Jump to content

Can Pgi Finally Scrap The Matchmaker?


65 replies to this topic

#41 Mackman

    Member

  • PipPipPipPipPipPipPip
  • 746 posts
  • LocationCalifornia

Posted 16 August 2013 - 10:04 PM

View PostIceLom, on 16 August 2013 - 02:22 PM, said:


You just proved why he is correct...

Because in that context of what players are available, the most balanced team is the strong + weak vs the 2x middle. Sure the one team might still be allot stronger but its the most fair possible combination that can be created given the available players.


Excellent point, and one that needs to be remembered.

#42 Mangonel TwoSix

    Member

  • PipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 238 posts

Posted 17 August 2013 - 08:36 AM

View PostDeadlyNerd, on 16 August 2013 - 01:02 AM, said:

It doesn't. I've had countless steamrolls in DoTA.

Also 1v1 can always mean premade team vs premade team, but that premade team is a single entity and players in it aren't counted individually.


Getting steamrolled as you put it does not prove a disparity in ELO or skill. If you think about it, games ending up with lopsided wins reinforces that it IS working.

Something happens to give a team an advantage. Player disconnects, someone gets a quick headshot kill, takes a few seconds to scarf down a hot pocket, your mom comes down to the basement to make you take a bath, whatever.

One side has a force multiplier. They are able to utilize that to get up on kills, the bigger their advantage the further they get up. The easier it is to maintain that advantage. The harder it becomes for the team that is down to come back.

Basically go play tug of war, and give one team a running start. The other has to stand still. That is what is going on.

#43 DeadlyNerd

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,452 posts

Posted 17 August 2013 - 10:18 AM

View PostRoknari, on 17 August 2013 - 08:36 AM, said:


Getting steamrolled as you put it does not prove a disparity in ELO or skill. If you think about it, games ending up with lopsided wins reinforces that it IS working.

Something happens to give a team an advantage. Player disconnects, someone gets a quick headshot kill, takes a few seconds to scarf down a hot pocket, your mom comes down to the basement to make you take a bath, whatever.

One side has a force multiplier. They are able to utilize that to get up on kills, the bigger their advantage the further they get up. The easier it is to maintain that advantage. The harder it becomes for the team that is down to come back.

Basically go play tug of war, and give one team a running start. The other has to stand still. That is what is going on.

This is not a tug of war and you, like many others, are again misunderstanding how Elo works and why it's used.
In tug of war each team can be seen as a single entity and players within each team can have their strength combined into 1 value. The whole system is actually 1v1 in the end.
This is NOT. This is a game where each player plays on his own and for himself. He doesn't have a symbiotic relationship with the team. Teams, however premade they were, are never going to be counted as a single entity until their members become constant and don't change.
It's a simple fact, yet nobody seems to understand it and just keeps reinforcing Einstein's definition of insanity.

Reason a matchmaking system is being used is to create similar matches and fair fights. The fact that it doesn't do that is exactly the reason why it's NOT working properly. I can't fathom why someone would think a matchmaker works if it's not making even matches.

View PostMackman, on 16 August 2013 - 10:04 PM, said:


Excellent point, and one that needs to be remembered.


Actually, that's a point that only A I D S (it's a ******* verb, fix your ******* filter) my argument. There shouldn't be a team where 1 player greatly deviates from the rest of the team.
Current matchmaker is doing just that, and it obviously isn't working.

View PostMackman, on 16 August 2013 - 10:01 PM, said:


But it's extremely unlikely (mathmetically speaking) that you winning despite your low skill would happen more often than you losing because of it. It's simple probability: Given a sufficiently large number of matches, you will eventually be placed with people of the same skill (or "effectiveness", if you bring in people who are intentionally gimping themselves).

Elo, if properly implemented, will work: That's why League of Legends has been using it successfully ever since it started out (you know, before it became the biggest game in the world by using a good Elo matchmaking system.).


He get's carried the first match and then loses a match for those of higher rating. He caused their rating to get ruined, it wasn't their fault. How do you explain and justify that? He wont be placed with people for the same rating because games are already in motion and the matchmaker already has to loosen the criteria to get more players. He'll ALWAYS be put into a team with better or worse players.

Elo doesn't "work", it's the matchmaker that "works". Elo is made for 1v1 and it can't be properly implemented. I understand ignorant, but ignorant and dense takes the cake.
LoL using it efficiently? Haha, that's a laugh. Tell that to my frustrated brother who gets beginners in high ranked games just cause the matchmaker decided to "fill out the blanks".

