Lol can’t believe you love mlg sweat fest halo 24/7
1 Like
this may be true, but its due to how bad it actually is for the fun of the game. It creates a toxic atmosphere of addiction similar to gambling. People keep putting quarters in because surely the next time will be the time it works in their favor.
Some might say that’s problematic and morally questionable.
You keep saying this but its the same thing. Your kills and deaths each match are what determines how MUCH you rank up. Winning is what is required to rank up at all, but winning with no kills is a single point of CSR at best. You have to slay out and win too. And the more you do that, the more you get saddled with worse teammates to balance out YOUR skill level to make it fair for the other team. That’s the basic premise of SBMM.
No. The person who gets 40 kills can still lose all these objective matches if he’s the only one doing anything of value.
If he stops slaying to do the objective, its the same story, because his teammates obviously aren’t going to support him while he’s on the line.
That’s what we’ve been trying to explain about team play. You have to be smart about the nuance, not just look at the numbers.
this right here is what annoys me the most
Why wouldn’t you try the same in every match? Why wouldn’t you want to win? More importantly why WOULD you want to LOSE?
Losing is bad and it isn’t fun. Getting killed a lot is bad and it isn’t fun. What is fun is building a throne of bones on top of your gold medal at the end of the match. There is no other reason to play this game, or any FPS competitive games.
1 Like
And I can’t believe you turn every game into a sweat fest.
Relax and have fun.
Some might. But it’s a very, very tenuous proposition.
It’s not.
If you go 4 and 2 you have a KD of 2.0
If the match goes for 2 minutes you have a KPM of 2.0. But if the match goes for 8 minutes you have a KPM of 0.5
Same KD. Very different KPM.
Playing to protect your KD tends to make your KPM worse. Which is why I think it’s very important to stress the difference.
I’m not sure you can say that. Your CSR change at the end of the game does not take into account your personal performance.
Again. For those people out there trying to work out the best way to maximise their MMR - the details are important.
This bit is correct.
Almost.
The basic premise of SBMM is to create an evenly matched game, where both teams have the same chance of winning.
If the match maker is forced to widen the skill ranks within the team (eg. when population is low) then you will find good players “saddled” with lower ranked ones. And this probably does happen more to the very good players (not many similar players about at the best of times).
But that is not the “premise” of what the ranking system is trying to do.
They could. But if they are slaying strategically they are giving their team the best possible chance.
Of course you can’t always control what your team mates are doing (sadly).
But that’s the principle of team games. The best team should win. That’s the nuance.
My question is why would you want to win when you didn’t have to try?
Because that is what you are essentially saying. You want to play at the 100% but not have to come up against a team that can fight back.
That doesn’t sound like fun.
But isn’t that what you want to inflict on your opposition?
When I was young enough to play competitive cricket - I played most of my career in the firsts (albeit lower leagues). You would sweat those games. You would agonise over every delivery; long run up, full pace, tailored to who was on strike, meticulous field placings. Continually trying to rev the team up. If someone made a mistake or dropped a catch it was distressing. It was still a lot of fun and rewarding - but also very mentally exhausting.
Sometimes, when I was working the wrong shifts, I had to drop back a grade or two. It was much more relaxing. Chill.
There is no way I would come off the long run and try hurt the players at that level (when I was young enough to actually bowl quickly). You would experiment with different deliveries. Try things you wouldn’t normally do (it was the only time I would actually try and bowl in-swingers for example).
Don’t get me wrong. You are still out there for the win. But having fun with a bunch of mates is the number one priority. You would never criticise your team mates for making a mistake - even if they dropped a simple catch off your bowling.
And if you started acting like a -yoink-, the first people to pull you into line would be your team mates.
People keep telling me social is for kicking around with friends. Mucking about. Having fun. And I agree. That is what I do.
You’re trying to tell me that Social is for going 100%, winning at all costs, and showing off your KD.
But that literally sounds like Ranked to me?
You complain that ranked and social feel the same - but freely admit that YOU play them exactly the same way. To the letter.
The only difference you seem to be putting on the two is the quality of your opponent. In ranked you are happy to work for your win. In social you want to be spoon fed players for the easy win.
You want to have fun stomping people. You say that losing is not fun - yet you expect your opponents to -yoink- it up and have fun doing just that.
2 Likes
what does this even mean
it absolutely does.
You have to try every match, the entire games uses SBMM
If they’re randoms I absolutely would. You can’t teach these people anything otherwise. Most of them barely have an IQ over their KD.
Yes, because I don’t want to lose and I hate going negative. Are you saying you play social just to lose and go negative?
You know exactly what it means. Players changing they way they play to protect their KD as opposed to helping the team to win. It’s a problem as old as time.
And nice deflection from the original argument.
