple kills then hide you should have a good KD that goes along with KPM. I don’t see any way you could have a high KPM and not have a good KD. Maybe at lower ranks it would be possible to trade over and over and have a bad KD then still have a high KPM but from my experience its not possible. So i guess im wrong about the specific there but in general get more kills and spend less time dead so that your KPM is higher thus making your KD higher as well.
There are three problems…
-
Poor communication from 343. They seem to be happy to allow misconceptions like “KD is important” to perpetuate.
-
The scale of the CSR. By making it 1-1800+ we are subject to the precision errors of the MMR. Everyone was happier in the old days with smaller scales (eg. 1 to 50). Small fluctuations in rank happened silently in the background. People happily played for days (or even weeks) with no change in their rank. Currently we are losing our minds over a handful of points after every game.
-
There is no XP rank to grind. So 343 feel compelled to award bits of CSR even if you don’t deserve it (ie. your MMR didn’t budge).
It took me a little while to wrap my head around it.
You pretty much can’t fake KPM. It is a pure reflection of your 1v1 ability vs the rank you are playing.
If I am a mid-Diamond and play a bunch of lower level players I will dominate and come out with a pretty good K, K/D, K-M, and KPM.
As I play stronger opponents it gets harder to maintain all of them.
My KPM will naturally start to fall to around 1.2 to 1.5 KPM as I get equal opponents (think of 12.5 kills as average in an 8 minute match).
If I play opponents better than me it will be virtually impossible to increase my KPM. I just can’t because I don’t have the skill to win the 1v1. But I can, through strategy, come out with a good K-D or K/D.
For example - I could eke out a few black widow kills and then camp to finish 6 and 2 for the game. A K/D of 3.0 and a K-D of +4. But a lowly KPM of 0.5
If you think of it that way KPM is a mirror of your (slaying) skill in the game.
Of course it’s only a weighting. But Josh Menke used to look up peoples KPM vs opponent ranks when people complained about their rank. He used to say things like “your KPM vs Diamond is around 1.5 but vs Onyx is 0.7 - so you are definitely Diamond”.
And I think the final misconception is people think that it’s a high KPM that defines their rank per se. It’s not. It’s having the same KPM as your opponent. If you want to be Onyx it’s about matching the KPM of the Onyx players playing against you (eg. 1.5 to 1.7 KPM). You are not ranking up by showing off and thrashing a bunch of players ranked below you.
A little bit.
There is also a weighting for DPM. Although it’s less than for KPM. But it would still kind of act like a K/D. Or at least a KPM/DPM.
And yes, your point is very valid. The more time you spend respawning is less time you get to maintain your KPM.
I wish there was a way for players to look up their KPM.
An auto generated graph of KPM vs rank.
I think it would be very illuminating. For everyone.
1 Like