An Elo based matchmaker works only in WoW, because those arena teams are counted as a single entity. An Elo based matchmaker would even work for football teams because they're a single entity, but it WOULDN'T work if those teams were comprised of random players that never played along side each other., the very reason why national teams can suck even if they're filled with premier league players.

View PostHauser, on 16 August 2013 - 02:42 PM, said:


Yes. Like I said before, I agree. People with significant Elo difference should not be on the same team. However that was not the point the OP was making.

That IS the point I am making. Thank you for being ignorant throughout the thread.

Edited by DeadlyNerd, 17 August 2013 - 10:39 AM.


#44 I am

    Member

  • PipPipPipPipPipPipPip
  • 542 posts

Posted 17 August 2013 - 11:52 AM

People with significant ELO differences should be on the same team. A good player can pick up the slack for a bad player. Here is a match I played today. I wish it had been a win, but oh well. As you can see, I got some so so mates. It started out 0-6, with mates cresting, dying, and joking about it in all chat. It happens. But imagine if I had been in a 4 man with the three best players on the other team, wouldn't that have been unfair? You have to mix the good and bad players. Sorry PGI cares more about everyone having a chance, than they do about your personal enjoyment at the expense of others. I also feel bad for anyone who would prefer their increasing personal enjoyment, at the expense of everyone having a fair chance. Everyone having a fair chance, makes this game more marketable.

Posted Image

#45 DeadlyNerd

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,452 posts

Posted 17 August 2013 - 01:13 PM

View PostI am, on 17 August 2013 - 11:52 AM, said:

People with significant ELO differences should be on the same team. A good player can pick up the slack for a bad player. Here is a match I played today. I wish it had been a win, but oh well. As you can see, I got some so so mates. It started out 0-6, with mates cresting, dying, and joking about it in all chat. It happens. But imagine if I had been in a 4 man with the three best players on the other team, wouldn't that have been unfair? You have to mix the good and bad players. Sorry PGI cares more about everyone having a chance, than they do about your personal enjoyment at the expense of others. I also feel bad for anyone who would prefer their increasing personal enjoyment, at the expense of everyone having a fair chance. Everyone having a fair chance, makes this game more marketable.



So basically PGI is giving everyone a fair chance by not giving them a fair chance with the faulty matchmaker. Fascinating.

Also, you can't really prove anything with a random game you were obviously waiting to do a screenshot on, because otherwise you wouldn't be posting it here.

And again, by increasing the number of players PGI increased the chances of good players actually entering in with the bad players and sort of compensating. This is a false sense of security that the system "works".
I already said, give us 32v32 and we'll all be happy as then rating wont matter at all. You don't call a matchmaker functional just because team sizes overrode some of its matchmaking criteria and somehow made the thing function. You don't MacGyver in programming and PGI is failing to see that in many aspects, not just the matchmaker, and it's backfiring.

My point is that without a matchmaking system we'd have instant matches and nobody could blame their losses on not being matched properly because everything would be random.

Edited by DeadlyNerd, 17 August 2013 - 01:23 PM.


#46 I am

    Member

  • PipPipPipPipPipPipPip
  • 542 posts

Posted 17 August 2013 - 01:49 PM

No it wouldn't be random, because collections of skilled players group up, have good gear, and verse players who are solo with sub-par builds. That's what you want, your pug farming back. Your not going to get it, so just stop.

#47 Hekalite

    Member

  • PipPipPipPipPipPip
  • 424 posts

Posted 17 August 2013 - 02:01 PM

Bad matches are not outliers, they are pretty damn common right now. I don't think the system should be scrapped, but it needs a heavy tuning pass.

#48 DeadlyNerd

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,452 posts

Posted 17 August 2013 - 02:20 PM

View PostHekalite, on 17 August 2013 - 02:01 PM, said:

Bad matches are not outliers, they are pretty damn common right now. I don't think the system should be scrapped, but it needs a heavy tuning pass.

I'd very much like to see someone actually make a good matchmaking system with Elo rating as base, but it's simply not possible because Elo is not designed for more than 1v1.
If they really wanted to balance out matches then they'd have to take into account wins, losses, k/d with the selected mech, damage per match with the selected mech etc. That, however, is a shitton of work to sort out and create an algorithm to process it.

View PostI am, on 17 August 2013 - 01:49 PM, said:

No it wouldn't be random, because collections of skilled players group up, have good gear, and verse players who are solo with sub-par builds. That's what you want, your pug farming back. Your not going to get it, so just stop.

