Pgi Messed Mm Again
#41
Posted 22 August 2016 - 09:44 PM
I would suggest they recycle the tier list and put all players back at tier 5. or for even more seperation say 10, next follow the example of other more competitive games and make a win equal to about 2 losses meaning any player not winning 30% of his matches will not progress past thier tier. and of course have that scale decrease untill at tier 1 the player has to win say 50% of thier matches to stay in the tier.
#42
Posted 22 August 2016 - 09:58 PM
The PUG matchmaking is particularly poor and some days I fantasize about uninstalling because of it..
- It's based on the idea that a couple of good players and a bunch of bad ones can have a balanced match against a team of average players. It doesn't work that way, and stomp after stomp ensues.
- The tiering system is biased so that players progress up the tiers, and much too quickly. So, Tier 1 has become meaningless, and leads the matchmaker to mix ultra good players with poor ones.
- The matchmaker makes no allowance for those levelling Mechs vs those playing ultra optimised, mastered Mechs.
- Tiers to be reset on every patch day. Apply a bell-shaped curve to players active in the past three months.
- Tier progression algorithm to be modified so that going down in the tiers is weighted equally with progression up.
- Tier Progression algorithm to be modified so that it takes more account of personal performance and less of team performance.
- Matchmaker to take into account unlevelled Mechs, and class of Mech being played (as the current challenge does).
I could write the maths/algorithm which would implement the above principles and create better matches. It's not hard or even very complicated. However, I have zero confidence that it would be adopted by PGI, so why bother.
Edited by Appogee, 22 August 2016 - 10:08 PM.
#43
Posted 23 August 2016 - 12:52 AM
I would suggest you focus on your own contribution first, if you are T1 then carry like a T1. 800+ dmg, 4+ kills on average at least. That would sway most of your matches in your favor.
#44
Posted 23 August 2016 - 01:21 AM
Cathy, on 22 August 2016 - 03:37 PM, said:
It tries to balance games so that you average out on a 1-1 win loss ratio, it looks at data and then tries to put a game together on that basis, if people have been carried for 8 wins in a row and then it decides they need to lose based on its code, and they play like muppets, rather than give a close fight, hardly MM's fault.
No it doesn't. The MM doesn't use your W/L as an input to sort players in matches. The MM uses a players PSR rating, the mech they dropped with and what servers they have ticked.
#45
Posted 23 August 2016 - 01:27 AM
I have read quite often now, that Tiers should be reset every season or so? What is the advantage of that? Cause even if I take a 6month brake I don´t forget the minimum/maximum ranges or how to shoot Uavs and so on. So it would suck incredibly hard to come back after a brake.
#46
Posted 23 August 2016 - 01:32 AM
Flitzomat, on 23 August 2016 - 01:27 AM, said:
I have read quite often now, that Tiers should be reset every season or so? What is the advantage of that? Cause even if I take a 6month brake I don´t forget the minimum/maximum ranges or how to shoot Uavs and so on. So it would suck incredibly hard to come back after a brake.
Since it can take months/100s of games to properly seed a player in a MM, reseting players regularly would defeat the point of a skill rating in the first place.
#47
Posted 23 August 2016 - 01:56 AM
Cathy, on 22 August 2016 - 03:37 PM, said:
It tries to balance games so that you average out on a 1-1 win loss ratio, it looks at data and then tries to put a game together on that basis, if people have been carried for 8 wins in a row and then it decides they need to lose based on its code, and they play like muppets, rather than give a close fight, hardly MM's fault.
I have yet to hear that PGI adjusts you relative mm based on wins and losses. War THunder tried this and got yelled down because the top players felt their W/L was being manipulated by factors not determined by their skill.
What I have noticed is that I think a tier has sub sections and I will be doing well for a couple of weeks and move up a bit and I think I got the Keystone cops on my team. The primary indicator is getting a game as soon as you hit go. Pretty much means the gates are open for another match and you can only hope you are the worse player with goods instead of one of the better trying to make silk purses from *****.
#48
Posted 23 August 2016 - 02:27 AM
Ghogiel, on 23 August 2016 - 01:32 AM, said:
No, the skill history is what would be used to redistribute players among the skill brackets. Noone (at least, noone with a brain) is proposing to 'reset everyone to Tier 5'.
A reset and rebalance would help create better separation between the 'super good who deserve to remain in the top tier' from the 'fairly good who ground their way up to the top tier' and the 'decent players who are better than noobs but nowhere near top tier'. It would similarly create bigger brackets at the bottom for those who just aren't good and should never have progressed up the tiers.
Everyone would win through better matching of skills.
