Jump to content

Potato Problem And The Match Maker


112 replies to this topic

#41 Zergling

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Angel
  • The Angel
  • 2,439 posts

Posted 07 August 2017 - 11:27 AM

View PostRussianWolf, on 07 August 2017 - 11:19 AM, said:

fine, increase it to 10 or 15 minutes. Then what? The point is you are getting games with T3s and below because the population isn't large enough to have "competitive" games all the time and queue times that are reasonable. What is your solution?


As above, you're not presenting a solution, and it is a null argument because you know Tier bracketing won't work.
I know it won't work either, which is why I've never presented it as a solution.

But here is my solution: turn off the matchmaker entirely, make it completely random. No player gets punished for being good, no player gets rewarded for being bad.
Good players will still lose battles due to bad teammates, but it will be less than now.



View PostGalenit, on 07 August 2017 - 11:19 AM, said:

Hating someone because he is less seriously playing a virtual computer game in his spare time then you?

If you dont understand whats wrong with that, you have a real rl problem ...


Ah huh, so you feel you can judge what other people should or should not take seriously, and that you are also qualified to say other people have 'a real rl problem'?

Nice kindergarten-level psychoanalysis there Dr Freud.



View PostGalenit, on 07 August 2017 - 11:19 AM, said:

Lots of games are able to seperate them to some degree or total,

using a working matchmaker, a comp-mode, leagues, ...


Guess what? All those options require a large playerbase, which MWO does not have, and will never have.

#42 Galenit

    Member

  • PipPipPipPipPipPipPipPip
  • 1,198 posts
  • LocationGermany

Posted 07 August 2017 - 11:42 AM

View PostZergling, on 07 August 2017 - 11:27 AM, said:

Ah huh, so you feel you can judge what other people should or should not take seriously, and that you are also qualified to say other people have 'a real rl problem'?

Nice kindergarten-level psychoanalysis there Dr Freud.


I dont judge what others should or should not take seriously,
but the people that call others "potatos" do that.

If you want to destroy players that take a game not seriously enough,
you have a rl problem, if you want to talk "freud":

Quote

Sigmund Freud defined hate as an ego state that wishes to destroy the source of its unhappiness

Maybe you mean you hate to play against players not matching your skill,
but then you would hate the matchmaker, not some random guy playing a game?


View PostZergling, on 07 August 2017 - 11:27 AM, said:

Guess what? All those options require a large playerbase, which MWO does not have, and will never have.

There is a leaderboard, their are average matchscores,
let only the first 15% play together and then the last 35% and then the middle.
Three dynamic tiers, no tier-crossover and no xp-bar or psr is needed.
At the beginning of each season you start with zero,
but your last average matchscore is used for your first match in the new season.
Simple system, not based on tier or playernumbers.

Edited by Galenit, 07 August 2017 - 11:48 AM.


#43 Brain Cancer

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • 3,851 posts

Posted 07 August 2017 - 11:52 AM

Quote

Okay, say PGI flips the switches and you only play T1 and T2 players (not that they are all "good") and your queue time increase to 5 minutes. Will you be happy?


No, because PGI deliberately stuffed T2,even T1 with potatoes. In fact, the PSR system "rewards" lots of play with higher tiers due to it's massive positive bias- it's an exp bar after all.

The first thing that needs to happen is PSR having harsher requirements to gain/lose/maintain as tier increases. If tiering is to matter in the matchmaker, it has to actually represent improved skills, not just "I have lots of time to play, randomly shoot people, and hope I get carried again."

The second thing is reducing games to the players available -before- opening the valves to let the spuds drop. Got 24 players of the right tier? Great, 12v12. Not enough? Try an 8v8 or even 4v4. Only after that fails should lower tiers be injected in. Besides, smaller group sizes also act to change match dynamics, meaning more variety in same.

The third thing is expanding the match data available to players. We need to see all the relevant scoring for players, not just W/L + K/D + match score. KMDD, is he an assist machine with lots of damage or a paint-scratcher farming score by max range ERLL fire? This sort of thing can be better understood,and fed back into the scoring system to better reflect what efforts really do reflect "git gud".

#44 RussianWolf

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 2,097 posts
  • LocationWV

Posted 07 August 2017 - 12:03 PM

View PostZergling, on 07 August 2017 - 11:27 AM, said:


As above, you're not presenting a solution, and it is a null argument because you know Tier bracketing won't work.
I know it won't work either, which is why I've never presented it as a solution.