Pit 4 man teams only against 4man teams and you have better matchmaking than what's currently in.
Btw, you do understand that random means any situation where a pattern is not developed? ...and that your argument is meaningless.
You're starting your own crusade against me personally here so stop it because it's not contributing to the topic at hand and is completely irrelevant whatsoever.

Edited by DeadlyNerd, 17 August 2013 - 02:24 PM.


#49 IceSerpent

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,044 posts

Posted 18 August 2013 - 09:23 AM

View PostI am, on 17 August 2013 - 11:52 AM, said:

People with significant ELO differences should be on the same team. A good player can pick up the slack for a bad player. Here is a match I played today. I wish it had been a win, but oh well. As you can see, I got some so so mates. It started out 0-6, with mates cresting, dying, and joking about it in all chat. It happens. But imagine if I had been in a 4 man with the three best players on the other team, wouldn't that have been unfair? You have to mix the good and bad players.


I don't follow your reasoning, especially the part about 4-man on the other team. Granted, a good player can pick up a slack, but it's the same as dropping 11 v 12. A bigger question is why do you think that player with vastly different Elo ratings should be on the same team instead of being in completely separate matches? From my perspective forcing a good player to essentially drop 1 man down (or more than one in a lot of cases) just because their Elo is high is definitely unfair.

#50 I am

    Member

  • PipPipPipPipPipPipPip
  • 542 posts

Posted 18 August 2013 - 09:38 AM

View PostIceSerpent, on 18 August 2013 - 09:23 AM, said:


I don't follow your reasoning, especially the part about 4-man on the other team. Granted, a good player can pick up a slack, but it's the same as dropping 11 v 12. A bigger question is why do you think that player with vastly different Elo ratings should be on the same team instead of being in completely separate matches? From my perspective forcing a good player to essentially drop 1 man down (or more than one in a lot of cases) just because their Elo is high is definitely unfair.


What I meant was, if I hadn't been solo on team L (losers), but instead had been in a 4 man on team W (winners), with their top three players. The match would have played out quite differently. Instead of it being the 8-12 finish it was. IF we had 30k players, and you could support herding the low elos together, and high elos together, maybe it could be considered. I doubt it though. Its a lot easier just to use the whole body of players.

You say that dropping one man down is unfair? IF your 5x better than anyone else on the opposite team, you'll hopefully, if the player is available, get a mate who is 5x worse than everyone on their team. Maybe him being 5x worse means he AFKs 4 out of 5 games. Example, good player is 5. Bad player is 1. Avg. player is 3. 5+1 = 3+3. I really don't understand why that is so hard to understand. Yes you get bad mates sometimes, but that is because your above average in terms of skill, and win more than they do as a result. What you, and many of the posters here want, is 5+5+5+5 vs 1+2+1+3, followed by L33T high fives on TS for being so dominant. THAT is unfair. Hope that clears it up for you.

#51 IceSerpent

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,044 posts

Posted 18 August 2013 - 10:36 AM

View PostI am, on 18 August 2013 - 09:38 AM, said:

What I meant was, if I hadn't been solo on team L (losers), but instead had been in a 4 man on team W (winners), with their top three players. The match would have played out quite differently. Instead of it being the 8-12 finish it was.


I hear you, but that's not what we are talking about - the idea is to put folks with different ratings into different matches, not shufle them into different teams within the same match.

Quote

IF we had 30k players, and you could support herding the low elos together, and high elos together, maybe it could be considered. I doubt it though. Its a lot easier just to use the whole body of players.


Yeah, it's certainly easier. One problem though - it completely defeats the whole purpose of having Elo to begin with. If you want to go down that road, you might just drop Elo matching completely and make it even easier.

Quote

You say that dropping one man down is unfair? IF your 5x better than anyone else on the opposite team, you'll hopefully, if the player is available, get a mate who is 5x worse than everyone on their team. Maybe him being 5x worse means he AFKs 4 out of 5 games. Example, good player is 5. Bad player is 1. Avg. player is 3. 5+1 = 3+3. I really don't understand why that is so hard to understand.


This is where you went wrong - in MWO 5+1 is not the same as 3+3, it's actually much, much worse due to the nature of the game. You can't add player skill like this in games where you have to rely on your teammates to do certain things in order for you to succeed. This works in something like shooting clay pigeons where you simply add up everyone's score, but doesn't work in games like football or basketball - a team full of average players would almost always win against a few exceptional ones trying to carry a bunch of rookies.

