Jump to content

Combined Queues - Final Discoveries


849 replies to this topic

#661 Nearly Dead

    Member

  • PipPipPipPipPipPip
  • Philanthropist
  • 274 posts

Posted 29 May 2020 - 02:46 PM

As time passes I am more and more concerned that they are going to cheap out and keep the tiers exactly as they are and just try to modify the existing formula for moving players up and down and call it zero sum.

It needs to be a new calculation that measures individual player contribution to victory much more accurately than the colorful boxes that have been posted here in the past, with big, coarse groups of "performance". We have computers now, saying you are going to do something like "250-299 = skill bracket +, 300-349 = ++" is a bit crude.

The issues I see are:

1) The resolution is too coarse. There are 5 tiers, how do you combine them to speed things up if you need to, but still have good (tighter skill groups) matches if you can? You can't, you need twelve, twenty four or some other large number of tiers so you can combine them when the MM is slow. Even better would be to program the MM to look at the players as a continuous spectrum and divide them up on the fly into as many or few groups as needed to match the desired drop time.

2) The skill curve is a J curve. I don't care if MM wait is a minute or even two or three, there needs to be separation of skill groups, especially the way that high individual skill + premade teams multiply each other.

A premade of casual, wanna have fun with by buds players in tier 3 is a far different thing from a premade of Jarl's list 97%ers. Those top players need to be playing other top players, not lumped in with tier 3 players. Motivation is different, talent is different, experience is different, mechs are different, C&C is different. All those things add up.

3) Any system accumulates error over time. Measure performance over a rolling period of time or number of games with the oldest games or months dropping off. Ability and experience change over time. People get older, (or grow up depending on where they are when they started), PCs get old, or replaced. Eyes get older, glasses get replaced, people discover gaming mice with sensitivity buttons. Make it long enough to be representative but short enough that people can move up and down to find their comfort zone in a reasonable time or number of games.

edit: Because spelling matters.

Edited by Nearly Dead, 29 May 2020 - 02:51 PM.


#662 evil kerensky

    Member

  • PipPipPipPipPip
  • Ace Of Spades
  • 143 posts

Posted 29 May 2020 - 03:13 PM

Yes, do the reset pull the lever on that immediatly.

#663 Collin Roseblade

    Rookie

  • Survivor
  • 3 posts

Posted 29 May 2020 - 03:14 PM

I also want to respond to this Dev comment:

Quote

Just want to touch on this. It's known that MWO has a small playerbase now. Even if we do start to separate players into more finite tiers, the release valves would still have to open up fairly wide to get matches to kick off.


I get that MWO playerbase has been diminishing for quite a while. I cannot speak for the consumer base as a whole, but I have at least 7 friends that I tried to get into the game who all stopped playing because the queue time for us to play together in the same match was 10+ minutes. A big draw to a multiplayer-only game with no solo campaign is the ability to play with friends. Until the combined queue goes through, casual play with friends just isn't viable.

So failing to make important changes to the system "Because we don't have enough players for that to matter right now" is definitely shooting yourself in the foot. Fix the system so that it will support growth, or else you're just waiting for your product to die.

And I really think the improvements you are making will help the playerbase grow. Already, I have several friends ready to give the game another shot once this is all hammered out. I'm sure I'm not the only one.

Edit:Punctuation

Edited by Collin Roseblade, 29 May 2020 - 03:16 PM.


#664 Carl Avery in casual attire

    Rookie

  • 2 posts

Posted 29 May 2020 - 03:49 PM

edit: lol whoops, wrong account. that'll teach me to browse the forum while logged in on my alt.

Edited by Carl Avery in an old mech, 29 May 2020 - 03:53 PM.