#49
Posted 23 August 2016 - 02:33 AM
#50
Posted 23 August 2016 - 02:43 AM
Appogee, on 23 August 2016 - 02:27 AM, said:
A reset and rebalance would help create better separation between the 'super good who deserve to remain in the top tier' from the 'fairly good who ground their way up to the top tier' and the 'decent players who are better than noobs but nowhere near top tier'. It would similarly create bigger brackets at the bottom for those who just aren't good and should never have progressed up the tiers.
Everyone would win through better matching of skills.
Not relevent.
I quoted this "I have read quite often now, that Tiers should be reset every season or so?"
#51
Posted 23 August 2016 - 02:52 AM
PhoenixFire55, on 23 August 2016 - 12:52 AM, said:
But, that's 'not necessarily'. And 'not necessarily' is not the same as 'regularly'.
As a player, you can certainly tell/differentiate those matches from the many matches which are stomps purely because of bad matchmaking. In 'well matched' games you can observe the majority of pilots demonstrating skill and knowledge in their positioning (grouping and flanking), personal piloting, effectiveness of loadouts, use of weapon systems, etc. In the 'bad matchmaking' matches, you can observe a predominance of pilots who derp unthinkingly into bad positions, neither regroup nor move in a way that complements the team's positioning, fire their weapons well out of range.
Ie. an experienced player can tell whether their team, or their opposition, is largely comprised of strong players, or not.
PhoenixFire55, on 23 August 2016 - 12:52 AM, said:
But we're all in this thing called 'Tier 1'. And the spread of skill you see in a 'Tier 1' match now is ridiculously wide... because the tiering system is weighted towards making people go up in the tiers... and there's nowhere for the guys who are genuinely at the top to go any higher.
Noone should be on a 'Tier 1' team who doesn't know not to fire Small Lasers at 600 meters. Or who alphas and shuts down over and over as their default playstyle. Nor should the matchmaker be creating a team that features a bunch of the best players in the game all on the same side, against lesser skilled people on the other side.
The matchmaker does that because we're all 'Tier 1' and supposed to be able to 'carry like a Tier 1'. Yet, not many people can carry like Proton, the top guys in KCom and ELH etc. So creating matches based on that flawed logic is dumb.
It is reasonable for all of us who aren't at Proton's level to be concerned that the 'matchmaker is broken' ... even while we do our utmost to focus, carry, and help our team's win.
Edited by Appogee, 23 August 2016 - 03:22 AM.
#52
Posted 23 August 2016 - 02:54 AM
Your tier level is not an expression of your skill, it is an expression of how much you play.
#53
Posted 23 August 2016 - 03:01 AM
#54
Posted 23 August 2016 - 03:16 AM
Sader325, on 22 August 2016 - 09:44 PM, said:
PhoenixFire55, on 23 August 2016 - 12:52 AM, said:
Flitzomat, on 23 August 2016 - 01:27 AM, said:
#55
Posted 23 August 2016 - 03:21 AM
adamts01, on 23 August 2016 - 03:16 AM, said:
What it should be then if there are 2 skilled players and 22 potatoes searching for a match? No match at all?
#56
Posted 23 August 2016 - 03:25 AM
PhoenixFire55, on 23 August 2016 - 03:21 AM, said:
That's a straw man around an edge case. It doesn't address the regularity of poor matchmaking that can be observed daily. And, if the player numbers are to be believed, then it would almost never be the case.
But anyway, let's play 'what if your straw man was real'. At the very least, the 2 most skilled players should be on opposing teams. And the rest of the 22 players distributed evenly based on skill across the two teams. Just like we did at school when picking lunchtime football teams.
But anyway, I take it from the ongoing arguments that you make, you're delighted with the quality of matchmaking at the moment. That's great. I'm glad to know the current system is working well for you.
Edited by Appogee, 23 August 2016 - 03:30 AM.
#58
Posted 23 August 2016 - 03:31 AM
Go.
Away.
D-bag.
Signed,
Everyone.
Edited by Lemming of the BDA, 23 August 2016 - 03:32 AM.
#60
Posted 23 August 2016 - 03:38 AM
Appogee, on 23 August 2016 - 03:25 AM, said:
What I like about you guys is that when its convenient for you the population is too low for MM, but then two hours and six threads later the population suddenly becomes sufficient.
Appogee, on 23 August 2016 - 03:25 AM, said:
Indeed ... and then each skilled player will have to single-handedly carry his team.
Appogee, on 23 August 2016 - 03:25 AM, said:
Delusional much? ... Scroll back a few pages in general discussion and look through MM-related threads. Better yet find MM-related threads from back when PGI announced it.
Edited by PhoenixFire55, 23 August 2016 - 03:38 AM.
3 user(s) are reading this topic
0 members, 3 guests, 0 anonymous users



