#52 I am

    Member

  • PipPipPipPipPipPipPip
  • 542 posts

Posted 18 August 2013 - 10:53 AM

You don't get it because you have an agenda behind your argument, reinstituting a non-elo based mm system which perpetuates pug farming. Not gonna happen. PGI won't push away droves of would be customers by creating a meta where you can farm them indefinitely. They did, for more than half a year. It's sad, because we might have a much larger community by now hadn't they. But it finally dawned on them, that they were killing their own player counter, and now we have ELO and 12v12 to lessen the impact of a skilled 4 man in a pub match.

No matter how you spin the issue, that's what it boils down to, and farming isn't coming back. If the bad player on your team, is so bad you cannot carry him to an even match result 8-12 9-12-ish, then your own elo is temporarily overstated. If you don't like that, and always want the odds in your favor, I'd suggest a single player game. If you want to play mp, do so seeking a challenge (a "good fight"), not simply a steamroll.

#53 Alistair Winter

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • Storm
  • Storm
  • 10,823 posts
  • LocationBergen, Norway, FRR

Posted 18 August 2013 - 11:33 AM

The reason ELO isn't working is because there are too few players. The servers need to find players within a certain group, ideally with certain weight classes, looking for a certain game mode, either solo or belonging to a premade group of certain skill and - in the long term - from a certain faction. So while the system works in an ideal world, the end result is not good right now.

I would prefer it if there were only 3 different available levels, call it beginner, intermediate and advanced. Beginners can choose any level, but as you progress and play x number of games, you can no longer play beginner level, and then no longer intermediate level. New players can choose to play in intermediate and advanced if they want, but they are not likely to do so, unless they want to play with their friends who are already at a higher level. There's no incentive of choosing a higher level, no higher rewards, so there's no reason for people to suicide farm advanced level or anything like that.

Failing that, at least keep the ELO of every player very visible during a match. This solves 2 things.
1) It takes some of the pain away when you get headshotted by a jump jetting spider from 2 km away, if you can look at the guy's ranking and realize that he's probably one of the best pilots in the game. A lot of players are frustrated with this gaming, asking "why do I suck?" when they get stomped by more experienced players. That would stop if they realized who they were up against and that it's not their fault.
2) It takes some of the pain away when you can see by the difference in ELO between teams, to see that it's not your fault, but your team was just outmatched. If the average ELO on your team is 500, and it's 1500 on the enemy team, you don't end up cutting yourself later that night for being such a bad pilot.

Also, I don't like the way ELO is calculated. I think a system based on kills and resources capped is better than matches won. Not everyone plays to win. I'm in this game to collect heads. Winning is just a bonus.

#54 IceSerpent

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,044 posts

Posted 18 August 2013 - 12:03 PM

View PostI am, on 18 August 2013 - 10:53 AM, said:

You don't get it because you have an agenda behind your argument, reinstituting a non-elo based mm system which perpetuates pug farming. Not gonna happen. PGI won't push away droves of would be customers by creating a meta where you can farm them indefinitely.


1/10 on the troll scale. Everybody on these forums is well aware that there's absolutely nothing to farm in MWO, therefore "pug farming" is a complete waste of time that provides no rewards whatsoever.

#55 Jman5

    Member

  • PipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 4,914 posts

Posted 18 August 2013 - 12:39 PM

View PostIceSerpent, on 18 August 2013 - 12:03 PM, said:


1/10 on the troll scale. Everybody on these forums is well aware that there's absolutely nothing to farm in MWO, therefore "pug farming" is a complete waste of time that provides no rewards whatsoever.

Call it Ego farming then. A lot of people get their kicks out of being a big fish in a small pond. beating up on noobs without really trying. Eventually they even start to convince themselves that they are highly skilled because they can consistently get 5 kills and 800 damage against day-1 beginners.

#56 IceSerpent

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,044 posts

Posted 18 August 2013 - 01:08 PM

View PostJman5, on 18 August 2013 - 12:39 PM, said:

Call it Ego farming then. A lot of people get their kicks out of being a big fish in a small pond. beating up on noobs without really trying. Eventually they even start to convince themselves that they are highly skilled because they can consistently get 5 kills and 800 damage against day-1 beginners.


Ok, I can buy that. Question is what in a world does it have to do with the curent topic? Nobody suggested that low Elo teams should be matched against high Elo teams. We are not talking about premades vs. pugs in the same queue either. What we are saying is that players within the same team should always have very similar ratings.