#665 Carl Avery

    Member

  • PipPipPipPipPipPip
  • 207 posts

Posted 29 May 2020 - 03:51 PM

View PostCollin Roseblade, on 29 May 2020 - 03:14 PM, said:

I have at least 7 friends that I tried to get into the game who all stopped playing because the queue time for us to play together in the same match was 10+ minutes. A big draw to a multiplayer-only game with no solo campaign is the ability to play with friends. Until the combined queue goes through, casual play with friends just isn't viable.


The queue merge happened a few weeks ago, and it remains that way, so grab your friends and hop in.

#666 ThunderKats

    Member

  • PipPip
  • Liquid Metal
  • Liquid Metal
  • 20 posts

Posted 29 May 2020 - 05:05 PM

To solve tonnage issue groups should be set to Bravo or Charlie base on how much tonnage you give them. What's truly hurting the game is having them on Alpha lance, for a start you are taking the scout lance away from the game and not only that but also in many occasions you see several assaults in alpha vs 3light+1medium on the other side of the team for the lack of a group. Tier 1? What's that? Because in my eyes doesn't represent anything close to PSR. Only way I imagine someone having skill on this game is base on his average score on the past 100 matches (real time skill balance and also easy to do since it has been done before when doing rank events). You don't even need to mess around with the tier system this way, you can keep your Tier 1-2 as a measure of experience and combine it with this average score, even if the thing is invisible for the players it should help matching the players better. Avg Score x Tier = PSR...
Note: Most broke thing about the tier system is how time affect the result, that's why old players coming back to the game have a hard time achieving higher tier despite they overall good performance
. I vote for a reset and some new system taking place


Posted ImagePosted ImagePosted ImagePosted Image

#667 Akillius II

    Member

  • Pip
  • 12 posts

Posted 29 May 2020 - 05:05 PM

Fix PSR algorithm to be based on players skill, & unlike the current match win/loss exp tracker.
Rename tiers: 5=A, 4=B, 3=C, 2=D, 1=E
Ensure QP, CW, and Solaris/etc all have and use their own separate PSR.
Then reset PSR.

Make a clear announcement regarding - At what player population number will allow PGI to change match making Ques to be tiered and group vs solo separate again (so that the tiers mean something).
- OR -
At what time of day would the Ques be by tiers, and when its returned to a single massive Que-ball.

Add at least 2 new maps.

Add something like Easy Anti-Cheat

Fix Faction Play (I stopped playing CW when changed to suck at end of 2016, so ask other players.)


Then make a big advertising splash and highlight the improvements:
- Player Ratings Reset & PSR Algorithm changed
- Join Quick Play matches based on tiers, or as a tierless solo+group free for-all.
- Added Easy Anti-Cheat system
- Faction Play made great fun again
- New Maps added



Only then can PGI find out if MWO gets its second wind when some old players return.
Or if new players join and stay.

Edited by Akillius II, 29 May 2020 - 05:18 PM.


#668 Bowelhacker

    Member

  • PipPipPipPipPipPipPip
  • Hero of Marik
  • Hero of Marik
  • 922 posts
  • LocationKooken's Pleasure Pit

Posted 29 May 2020 - 05:19 PM

Just reset it it then keep everyone's tier a permanent secret from everyone.

#669 BROARL

    Member

  • PipPipPipPipPipPip
  • Marshal
  • Marshal
  • 304 posts
  • Locationcommunity warfare

Posted 29 May 2020 - 05:23 PM

just make mech5 PvP and you can riddle all this out somewhere that matters, somewhere with a future?
spending staff time here is an expense, spending time in mech5 is an investment, just make it PvP for goodness sakes...

