Jump to content

Elo...writing On The Wall


148 replies to this topic

#61 DaisuSaikoro Nagasawa

    Member

  • PipPipPipPipPipPipPip
  • Mercenary
  • Mercenary
  • 973 posts
  • Google+: Link
  • LocationTaipei, Taiwan

Posted 13 March 2013 - 10:56 AM

View Postjay35, on 13 March 2013 - 10:15 AM, said:

Question: If someone frequently runs into well-known folks here like V. Rain, Wispsy, and Broceratops in their PUG matches, is that a sign of a high ELO?



No, Yes, Yes.

Edited by Daisu Saikoro, 13 March 2013 - 10:57 AM.


#62 Aethon

    Member

  • PipPipPipPipPipPipPipPipPip
  • Mercenary
  • 2,037 posts
  • LocationSt. Louis, Niles, Kerensky Cluster

Posted 13 March 2013 - 10:58 AM

View PostYokaiko, on 13 March 2013 - 10:46 AM, said:



That is one of my main gripes with ELO. You are at the mercy of your team, I've lost games where there were two guys with over 1000 damage (not LRM boats) but FIVE under 100.

Four or five times a day I end up with a team where I'm top damage and account for 2nd through 5th to come up with an equal number.

.....and take the elo hit for it.

It actually punishes you, which is ******* annoying.


This sums up our matches pretty well. Whichever one of us is in 4th place typically deals more damage than all four of our pugs combined. I realize damage is just one aspect of a player's performance, but when the guy in 4th place does 250 damage, and the people in 5th-8th place do between 0 and 60 damage, something is wrong.

#63 Yokaiko

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 6,775 posts

Posted 13 March 2013 - 10:58 AM

View PostMegurine Luka, on 13 March 2013 - 10:52 AM, said:

They did say that they were still doing a balance pass on ELO. I certainly hope that it improves the quality of games during off-peak hours. For now, playing my lower ELO weight classes does help during the 'bad' times.



No off times are when you get trails in with everyone else. I play like 8am eastern US, I see a fair about of it. On the few times that I play US prime (maybe once every other week) I see a bunch of very good players.

#64 Tickdoff Tank

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 2,647 posts
  • LocationCharlotte NC

Posted 13 March 2013 - 11:00 AM

IMO, Elo alone will not be sufficient to balance the MM in MWO. We need to have some sort of a "BV" (Battle Value) system which works with the Elo system.

A pilot's effectiveness in a match is not determined purely by skill, His choice of mech and weapons also plays a part. Ignoring either of these aspects will cause problems.

We will not be able to use the base BV system from TT. The weapons and mechs work much differently in MWO for the TT BV system to be accurate. In order to develop an effective BV for MWO you will have to look at more than just the base numbers for a weapon or equipment. As a previous poster has stated, in TT all weapons fired at the same rate, all weapons did their damage instantly, and all weapons hit random locations. That is not how MWO works, so we must develop our own BV metric, one which would give a much higher value to a SSRM2 than an AC2. (In TT the SSRM2 has a BV of 30 and the AC2 has a BV of 37)

#65 Glythe

    Member

  • PipPipPipPipPipPipPipPip
  • 1,566 posts

Posted 13 March 2013 - 11:05 AM

View PostYokaiko, on 13 March 2013 - 10:46 AM, said:



That is one of my main gripes with ELO. You are at the mercy of your team, I've lost games where there were two guys with over 1000 damage (not LRM boats) but FIVE under 100.

Four or five times a day I end up with a team where I'm top damage and account for 2nd through 5th to come up with an equal number.

.....and take the elo hit for it.

It actually punishes you, which is ******* annoying.


There's so many people that apparently are having fantastic match making. And I see posts like this where someone is having a mirror experience to me (and then another and another).

I'm happy some of you are having a nice time with the new MM ( if you are in this crowd you probably either play at peak times or you were getting stomped by pemades before and are now matched properly). But there's a number of us that are getting screwed over here and we can't figure out why.

#66 jay35

    Member

  • PipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 1,597 posts

Posted 13 March 2013 - 11:14 AM

View PostYokaiko, on 13 March 2013 - 10:30 AM, said:


My KDR jumped a full point after ELO......even though my win loss is dropping like a stone.

All that really says is that you're probably an above average player and when ELO launched, everyone started with an average ELO (1500 or whatever), so as you worked your way up to your accurate ELO, you played a lot of games that gave you a high KDR. Now your W:L is dropping as you're leveling out at an appropriate ELO, provided that's actaully how it works. Then again, over time, ELO has definitely gotten worse with matchmaking. I've seen more lopsided (winning 8-0 or losing 0-8 or close to it) and incomplete-team games (6v8, 7v8,etc), in the past few days than ever before.