#57 Jman5

    Member

  • PipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 4,914 posts

Posted 18 August 2013 - 01:57 PM

Quote

Ok, I can buy that. Question is what in a world does it have to do with the curent topic? Nobody suggested that low Elo teams should be matched against high Elo teams. We are not talking about premades vs. pugs in the same queue either.

OP wants to scrap the entire skill-based matchmaking in lieu of a simple tonnage system.
Quote

Quote

What we are saying is that players within the same team should always have very similar ratings.


That's exactly how it works currently. The matchmaker tries its best to match you with and against players as close to your Elo as possible. The longer it takes the looser it gets until it reaches a threshold and "fails to find match." It doesn't just load you up with scrubs because you're an ace. I don't match with beginners because my Elo is simply too high. The system will fail to find a match long before I play with them.

Quote

How does the match maker compose a teams Elo rating, is it average rating or closest to a target?

It's closest to a target value, so the match maker starts trying to make a match for an Elo of say 1300 and will pull in players to those teams closest to those values; however, as mentioned earlier within growing thresholds and those curves will be tuned.


#58 IceSerpent

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,044 posts

Posted 18 August 2013 - 02:20 PM

View PostJman5, on 18 August 2013 - 01:57 PM, said:

That's exactly how it works currently. The matchmaker tries its best to match you with and against players as close to your Elo as possible. The longer it takes the looser it gets until it reaches a threshold and "fails to find match." It doesn't just load you up with scrubs because you're an ace.


No, it's not how it works - "the longer it takes the looser it gets" part is a deal breaker. It should never, no matter what, with no exceptions whatsoever put players with a huge Elo difference on the same team.
Whether this "the longer it takes the looser it gets" scheme should apply to matching between already composed teams is open for a debate. Personally, I'd rather not have it, but YMMV.

Quote

I don't match with beginners because my Elo is simply too high. The system will fail to find a match long before I play with them.


Yes, you do. We end up in the same matches quite often and there are plenty of rookies there with us. You just don't pay close attention to what they say and do during the match.
Here's a hint: when you have a team mate in a trial mech asking "how do I cap the base?", that's one of 'em new players... :P

#59 Jman5

    Member

  • PipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 4,914 posts

Posted 18 August 2013 - 02:39 PM

Quote

It should never, no matter what, with no exceptions whatsoever put players with a huge Elo difference on the same team.


It doesn't, that's my point. It loosens things because there is not enough players in the queue, but within reason. If you make it too tight, you will spend more time in queue than playing the game. How are you so certain that you're matching with people with huge Elo differences when those numbers are hidden? Are you just assuming based on bad games, because I've got loads of those.

I mean what would be the point of the developers creating an Elo system if they just wound up throwing experienced vets in with noobs? It doesn't really add up. If your Elo is high for the chassis you were playing when you saw someone in a trial mech claiming to be new, then odds are he was probably joking. Either way it doesn't matter. People who want to believe the matchmaker doesn't work wont listen no matter what anyone says. It's the same complaint in every game with an Elo.

#60 IceSerpent

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,044 posts

Posted 18 August 2013 - 02:55 PM

View PostJman5, on 18 August 2013 - 02:39 PM, said:

It doesn't, that's my point. It loosens things because there is not enough players in the queue, but within reason.


Not within reason, apparently it goes all the way to newbie Elo (1100 iirc).

Quote

If you make it too tight, you will spend more time in queue than playing the game.


I wouldn't mind. One good match per hour is better than 5 lousy ones, as far as I am concerned. Besides, I think queue time is affected by Elo matching between teams more than by Elo matching within a team. In other words, finding 12 players with the same Elo for one team is easier than finding 24 for two teams. Which is why I am not entirely sure whether matching between teams should be strict or not.

Quote

How are you so certain that you're matching with people with huge Elo differences when those numbers are hidden? Are you just assuming based on bad games, because I've got loads of those.


No, not assuming based on outcome. Just observing who is in the match, what they say in chat, and how they act. It's usually very easy to spot a complete newbie and most high Elo players have been around for a while, so I just recognize the names. Not to mention that those players are also quite easy to spot by how they act in game even if I don't recognize the name.

Quote

I mean what would be the point of the developers creating an Elo system if they just wound up throwing experienced vets in with noobs? It doesn't really add up.


They did a lot of things that "don't really add up", so there is no point. It's just PGI being PGI. What was the point of preventing 5-7 player groups from dropping in any queue? (rhetorical question)





6 user(s) are reading this topic

0 members, 6 guests, 0 anonymous users