I assume you concede that KD and KPM are, in fact, inherently different stats.
And that to rank up a player needs to specifically buff their KPM and DPM. Not just their KD.
Occasionally you get an objective match, particular 4v3 or worse, and the opposition try to milk their stats by prolonging the match. The classic one is when they drop the Oddball on 99 seconds and then try and bait you to come and get it. I love hiding at the other end of the map - they think by getting a couple of extra total K and/or increasing their KD they are ranking up - but the joke’s on them, the extra time ticking away is ruining their KPM. Especially because the number of players in each team is factored into the algorithm!
It doesn’t.
It definitely didn’t in Halo 5. We know that. It was explained multiple times.
It’s a little less clear exactly what have done with Infinite. They said in the ranking blog that it can contribute. But we don’t know how. And when I ran over 500 game results through a spread sheet there was weak to no correlation with any of the in game stats (including KPM and DPM).
We know KPM and DPM affect your MMR and that is by far the strongest influence of personal performance on your CSR - the “leaking” of MMR back in the next few games. And I suspect this is actually the relationship 343 were “dumbing” down in their blog post.
And besides, you are essentially contradicting yourself. You keep saying it does - but then spend the next ten posts ranting about how you “carried” the side but didn’t get any CSR for it. So which is it?
I play social to have chill and have fun. You should try it.
1 Like
They aren’t. If you have a low kd every match, you have a low KPM. If you have a high KD every match, you have a high KPM.
They have to go very positive and win the match, yes. That’s why KD is important.
It does. If you perform “better than expected”, IE getting far more kills than the algorithm expects you to, then you rank up a lot more each match.
You haven’t explained what that even means. Do you mean you just betray your teammates and quit out of games? Get high and drive warthogs into fusion coils? Lose on purpose?
Because people hate those kinds of people. People want to win games and do well. That’s the point of video games.
@Darwi From my H5 testing back in the days, DPM wasn’t taken into account which was a bit annoying considering it is the most consistent data.
Did 343 communicate about it? Have you done testing to comfirm it is used to calculate MMR?
1 Like
Again. And this time I’ll let you do the math;
8 kills and 4 deaths.
What is your KPM and KD if the match went for 8 minutes?
And then…
What is your KPM and KD if the match went for 4 minutes?
Results in your own time. But spoiler alert - they are not the same. There will clearly be a different push on the MMR from these two scenarios.
And similarly. A high KPM doesn’t guarantee a high KD. If I go (and I wish) 50 and 49 that is going to be very impressive KPM. As for the KD - not so much.
KPM trumps KD.
If you have a team’s with matching KD’s across the board - the team with the higher KPM will get to 50 first.
Game over.
More the MMR and then a stronger pull on the CSR in the next game.
No.
It just means I play at a certain level of intensity. I tend to run a bit more solo or with my mates. I grab equipment and weapons I wouldn’t usually use in ranked (eg. Sniper). I’ll play far more aggressively and take risks (as opposed to camping in a spot and waiting for all my team-mates to be in the best possible position).
The push is still to win. I could count on one hand the number of times I’ve quit or betrayed a team mate (in 20 odd years of playing Halo).
My team mates should be very happy.
Are they getting the possible Darwi? No. They are getting Darwi-lite. But I am playing as advertised. I am living up to my Social MMR (which is used to organise and match the teams).
And I’ve checked my expected kills and deaths. I am playing very chill and still getting filling my own boots.
1 Like
I haven’t seen a lot about it.
The TrueSkill2 paper mentions it - and the graph shows both kill and death rates with variances (page 15 if you need to refresh it). It discusses that the kill data has a stronger correlation. But the table at end only demonstrates the improvement in ranking from the kill data.
When it has been discussed they have mentioned that the weighting for kill data is higher. But we don’t know how much higher, or what even the KPM weighting is.
Neither KPM or DPM has correlated with my CSR changes post game in the data I’ve collected (as expected). But I have no way of checking to see how they affect my MMR. And even if we had the raw figures it would be very hard to retrograde the changes in MMR (it’s a lot of multi-pass comparisons vs opponent skill that is above my pay grade).
The fact that waypoint stats highlight both KPM and DPM (via expected kills and deaths) suggests that both are important?
My current data collection for the second half of season 2 is going to be matching opposition MMR to KPM and DPM. Hopefully I can show that they both correlate - and maybe get an idea of the relative weighting. But one of the problem’s with a relatively small data set is going to be the lack of precision with the rounding of expected kills and deaths to integers.
As an interesting aside - I just played two placement games vs similarly ranked opponents. In the first the game finished normally and my expected kills and deaths was around 15. In the second I had a team-mate quit early, followed by another about half way in. So a chunk of the game was 2v4. My expected kills and deaths for that game was only around 6. So it appears to take team numbers into account (which is actually what is described in the TrueSkill2 paper).