But here is my solution: turn off the matchmaker entirely, make it completely random. No player gets punished for being good, no player gets rewarded for being bad.
Good players will still lose battles due to bad teammates, but it will be less than now.


Randomize it? wouldn't that basically guarantee that you get T5s on your team? Not seeing anything that makes sense there if you are wanting more competitive matches? So I'm no longer sure what you even want.

I didn't claim to have a solution because I understand that the problem isn't solvable from the MM.

View PostBrain Cancer, on 07 August 2017 - 11:52 AM, said:

No, because PGI deliberately stuffed T2,even T1 with potatoes. In fact, the PSR system "rewards" lots of play with higher tiers due to it's massive positive bias- it's an exp bar after all.

The first thing that needs to happen is PSR having harsher requirements to gain/lose/maintain as tier increases. If tiering is to matter in the matchmaker, it has to actually represent improved skills, not just "I have lots of time to play, randomly shoot people, and hope I get carried again."

The second thing is reducing games to the players available -before- opening the valves to let the spuds drop. Got 24 players of the right tier? Great, 12v12. Not enough? Try an 8v8 or even 4v4. Only after that fails should lower tiers be injected in. Besides, smaller group sizes also act to change match dynamics, meaning more variety in same.

The third thing is expanding the match data available to players. We need to see all the relevant scoring for players, not just W/L + K/D + match score. KMDD, is he an assist machine with lots of damage or a paint-scratcher farming score by max range ERLL fire? This sort of thing can be better understood,and fed back into the scoring system to better reflect what efforts really do reflect "git gud".

I addressed that in post #40... try to keep up. ;)

#45 Atomic Hamster

    Member

  • PipPipPipPipPip
  • Bad Company
  • 108 posts
  • LocationUK

Posted 07 August 2017 - 12:09 PM

Tbh as a proud potato myself, I found myself chuckling at OP because it gave an amusing description of a lot the crazy things you see over and over again in PUGs. I've seen potato-hate and that isn't it. The only thing is, the players doing those things may not know any better unless someone tells them (nicely!). Maybe every loading screen should have a prominent tip-list of basic things not to do in a match..

As for totally segregating players by skill level, apart from the impracticalities pointed out in the posts above, I've (occasionally) seen a disorganised rabble of confused newbies turned into a winning force by a few experienced callers who had the good humour and patience to do this; for those newbies, 1 match like that could teach them more than dozens of matches where everyone on the team just does their own thing. Absolute segregation by skill would remove that possibility.

Edited by Atomic Hamster, 07 August 2017 - 12:17 PM.


#46 Zergling

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Angel
  • The Angel
  • 2,439 posts

Posted 07 August 2017 - 12:14 PM

View PostRussianWolf, on 07 August 2017 - 12:03 PM, said:

Randomize it? wouldn't that basically guarantee that you get T5s on your team? Not seeing anything that makes sense there if you are wanting more competitive matches? So I'm no longer sure what you even want.

I didn't claim to have a solution because I understand that the problem isn't solvable from the MM.


The problem isn't competitive matches, because matches aren't competitive right now, and they won't ever be. The only way to make competitive matches is with premade teams, or some very restrictive player bracketing (and bracketing won't work in MWO, as discussed previously).

The problem is high skill players are deliberately lumped with a higher than normal proportion of bad teammates. It is quite frustrating for any player to be performing well and still lose the battle because of poor performing teammates; for high skill players, that occurs much more frequently.

In a random matchmaker, players will still rant and rave about bad players doing stuff that makes their team lose or detracts from their enjoyment of the game, but it wouldn't be as bad as it is now.

Edited by Zergling, 07 August 2017 - 12:14 PM.


#47 RussianWolf

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 2,097 posts
  • LocationWV

Posted 07 August 2017 - 12:44 PM

View PostZergling, on 07 August 2017 - 12:14 PM, said:


The problem isn't competitive matches, because matches aren't competitive right now, and they won't ever be. The only way to make competitive matches is with premade teams, or some very restrictive player bracketing (and bracketing won't work in MWO, as discussed previously).

The problem is high skill players are deliberately lumped with a higher than normal proportion of bad teammates. It is quite frustrating for any player to be performing well and still lose the battle because of poor performing teammates; for high skill players, that occurs much more frequently.

In a random matchmaker, players will still rant and rave about bad players doing stuff that makes their team lose or detracts from their enjoyment of the game, but it wouldn't be as bad as it is now.

still not understanding your logic.

