Would love the losing spree anouncement! Hopefully to everyone playing so they know how awesome you’re doing. i feel like a loser tag instead of your name tag would also be appropriate
2 Likes
Does this take assists into account?
1 Like
Well, I question the usefulness of it…
KPM and DPM are valuable for quantifying player skill and evaluating performance. In turn they’re useful for matchmaking. However, they’re very general metrics. A large number of variables will influence those values. It’s likely a big reason they’re useful in the first place.
The trouble is that general nature makes them far less useful for a player looking to improve. If KPM and DPM are worse than expected what exactly do you take away from it? What exactly do you look at to “fix” here? This scenario basically amounts to play the game better.
What is the takeaway when KPM and DPM were better than expected? I’d think anyone invested in their performance would view this scenario the same as the one where they under-perform. They’d be looking to improve their performance. Find ways to play the game better.
There is a difference between what is useful for the MM vs the player for evaluating performance. The MM is attempting to remove variables and simplify the process. Isolate the information it needs down to the least amount possible while remaining accurate.
I seem to recall the trueskill2 “paper” finding, or stating, it’s ideal to rely on data a player has minimal control over to quantify skill and evaluate performance. It makes a good bit of sense for the matchmaking. For the player it’s a very different picture. If they have minimal control over it then it’s not very actionable.
1 Like
i went 13 and 4 in fiesta on cliffhanger and it said I underperformed in deaths 
2 Likes
Due to my vast underperforming I’ve been placed in rank as aluminum 2…… did anyone else know it goes that low 
. The emblem and name plate is a trash can if anyone is wondering.
All stats have their pros and cons.
But I’ve really come to love and appreciate the value of KPM (and DPM).
When I play with my sons, who are both way better than me, I’m in a harder lobby than I’m normally comfortable with. I have to alter the way I play.
If I camp like a -yoink- I can match them on said KD. I can push a respectable KDA. I can hold the oddball and feel like I’m doing something. I can jump in at the last second and help return the flag.
But so can they. They could do all those things better, AND still get more kills. The only think I can beat them in consistently is assists - and while it suggests I am a good team player - it also tells a story of someone who is out of their depth.
What I can’t match them in is raw kills (and hence KPM). Not even close. It is literally the skill gap between me and them (and the lobby).
So, like it or not, KPM is a reflection of 1v1 skill.
And yes. It is a general metric. We know that not all kills are the same. Coming across someone reloading with no shields while you have the rocket launcher is not the same as a bulltrue vs a camo wearing sword wielder.
One day we may have an AI that analyses each play!
But overall - if you are looking to improve. KPM is it. At least in terms of the simple (and low resource intensive) stats. It is the stat that best summarises the differences between ranked players.
It doesn’t lie to you like a KD can. It’s not buffed with disappointment like a KDA is.
A simple graph of KPM vs opponent skill level will tell you everything you need to know. The drop off will happen just above your rank - where you start losing the 1v1s. And how steep the drop off is will relate to how close you are to ranking up.
It’s no surprise that it was found to improve the speed and accuracy of ranking in TrueSkill2.
But keep in mind that it’s not about getting a bigger and bigger KPM. It’s about matching the KPM of the rank you want to be. If you want to be an Onyx player you need to match the KPM of other Onyx players vs Onyx players.
Josh once said something along the lines that if you knew the rank of the lobby you look at each player’s KPM and predict their rank. It was that reliable.
Going forward. Apart from more focus on KPM and DPM, I would love them to look at damage. Ratios of effective (leading to a kill) and ineffective. Proper assist damage ratios. Damage over time. And so on…
Bro I got tin foil comp rank…
3 Likes
Your attempt to rank higher was foiled?
They stuffed up the UI. It’s the wrong way around for deaths.
Even if it’s assumed KPM illustrates 1v1 ability alone, which I wouldn’t agree with, a lot goes into it. Which area do you seek to improve to win more 1v1’s? Which one do you prioritize? How do you know which area to direct your focus toward to elevate your KPM? The underlying point is you don’t know. The stat doesn’t point in a specific direction when viewed in isolation.
The fact that there has to be this much discussion online to understand how it works is the perfect indicator that it is not suitable.
God I wish we could bring back halo 3 matchmaking. Get stomped sometimes, win sometimes, and on the whole do pretty well if you were a good player.
Quick play is the only arena playlist even worth playing in infinite.
1 Like
After playing a few games I like the idea of the game telling you that you performed above expectations. It’s kind of nice in games where you feel like you didn’t do very well but then find out the enemy was expected to wipe the floor with you and you actually did pretty well.
In other games when you feel you played great but it tells you you only performed at the level expected, it brings you back to earth a little.
The big catch is when you have a bad game it really should just say nothing because you are probably already well aware you sucked in that game.
So in my opinion, fix the ‘underperformed’ issue for deaths and only highlight to players when they have actually performed above expectations.
2 Likes
Not necessarily 1v1 ability “alone”.
Every kill is different. Or in my case just surprising.
But it is the best of the current basics stats that reflects 1v1. Far better than KD, K-D, KDA, or any other combination thereof.
The time denominator is the key. The only way you can improve your KPM is to get out there and kill more. And that takes skill. The tougher the opponent the more skill it takes.
Unlike those other metrics - you can’t manipulate or fake them (eg. KD).
The same could be said about improving any other stat?
But in terms of KPM - the obvious answer is to focus on killing more enemies. Which is sage advice for anyone who wants to improve - or help their team more.
It’s not that difficult.
But the problem is that 343 have put zero effort into it. They pretty much introduced expected kills and deaths on waypoint with no post it notes. We were left to fend for ourselves. And now they’ve appended a performance review on the front page of the post game report with zero context.
Just a little explanation would go a long way to making this stuff useful.
The problem here is that you didn’t necessarily suck at all.
If your expected deaths was 9.5 and rounded to 10… and you got 9 kills… you only fell short by 0.5 kills - but the game is telling you that you “under performed”. Which is kind of true… but not helpful.
There should be a plus or minus 15 to 20% for a “par” performance.
Maybe even a traffic light colour code systen. Make the value orange if you are +/- 2 of expected. Green if you are 3 or more above. Red if you are 3 or more below.
There’s already some degree of traffic light system just not in those colours. It must have some sort of tolerance before it declares you over or under performed because I frequently have neither applied to my summary or only one of either kills or deaths. I don’t think I’m hitting the exact mark expected of me that often.
Either way I still don’t think people want to have it highlighted that the played badly. It’s really just not helpful.
It doens’t even work correctly,
I got 2 deaths in a game and it told me I underpreformed…
I’d say only 2 deaths in a gamer is quite good…
It’s a known issue. ‘Underperforming’ on deaths is doing well.
2 Likes
I think you would be surprised how often you do.
We are anything but unpredictable.
And I was only 1 under / over in the game I checked for “under performing”.
Agree. With no context it’s no wonder people are bemused.
1 Like
I over performed on deaths. 20 deaths a game lets GO!
Yeah I think from the amount of feedback we got in here the general consensus is that the current version is not good whether it’s correct or incorrect. People know when they’ve played bad so they don’t need that to be reinforced in any way, shape or form or accidentally reinforced. It’s probably fine if they leave it as a you did better than we expected you to, but they should never tell you that you did worse than they expected you to. That is quite literally just going to make everyone who did bad feel like utter crap and they already feel like other crap because they played poorly. Yes let’s kick the man while he is down.
1 Like
Why the -Yoink!- would they add this to the game?! I don’t want to see this at the end of a game. It is just not necessary at all.
This company really has no idea!
1 Like