1 Like
The same number
They’re the same thing. If you kill 5 people in the first 30 seconds of the match or if you kill 1 person every minute for 5 minutes it’s still 5 kills.
So you play social like you’re supposed to be playing ranked?
Those numbers are nonsense. They don’t appear to have anything to do with anything. They’re randomly generated. It will say you should get 30 kills against better players and 10 against worse players.
it’s not clear from your message wheter you think that “rapid-change MMR” is a good idea, but it abosolutely should be a thing.
Otherwise people who are having both good days and bad days make that 50% chance of winning/losing very inaccurate and unreliable
And as you said rapid change MMR is also good against smurfs
He’s probably right. Some people are “chosen” by the MMR system to always carry, or almost always… it happens even in platinum/diamond matches, it’s not just an “onyx who has to carry gold teammates” problem
We want a higher win rate. I’m not saying it has to be 90% like in the halo 4 days, but at least 55-60%. Reward your dedicated players who take time to master the game
I know what you’ll say: " OMG MORE THAN 50% WIN RATE = BAD ABSOLUTELY UNACCAPTABLE (with 343 guility spark’s voice), you just want to stomp bad players!" No, it wouldn’t be like that. I just want to be on the side of 50-60% win CHANCE instead of 40-50% chance. That’s why many people sweat. They want to be the difference between a 55% win chance and a 45% one…
I’ll keep playing social even as a sweat. Ranked is stressful and even sweatier and many people don’t agree with their changes in rank match after match. Wheter they’re mistaken or not, people want to see their rank move in the direction they think is the right one (based on what they believe is the right rank for them) and they don’t like infinite’s system
Thanks for the answer. I believe I got the wrong definition of DPM.
When you say DPM you mean Death per minute or Damage per minute?
I thought it was damage because it’s a very consistent data that wasn’t taken into account to calculate MMR.
Seriously. Or are you just trolling again?
8 kills and 4 deaths.
For the 8 minute game it’s a KPM of 1 and a KD of 2.
For the 4 minute game it’s a KPM of 2 and a KD of 2.
Quite clearly the second game is going to have more influence on your rank than the first. Despite the same KD of 2.0
And if you go with 50 and 49 in a ten minute game (somewhat unrealistic obviously) - but a massive KPM of 5 - despite a KD barely above 1.
If you’re maths isn’t up to it I apologise - I would never belittle anyone for that. But if you are just being belligerent and not take well intentioned advice on how to rank up then you deserve to be stuck in ELO hell (I think that’s the old fashioned term).
5 kills in 30 seconds is actually 10 KPM.
In a 1v1 match they would hit 50 kills in 5 minutes. Your hapless opponent going at 1 KPM would only be on 5.
That’s a pretty resounding win. 50-5.
They are generated against your past performance (or that of your peers - that bit isn’t entirely clear to me yet).
But not nonsense. And if you get over your prejudice - they should be a valuable tool in working out how far you are from ranking up.
They look like they are all over the place because the KPM underpinning them is multiplied by the time of the match.
And it looks like you need to adjust further if player(s) drop out - that seems to be taken into account as well.
I’m still getting my head around it. And I only have a handful of images from old Waypoint posts to go on (wish we could easily post them here). And some of them have a more linear MMR - but are called “normalised” - so I don’t know how that changes the interpretation.
As for the MMR being able to swing. I think it’s a good idea. We know form can hinge on a lot of things - and not all under the player’s control. So having a system that can adjust, especially in Social is great. You just have to work through that first game or two which is still making assumptions from the previous sessions.
And this would explain why 343 are sticking to their guns with the CSR. It has a slower trajectory so can average out your MMR’s zig zags. Almost giving you and average MMR path.
Josh was adamant that Smurfs were made redundant in one or two games. An MMR that was free to swing on an extreme KPM would certainly fix that.
It probably varies around the current server population. And if you are in that group of “senior citizens” at the time - you are going to get batched into games where you have to carry.
The key is to embrace it. Enjoy. And the way to do this is take reality check the expected kills on Waypoint. If you are meeting those then you’ve done your bit. You can blame your team mates in peace (knowing that your MMR will not take a big hit in the process).
No, I don’t have a problem with that.
From a ranking point of view 50:50 is important.
But 45-55 is ok. Maybe even a bit more (I’m a bit on the spot here, haven’t really thought about it).
The key is that both sides need to feel they had a chance. I feel somewhere around the 40-60 mark that probably gets lost?
But I imagine a lot of your “colleagues” would be very unimpressed with even this concession. They want absolute right to stomp.
And that’s on 343.
We have no boundaries set.