If we accept that better players will more quickly rise to T1 regardless of win/loss (since they will likely rise even in losses).

And we accept that under the current MM, your odds of being in a match with more T1 players is greater than if random.

How would randomizing the MM provide you less often with the "bad" players?


Yes, Tier system is screwed. Has been since day one as I and many others have pointed out. But "Better" players rise more quickly than "bad" players even in PGI's screwed up system.

#48 Zergling

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Angel
  • The Angel
  • 2,439 posts

Posted 07 August 2017 - 12:55 PM

View PostRussianWolf, on 07 August 2017 - 12:44 PM, said:

How would randomizing the MM provide you less often with the "bad" players?


The matchmaker works by trying to average out the skill level of the teams. It will balance the few high skill players on a team, by teaming them with much lower skill level players, to produce an average skill level to match the enemy team.

While that means each team would (theoretically) have an equal chance of winning, it means the high skill players will usually be teamed with a greater number of low skill teammates than normal.

Eg, an average player would, on average, get teammates that are about 50% good, 50% bad.
A high skill player however, might get something like 30% good, 70% bad.

So while each team theoretically has an equal chance of winning, what it means is that a high skill player has to play some level above average just to maintain a 50/50 chance of winning.
In effect, a high skill level player has to put in more effort to receive the same rewards (winning) than a player of lesser skill.

Edited by Zergling, 07 August 2017 - 12:57 PM.


#49 Jack Booted Thug

    Member

  • PipPipPipPipPipPipPip
  • IS Exemplar
  • IS Exemplar
  • 549 posts
  • LocationSan Diego

Posted 07 August 2017 - 12:59 PM

View PostSpr1ggan, on 06 August 2017 - 07:04 AM, said:

You forgot the moron that is last alive with no chance of doing anything. That simply runs around and hides in hopes of protecting his K/D. Locking everyone's mechs in that match and wasting their time.


There are names for this too: [redacted]
If you ever seen either of these two in a match, you'll know what i'm talking about. They also qualify as "snipers".

Edited by McValium, 07 August 2017 - 01:13 PM.
no Naming/Shaming


#50 OrmsbyGore

    Member

  • PipPipPipPipPipPip
  • 200 posts

Posted 07 August 2017 - 01:07 PM

View PostBelacose, on 06 August 2017 - 11:41 AM, said:


The good race car drivers know that the proper role of an assault is to be the proverbial lamb on a stake as it can in turn offer up some juicy back shots upon the enemy's own lambs. You know, sacrifice your wing man for glory kills.


No, the good drivers go left and support their assaults, while preparing to meet the enemy's 8 fastest mechs with 12 mechs in defensive positions

#51 OrmsbyGore

    Member

  • PipPipPipPipPipPip
  • 200 posts

Posted 07 August 2017 - 01:14 PM

View PostInfinityBall, on 06 August 2017 - 04:02 PM, said:

I was in a QP game last night populated by an incredible bunch of p- ... cowards. Manifold domination, and someone on our team convinced everyone but me to sit in the basement. I wound up fighting outside, 1 kill before dying.

So I spectated. For the entire game. They sat there in the basement doing nothing, and the enemy never attacked until a couple sniped one of them in the last minute. The 11 of them got no kills or KAs "playing" for 15 minutes


I don't know what it is about this map, but people absolutely refuse to help take the top, and a stupid nascar usually breaks out while the enemy team just shoots down on us and kills us 1 by 1. at least that's how it seems to me; ut probably happens a bunch to the enemy team and my team stomps them, but i don't notice those as much and they bother me a bit less :)

View PostAlexander of Macedon, on 06 August 2017 - 10:18 PM, said:

"Let he who has a WLR over 50% point the first finger," eh? Posted Image


eh, don't read too much into stats, it's a team game so often your performance is out of your control.

#52 OrmsbyGore

    Member

  • PipPipPipPipPipPip
  • 200 posts

Posted 07 August 2017 - 01:23 PM

View Postmike29tw, on 07 August 2017 - 01:45 AM, said:


This one I hate the most.

"Push" isn't the magic I win button. You don't push when the terrain favors their position. I've seen countless time a friendly assault calls a push and wade straight up into an enemy firing line and blame his team for not following.


This is an interesting point. If the enemy team has the favorable terrain, one might infer, then it is impossible for the opposing team to win unless the team with the favorable terrain voluntarily leaves said terrain. This, of course, is not true.