#670 L1f3H4ck

    Member

  • PipPipPipPipPipPipPip
  • 738 posts

Posted 29 May 2020 - 05:57 PM

View PostKilowatt Girl, on 29 May 2020 - 11:47 AM, said:

My first post if it works. Thanks for the game. Now it works a lot better with my fiber optic cable than Mw3 did over my rural telephone line. I have been in the max of tier one for quite a while. I still have considerable to learn about being a good team player but being in tier 1 has helped me with that. Thanks for that too. When I do something dumb for the 10,000th time I watch my mates around the world planning, positioning and good shots while I just finished spraying all over the place landing in a bad spot to meet my early death on to spectator view anxious to see that I didn't damage my mates with that last stupid move. When the bar first entered tier 1 on my screen I started trying to do better for the first time in years to spite lack of skill. My problem is that I played Mw3 on to now because I like to build mechs and try em out. That is it for my interest. I have built a lot of mechs. I can just build them all day. I do more of that offline now. Lately I found myself building 3 and 4 AMS supports and sticking right in the middle of the highest density of my mates on my mini map and trying but failing to wait until the time is right and learning. When I get disappointed with my very low damage scores I switch to my medium mech that can go off by itself and at the very least take you with me if there aren't more than two no matter what tier your in. Even here I try to stay with the assaults in density but usually never end there. I have a lot to learn. This is just how the game is for me. Mates were often rude to me in the lower tiers but since I got in tier 1, these players have manners and are fun. I love a match with chaotic harmony or whatever he is. I don't care what tier I am in but I know I will miss the nice players having fun in any tier. Thank you. Where do we go from here? I might have the most double heat sinks and hope to see you there.


Ey, hope to see you there too, glad to have you aboard. What a great time to be alive!

#671 Warning incoming Humble Dexterer

    Member

  • PipPipPipPipPipPipPipPip
  • Bridesmaid
  • 1,077 posts

Posted 29 May 2020 - 06:14 PM

View PostEast Indy, on 29 May 2020 - 11:12 AM, said:

So, in workweek bustle I skimmed what Paul originally wrote and assumed the "formula" would be something like Trevelyas' 5-year-old suggestion to score players in context with each other's performance each match, then apply that over time to the entire population.

No matter how unbalanced a match is, what all 24 players players have in common with their 11 teammates, is that each one is facing the exact same 12 enemy players.

So an easy and fair way to rank a player is to check how well he performed within his own team when facing the exact same opponent team, by ranking the individual players from each team from 1st to 12th.

Meaning a player with an average rating around 1 is always performing near the top performance of his team, and a player rated around 12 is always performing near the bottom performance of his team.

In which case you're correctly singling out and evaluating the performance of the individual player within his team, instead of incorrectly praising or blaming the individual player for the good or bad collective performance of his team.

So you get an average performance value that is strictly between 1 and 12, which can even be directly translated into a percentage value (1=100%, 12=0%) : It doesn't get easier, fairer or even clearer to understand exactly what a ranking means, then something like this.

Also you only take into account the last ~100 matches, to make that value stick close to the current level of performance.

Edited by Warning incoming Humble Dexterer, 29 May 2020 - 06:28 PM.


#672 Scout Derek

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Divine
  • The Divine
  • 8,022 posts
  • Google+: Link
  • Twitch: Link
  • LocationSomewhere where you'll probably never go to

Posted 29 May 2020 - 06:21 PM

View PostAndrzej Lechrenski, on 29 May 2020 - 03:37 AM, said:


I don't think you care about "new players". I think you care about new worshipers. It's sad.

Matchmaking simply IS bad right now. But stooping to lying about it? WTF? WHY?

I do though, it's just unfortunate that you're still keeping a closed mind about it. And worshipers? what? you're actually making me laugh, but please, do go on about that, I find it amusing since like, literally most stuff that comes from that discord is actual meme content for the most part. There IS actual talk about specific things if asked, but mostly just joking and banter there. It's like a pub, people go there to socialize and start stuff too at times :D

And why would I lie about MM? I have nothing to gain from it, unless you think I have some ulterior motive to sabotage MM which clearly a majority of people know I'm not that kind of person. But you wouldn't know that would you? To you I'm probably just some cheater :D

#673 Collin Roseblade

    Rookie

  • Survivor
  • 3 posts

Posted 29 May 2020 - 08:15 PM

View PostCarl Avery, on 29 May 2020 - 03:51 PM, said:


The queue merge happened a few weeks ago, and it remains that way, so grab your friends and hop in.