You see your rank - but it’s hard to judge where you sit. As an observational bias you see the matches you do well in isolation - especially when there is an opponent or two at the rank you are coveting.
And it’s a fine line. A Platinum player will play as an Onyx 5% of the time. A Diamond will beat an Onyx 25% of the time. There’s just enough to give you a false sense of expectation.
But sadly we have no way of looking this up for ourselves. So we have no sense of perspective. So we just frustrated
2 Likes
Yeah, sorry if I caused any confusion.
DPM is deaths per minute.
I know they looked at damage in the TrueSkill2 paper. I don’t know if they specifically looked at damage per minute - but off hand I would imagine it wouldn’t give any extra information above kills per minute.
I would love to look closer at various damage ratios.
If we only had an API to access this stuff.
1 Like
Mr wulfy is legit a good player AND plays the objective. That is very rare and hard to find. I’ve lurked a couple times…, Respect bro. I almost joined last night but you were up against some serious comp I didn’t want to be a crutch
1 Like
Didn’t 343 say that some day the old waypoint posts would be available again? If you could get more info on this topic, it would be nice
We definitely agree on this one
I think this is one of the main reason players are confused. For example, an onyx having a bad day could play like a platinum, but his csr stays onyx. Then real platinum players could definitely get confused " that onyx guys should have carried me! Why are diamonds on the other team playing better than him??"
Maybe in the short term it can feel good to win after carrying everyone. But long term it discourages me.
Knowing that if i do something wrong the match is lost gives me a bit of anxiety. Lots of pressure on your shoulders
It feels good to have teammates you can count on, especially when even the enemy team knows what to do
I think These ones are a “dying breed”. Almost every game has sbmm and MMR. I know many old friends of mine who don’t play anymore because they don’t like the system and they can’t find any “unprotected” games that they also like
For me, at the end of the day, seeing something like - 7wins, 5 losses- is good enough
The worst days that don’t feel good at all are like - 2wins, 6 losses-
I think ranks should be wider. There should be more difference between a onyx and a diamond. The result is people getting apparently closer matches, even when the chance is the same. This should increase the chance of “one division only” matches, eg. a match where everyone is platinum, instead of 2 diamonds, 1 platinum, 1 gold vs a mix of platinums
Also, one more thing. I don’t like streaks when SBMM is active at all. Streaks of losses because i like to win, i don’t feel good after losing. Streaks of wins because i know that BIG LOSSES are coming after some time and there’s almost nothing i can do, because i’m already playing at my max, it makes me feel under pressure
I love a close match too, but not when it’s every single match in a playlist that is supposed to be casual it gets tiring real fast. Claiming it only annoys the “top 2-3% of players” is also incredibly disingenuous considering this game has had an incredible drop in playerbase across the board. Yes a lot of this is due to lack of content and desync, but to close your eyes and claim that strict sbmm in casual isn’t affecting the average player just because you don’t have a problem with it while there are thousands if not hundreds of thousands of complaints about it spanning the entire community is deceitful. You are the exception to the rule, not the rule.
That evidence is super short-sighted, because it doesn’t take into account the long-term consequences of sbmm. Yes in the short term if one has a terribly unbalanced match (which again, are not that common with loose sbmm) they will be more likely to log off that day. But if every day they log on and every match is a super close sweat fest, they are more likely going to drop the game in the long-term, because playing games in a social setting should be mostly just that, social.
3 Likes
Where, may I ask, are you discovering these old posts?
1 Like
I don’t troll. I’m often sarcastic, but I’m also bluntly honest.
You miss the point. The match lasts the same either way, you’ve gotten 5 kills in a 5 minute span either way. My MATH is fine.
What you’re trying to say is that an overkill at the start of the match and no kills the rest of the match is a bigger deal than kills throughout the match.
No they aren’t. It bases it on how your MMR compares to who you’re against. If you’re against people with a lower MMR it expects you to get a high volume of kills. Hence why it keeps thinking I should dog walk low tier onyx as a high Diamond, so it ironically forces a platinum on my team to keep me from winning that match or meeting the kill threshold so I don’t rank up…which is GATEKEEPING.
Nah man anytime you want to play some hit me up. Some friends and I were going for our Open placement ranks in a 4 group yesterday, it was kind of a sweaty series too.
I’ll be on a good bit this weekend if you want to try some matches
No. What he’s saying is that the kill rate (kills/minute) needs to remain high by the end of the match.
He referenced two hypothetical matches with the same KD ratio, but with the matches lasting different lengths of time. One took 8 minutes the other 4 minutes. Same KDs, but different KPMs & DPMs. The skill assessor is going to reward more for the better KPM. It’s concerned about performance over time and of course, most importantly, the match outcome.
1 Like