You are absolutely correct that "push" isn't the magic win button, however many seem to assume that push= walk in a straight line towards the enemy firing line while out of cover. This is likely a failing on the part of those of us who call for pushes (myself included) not explicitly stating the route that should be taken

#53 Alexander of Macedon

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,184 posts

Posted 07 August 2017 - 01:38 PM

Indeed. If the enemy team has superior terrain, pushing is almost invariably the proper move. The only problem is with people who equate 'push' with 'charge in a straight line directly at the enemy'. You should know maps well enough to pathfind through the best cover and concealment available before breaking out and closing for a brawl, or setting up a firing line in an advantageous position.

Polar is a prime example of this, it's full of good avenues to push, but people hate and fear that map because they don't understand how easy it is to use those innumerable folds, valleys, and hills to move without being seen or shot.

View PostOrmsbyGore, on 07 August 2017 - 01:14 PM, said:

eh, don't read too much into stats, it's a team game so often your performance is out of your control.

I've gone on at length about why I don't generally pay much attention to stats on the positive side of things, but games like this, statistically speaking, will normalize at 1:1 W/L over time. If your WLR is significantly worse than 50% over multiple months you can't blame that on external factors.

View PostGalenit, on 07 August 2017 - 11:42 AM, said:

There is a leaderboard, their are average matchscores,
let only the first 15% play together and then the last 35% and then the middle.
Three dynamic tiers, no tier-crossover and no xp-bar or psr is needed.
At the beginning of each season you start with zero,
but your last average matchscore is used for your first match in the new season.
Simple system, not based on tier or playernumbers.

It's like you're not even trying to understand. There are two possible outcomes from matchmaking systems when the player population is as low as MWO's: You allow increasingly loose matchmaking to get people into games (thus creating skill imbalances in matches), or you maintain hard-locked matchmaking tied to whatever metrics you please (thus forcing everyone to wait for far longer to find matches).

I think most of us would rather have variable quality in our matches than have to sit around for ten or fifteen minutes just to get into one. I've played games like that in the past; I no longer play them. That should speak for itself.

Edited by Alexander of Macedon, 07 August 2017 - 01:41 PM.


#54 Albino Boo

    Member

  • PipPipPipPipPipPip
  • Shredder
  • Shredder
  • 281 posts

Posted 07 August 2017 - 03:04 PM

View PostOrmsbyGore, on 07 August 2017 - 01:07 PM, said:


No, the good drivers go left and support their assaults, while preparing to meet the enemy's 8 fastest mechs with 12 mechs in defensive positions


No good players do not go left to support 100 ton 35kph lrms boats that stand still. No amount of support can solve your poor choices in tactics and builds. It's sure sign of potatoedom complaing that you left the assaults behind. If you stand still in quick play you lose because there are no choke points that cannot be moved around. If you haven't pressed the w key and you are getting shot in the back by a light then it's your problem. Passivity it's not a tactic.

Edited by Albino Boo, 07 August 2017 - 03:55 PM.


#55 OrmsbyGore

    Member

  • PipPipPipPipPipPip
  • 200 posts

Posted 07 August 2017 - 04:44 PM

View PostAlbino Boo, on 07 August 2017 - 03:04 PM, said:

No good players do not go left to support 100 ton 35kph lrms boats that stand still. No amount of support can solve your poor choices in tactics and builds. It's sure sign of potatoedom complaing that you left the assaults behind. If you stand still in quick play you lose because there are no choke points that cannot be moved around. If you haven't pressed the w key and you are getting shot in the back by a light then it's your problem. Passivity it's not a tactic.


Yes, but a well built 100 ton assault brawler that just barely hits 60 kph can also get left behind by lights, mediums and fast heavies, even if they advance to the front right away. In fact, on some maps where charlie lance spawns on the team's left, nascaring as hard as you can to the right isn't really a great strategy: either you mirror the other team and derive no advantage, or the other team goes left and the 8 fastest mechs on your team run into the entire enemy team, including those well armed and armored assaults, and you are at a disadvantage. Additionally, since this is a TEAM deathmatch game, losing your assaults early (or being outmanned and outgunned quickly) is, in fact, your problem.

While I agree strongly that passivity is not a tactic, running headlong into the enemy in a counter-clockwise circle in every match is a silly tactic, as it is generally undesirable to allow the enemy to successfully predict your actions

#56 Alexander of Macedon

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,184 posts

Posted 07 August 2017 - 05:09 PM

Mining Collective is a great example of that at work. If one team all moves to push the enemy assault lance and the other doesn't, the match is decided within the first 3 minutes. This is most clearly observed during Conquest matches where the cap lures stupid assault jocks to linger until seven or eight 'mechs scythe in and pin them to the wall.

