357 replies to this topic

[Swamp Ass MkII]

Thinking about it, and it just dawned on me. Make PSR adjustments per round, based on match score rather than damage done...

[Horseman]

RRAMIREZ, on 07 October 2020 - 01:17 AM, said:

Totaly agree, except that IIRC T1 (and at that moment they were not as many as today) complained the most... (and T5 in a lesser proportion)
So maybe change to "people whined they don't care if they sealclub T3 Cadets"...
or just maybe "people whined"

No, idiots whined about matchmaking not being fast enough when it was good. So they got fast matchmaking that is ****. Now they're whining the matchmaking is ****.
Like, *******, please... Fast or good, pick one.

[Kodyn]

Horseman, on 05 November 2020 - 01:52 PM, said:

No, idiots whined about matchmaking not being fast enough when it was good. So they got fast matchmaking that is ****. Now they're whining the matchmaking is ****.
Like, *******, please... Fast or good, pick one.

The irony being that not only were matches faster to get into, but better matched...now everything is worse. Used to be you could just re-queue after 20-30 seconds if you didn't insta-pop, and you'd get a match near immediate. Now no matter what, you wait. Doesn't matter what % weight class is lowest or highest, doesn't matter if you restart the queue, you wait. And then you have a generally horrible, lopsided match, win or lose. God I miss the early spring...

[RRAMIREZ]

Horseman, on 05 November 2020 - 01:52 PM, said:

No, idiots whined about matchmaking not being fast enough when it was good. So they got fast matchmaking that is ****. Now they're whining the matchmaking is ****.
Like, *******, please... Fast or good, pick one.

Why do you start by "No"? Edit: after quoting me
On the point that it was a bad move to open valves wide to fasten match making, we agree.
I was on the side of good > fast (I'm not sure it would change something now).

I was just giving some perspective on "who" asked for it.
At that time T1 and T5 was low pop (and at that time we had metrics about tier population) because "climbing" was still on going.
And people complaining of MM time was from these population rather than the ones who "don't care if they get stomped". That's all I'm stating.

On a general point of view, I would appreciate current Tier population info by the way.

Edited by RRAMIREZ, 06 November 2020 - 02:14 AM.

[Horseman]

RRAMIREZ, on 06 November 2020 - 02:14 AM, said:

I was just giving some perspective on "who" asked for it.

And so was I. What I've observed was mainly casual mechdads asking for it, with their usual depth of understanding and comprehension of long-term consequences (read: negative).

Quote

And people complaining of MM time was from these population rather than the ones who "don't care if they get stomped". That's all I'm stating.

We seem to remember that differently. The people stating that they don't care if stomps increase mostly weren't T1s. It's the usual lack of foresight: they expected an increase in imbalanced matches to somehow favor them rather than throw them into the grinder. We all know which of those actually happened.

[RRAMIREZ]

Horseman, on 06 November 2020 - 05:02 AM, said:

We seem to remember that differently. The people stating that they don't care if stomps increase mostly weren't T1s.

True ^^
But I was stating about those who complained about time to drop...

[Vlad Striker]

'till MM will not respect weapon accuracy individual to playes there will no success I think.PSR have to be corrected on each weapon confuguration chosen to game match. Some guys good with AC, some with Lasers. When somebody good with lasers and bad with AC chooses AC, PSR need to be lowered, for example. Also some guys plays well on ligh mech and bad on assault mechs and so on.
Average temperature on hospital is a dead end.

[Gagis]

Vlad Striker, on 07 November 2020 - 11:28 AM, said:

'till MM will not respect weapon accuracy individual to playes there will no success I think.PSR have to be corrected on each weapon confuguration chosen to game match. Some guys good with AC, some with Lasers. When somebody good with lasers and bad with AC chooses AC, PSR need to be lowered, for example. Also some guys plays well on ligh mech and bad on assault mechs and so on.
Average temperature on hospital is a dead end.

Those differences are so small that they make absolutely no difference in a system with just 5 tiers. Maybe we'd see some difference if the system has 500 tiers instead. Then someone might be tier 465 with lasers and 466 with dakka.

The skill gaps between players in MWO are so astonishingly wide that nuance such as skill with specific weapons or different weight classes don't really come to play much outside of maybe Div A competitive play.

[DeadWeight18]

When will the Match Score triggers and their values be reviewed?