Oh! I assumed it was still in public test! Time to rally the gang!

#674 David Sumner

    Member

  • PipPipPipPipPipPip
  • FP Veteran - Beta 1
  • 470 posts
  • LocationAuckland, New Zealand

Posted 30 May 2020 - 12:13 AM

Could we make it a "betting exchange"? Say max 1,000 PSR.
Team 1 all put up 1% of their PSR = 72 points
Team 2 all put up 1% of their PSR. = 83 points

Say if you are in a group, you have to put up an extra 0.1% for every member in the group.
So a four man each put up 1.4% instead of 1%. Not sure you'd be able to recoup that on a win very well though.


If you win, your team keep 10% of what you put up.
Team 1 wins, they are now putting up 64.8 points

Total match score for team 1 = 2,700
Total match score for team 2 = 1,900

Team person in team 2 loses their 1%, then gains 64.8 * match score / 1,900
Team person in team 1 loses their 1%, then gains 83 * match score / 2,700


Assuming no rounding errors, that's zero sum.
And you could run a job say weekly to tweak the average back to a target point if you wanted.

Winning players lose less and gain more per match on average.
Losing players lose more and gain less per match one average.
Better players risk more per match on average.
Worse players risk less per match on average.

"Better" players on each team gain more than the "worse" players.

Edited by David Sumner, 30 May 2020 - 01:48 AM.


#675 Mister Bob Dobalina

    Member

  • PipPipPipPipPipPipPip
  • 674 posts

Posted 30 May 2020 - 01:05 AM

I am all in for a Tier reset and even a complete PSR overhaul. Stuck in T1 and simply not capable to play anything above T4-T3.

As for more desired changes I stick more or less with NGNG:

https://www.youtube....eature=youtu.be

(and of course: #MakeTheHatchetman Posted Image )

Edited by Hurbie, 30 May 2020 - 01:08 AM.


#676 Paul Meyers DEST

    Member

  • PipPipPipPipPipPipPip
  • CS 2024 Top 25
  • CS 2024 Top 25
  • 545 posts

Posted 30 May 2020 - 01:25 AM

Well, after some playing and having fun with my mates in qp, i still think it would be much better to let solos get into group queue instead of the other way round. The impact of the group in solo queue ist to much: weight, skill and group setup is bugged.

#677 AnAnachronismAlive

    Member

  • PipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 448 posts

Posted 30 May 2020 - 01:41 AM

View PostBROARL, on 29 May 2020 - 05:23 PM, said:

just make mech5 PvP and you can riddle all this out somewhere that matters, somewhere with a future?
spending staff time here is an expense, spending time in mech5 is an investment, just make it PvP for goodness sakes...


THIS!

#678 FRAGTAST1C

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Fighter
  • The Fighter
  • 2,946 posts
  • LocationIndia

Posted 30 May 2020 - 03:56 AM

Still no word on how PGI will be assessing the threat value of a group.

Still no mention on whether the spawns will be fixed or not (re: Assaults spawning in Alpha lance despite not being in a group)

Still no intention to solve the lopsided tonnage balance.

Still no interest in keeping Siege in all phases of a conflict in FW alongside other modes.

#679 Password1234

    Member

  • PipPip
  • The Referee
  • The Referee
  • 29 posts

Posted 30 May 2020 - 04:30 AM

Do it. Its long overdue.

#680 Anharn

    Rookie

  • Little Helper
  • Little Helper
  • 4 posts

Posted 30 May 2020 - 06:01 AM

Variation on the 8/8/8 idea, you could scale the points within the 8s

Top 4 double point gain
Next 4 Normal gain
Next 8 remain unchanged
Next 4 Normal point loss
Bottom 4 double point loss

:)





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users