#57 Albino Boo

    Member

  • PipPipPipPipPipPip
  • Shredder
  • Shredder
  • 281 posts

Posted 08 August 2017 - 01:12 AM

View PostOrmsbyGore, on 07 August 2017 - 04:44 PM, said:


Yes, but a well built 100 ton assault brawler that just barely hits 60 kph can also get left behind by lights, mediums and fast heavies, even if they advance to the front right away. In fact, on some maps where charlie lance spawns on the team's left, nascaring as hard as you can to the right isn't really a great strategy: either you mirror the other team and derive no advantage, or the other team goes left and the 8 fastest mechs on your team run into the entire enemy team, including those well armed and armored assaults, and you are at a disadvantage. Additionally, since this is a TEAM deathmatch game, losing your assaults early (or being outmanned and outgunned quickly) is, in fact, your problem.

While I agree strongly that passivity is not a tactic, running headlong into the enemy in a counter-clockwise circle in every match is a silly tactic, as it is generally undesirable to allow the enemy to successfully predict your actions

A well built 65kph brawler assault is an oxymoron in quickplay. 11 other people are not going to give away the speed of their build choices becasue of your poor choice of a slow mech. The Kodiak the Marauder IIc and the Madcat II can be configured to brawl at much faster speeds. Its your choice to enter the quick play queue with a slow mech and you have to take the consequences. 11 other people dont not have change their play style because of what you want to do. A 65 kph assault is wastage of tonnage when you have the possibility of playing conquest, domination and escort.

Edited by Albino Boo, 08 August 2017 - 01:13 AM.


#58 Wil McCullough

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,482 posts

Posted 08 August 2017 - 01:50 AM

what most thick-headed players seem to ignore is that "gitting gud" is often as simple as listening to the damn calls or calling enemy targets out themselves.

communication trumps good aim every time.

the ability to listen and actually TALK is what separates the new players from the potatoes.

#59 Racerxintegra2k

    Member

  • PipPipPipPipPipPipPip
  • The God
  • The God
  • 801 posts

Posted 08 August 2017 - 02:05 AM

View PostBrain Cancer, on 07 August 2017 - 11:52 AM, said:

No, because PGI deliberately stuffed T2,even T1 with potatoes. In fact, the PSR system "rewards" lots of play with higher tiers due to it's massive positive bias- it's an exp bar after all.

The first thing that needs to happen is PSR having harsher requirements to gain/lose/maintain as tier increases. If tiering is to matter in the matchmaker, it has to actually represent improved skills, not just "I have lots of time to play, randomly shoot people, and hope I get carried again."

The second thing is reducing games to the players available -before- opening the valves to let the spuds drop. Got 24 players of the right tier? Great, 12v12. Not enough? Try an 8v8 or even 4v4. Only after that fails should lower tiers be injected in. Besides, smaller group sizes also act to change match dynamics, meaning more variety in same.

The third thing is expanding the match data available to players. We need to see all the relevant scoring for players, not just W/L + K/D + match score. KMDD, is he an assist machine with lots of damage or a paint-scratcher farming score by max range ERLL fire? This sort of thing can be better understood,and fed back into the scoring system to better reflect what efforts really do reflect "git gud".



You Sir win. Best solution ive ever heard of.

#60 Zergling

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Angel
  • The Angel
  • 2,439 posts

Posted 08 August 2017 - 02:38 AM

View PostAlbino Boo, on 08 August 2017 - 01:12 AM, said:

A well built 65kph brawler assault is an oxymoron in quickplay. 11 other people are not going to give away the speed of their build choices becasue of your poor choice of a slow mech. The Kodiak the Marauder IIc and the Madcat II can be configured to brawl at much faster speeds. Its your choice to enter the quick play queue with a slow mech and you have to take the consequences. 11 other people dont not have change their play style because of what you want to do. A 65 kph assault is wastage of tonnage when you have the possibility of playing conquest, domination and escort.


64.8 kph is the top speed of a Kodiak without MASC (which only the Spirit Bear has) or speed tweak nodes.

Aside from that, both the Marauder IIC's and Mad Cat Mk II's best brawler builds only hit 64.8 kph too; and they are frequently built to be a bit slower than that.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users