[crazytimes]

DeadWeight18, on 10 November 2020 - 05:07 AM, said:

When will the Match Score triggers and their values be reviewed?

I believe in about 90 days™.

[RowanE83]

DeadWeight18, on 10 November 2020 - 05:07 AM, said:

When will the Match Score triggers and their values be reviewed?

Is there any reason that matchmaking is not simply based on win/loss ratio? Its uncomplicated, so easily to understand and presumably to implement.

Obviously some refinement like minimum number of matches before calculating a ratio higher than 1... some sort of weighting to make the most recent say 50 matches, or matches played this month, to count more than matches from 5 years ago.

Something like that.

[Horseman]

RowanE83, on 15 November 2020 - 12:30 AM, said:

Is there any reason that matchmaking is not simply based on win/loss ratio? Its uncomplicated, so easily to understand and presumably to implement.

Sample size. When you have players who play an hour every other sunday, the sample size is too small to provide much more than noise

[GARION26]

RowanE83, on 15 November 2020 - 12:30 AM, said:

Is there any reason that matchmaking is not simply based on win/loss ratio? Its uncomplicated, so easily to understand and presumably to implement.

Obviously some refinement like minimum number of matches before calculating a ratio higher than 1... some sort of weighting to make the most recent say 50 matches, or matches played this month, to count more than matches from 5 years ago.

Something like that.

It's also a model that causes regression to the mean of the matchmaking statistic. You can't differentiate in that theoretical matchmaker between a current tier 1 who has a 50% win loss ratio playing against other Tier 1's to a tier 5 playing tier 5's with a similar WLR.


It's not a great matchmaker in that sense.

[Sawk]

well i am hoping we get the KICKERS, installed soon, the new year JAN, see how it goes, the match maker, is acting really weird, or there changing something, on the fly testing, the core values were left blank, i think --
i seem to get a string of good matchs, then the bottom drops out, i am thinking there playing with weight class some, or maybe opening valve too tiers, or both, the other thing is 4 man drop, when i hit wall, i see less 4 man groups, the win/loss goes bad for me, or i get a group of 4 in lights, and they can not hit crap.

Sawk

[Horseman]

Sawk, on 10 December 2020 - 04:36 PM, said:

well i am hoping we get the KICKERS, installed soon, the new year JAN, see how it goes, the match maker, is acting really weird, or there changing something, on the fly testing, the core values were left blank, i think --
i seem to get a string of good matchs, then the bottom drops out, i am thinking there playing with weight class some, or maybe opening valve too tiers, or both, the other thing is 4 man drop, when i hit wall, i see less 4 man groups, the win/loss goes bad for me, or i get a group of 4 in lights, and they can not hit crap.

The match maker has been deteriorating since they opened up the tier valves. That impacted the PSR changes because high ranked players are mixed with much lower ranked ones, the latter driving down the match score average and making it easier for the merely mediocre to climb up the tiers - effectively destroying the entire point of the tier system.

Edited by Horseman, 10 December 2020 - 11:24 PM.

[Sawk]

ok horseman, thats sounds good, but that is not what i am seeing.
When my wall hits i am not, getting higher tier players, i am getting lower tier players, maybe the other team gets loaded with higher tier, but not the team i am on, its crazy i saw 1 team in 2 lights and 2 meds, just drive in to kill and were dead, in 3 minutes of start, no one was saying word, i tryed shooting, no one else moved, and they hunted me down, and i was like 13th on the list pew pew.
other times its a straight up over run, like i have seen with guys that can build, some dakka, med and high heavies speed CLAN mechs, very good shots, i have played with guys that can just rip a flanging movement apart, but now i play solo and i think maybe its based in the tonnage i play, and amount of players for that night.
so i disagree with you is letting folks go up in TEIR level, its more down thing, and why its flipping this weird 1 lopsided drop thing, i have no props playing guys that are good shots - real poke and shoot pilots, in mechs that hit for more 30 points or more, and they move or share anyway.
it maybe time to stop playing a heavy or bigger, when i hit the wall, and move to some high speed, or med with 6 med lazers, ect ect ect

SAWK

[MrFatBard]

i think you guys should release the fox mech to the game since you're all doing IIC variants and stuff since it is "mixed tech" you could have like clan and inner sphere variants and probly everyone would be pretty hyped