Then again, I think an unintended consequence of last weekend's contest is that everyone who tried to get in as many matches as quickly as possible to try to win the assists challenges, artificially dropped their ELO ratings severely (because they basically just ran in a Light, tagged all eight enemies, and died as quickly as possible), so a lot of players are much better than their current ELO and it will take time for their rankings to balance back up.

#67 Targetloc

    Member

  • PipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 963 posts

Posted 13 March 2013 - 05:21 PM

I'm experiencing the exact same thing as Glythe.

About 1 game in 4 I'll see a pretty decent team on both sides and maybe a name I recognize, but most rounds end with 2-3 players doing under 100 damage. Most rounds end with the top 2 players doing 600-800 damage and the rest doing 200 or less.

I managed a 70% win rate over 86 games in my heavies during the Be a Hero tournament when all this elo was supposed to be shaking itself out and there were lots of high value targets around, so I assume my heavy elo is at least decent.

But I'm still seeing trial mechs in my games. And LBX10 Altai.

Frankly I think the most telling factor of how a match is going to go is looking around at the start and seeing how many mechs are default green camo.

Though I will applaud PGI because since elo, games have definitely been far more challenging.

Posted Image

#68 Davers

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 9,886 posts
  • Facebook: Link
  • LocationCanada

Posted 13 March 2013 - 05:31 PM

1. Elo does not guarantee good teammates. You are not being put in an Elo 'bracket' and matched with players in that bracket. Just that the team's average Elo will be similar.

2. With so many high alpha strike builds out there, one wrong move can kill even the highest ranked player in one or two alphas. Low damage does not always mean bad team mates.

#69 Bleary

    Member

  • PipPipPipPipPipPip
  • Survivor
  • 365 posts

Posted 13 March 2013 - 07:40 PM

View PostDavers, on 13 March 2013 - 05:31 PM, said:

1. Elo does not guarantee good teammates. You are not being put in an Elo 'bracket' and matched with players in that bracket. Just that the team's average Elo will be similar.

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

A: 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. Currently it may be a bit 'sloppy' about how it's filling those buckets but over time it will be tuned to be much more precise."

This has already been posted once. Bad PuG players are not thrown in to balance out a good 4-man. ELO is not calculated by team average. The system does not work that way.

Though I expect dropping as a 4-man complicates matters, as you're forcing the matchmaker to put you all in the same match regardless of whether your clanmates' ELO scores are similar. It wouldn't surprise me if that's the real problem.

Edited by Bleary, 13 March 2013 - 07:46 PM.


#70 Davers

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 9,886 posts
  • Facebook: Link
  • LocationCanada

Posted 13 March 2013 - 07:50 PM

View PostBleary, on 13 March 2013 - 07:40 PM, said:

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

A: 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. Currently it may be a bit 'sloppy' about how it's filling those buckets but over time it will be tuned to be much more precise."

This has already been posted once. Bad PuG players are not thrown in to balance out a good 4-man. ELO is not calculated by team average. The system does not work that way.

Though I expect dropping as a 4-man complicates matters, as you're forcing the matchmaker to put you all in the same match regardless of whether your clanmates' ELO scores are similar. It wouldn't surprise me if that's the real problem.

But how is the target value found? Is it based on the first guy to hit 'launch'? Does it make a whole team, THEN tries to find the equivalent so the first team might have wildly different Elo's?

#71 Adrian Steel

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 545 posts

Posted 13 March 2013 - 08:12 PM

I can corroborate what the OP is saying.

I often run a 2 man pre-made with my brother, Ryan Steel. We're of comparable skill. We coordinate on comms. We can be matched with bads on our team that do everything wrong, while the other team appears to be chock full of decent players.

When we each solo pug, there never seems to be the same disparity. Every bad on our team correlates to a bad on the other team.

We have a sneaking suspicion that the match making code is different for pre-mades than solo pugs, and that this is most noticeable in non-primetime hours where the population at the extreme ends of the Elo bell curve will be minimal. It's very possible that pre-mades are assumed to have a communication advantage in the match making code, and are purposefully matched with team-mates of a lower Elo in order to compensate.

If you're incredulous at this claim, ask yourself how many modifications to the game haven't been in the patch notes... There are entire threads devoted to discussing these easter eggs.

#72 Targetloc

    Member

  • PipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 963 posts

Posted 13 March 2013 - 08:33 PM

View PostAdrian Steel, on 13 March 2013 - 08:12 PM, said:

I can corroborate what the OP is saying.

I often run a 2 man pre-made with my brother, Ryan Steel. We're of comparable skill. We coordinate on comms. We can be matched with bads on our team that do everything wrong, while the other team appears to be chock full of decent players.

When we each solo pug, there never seems to be the same disparity. Every bad on our team correlates to a bad on the other team.

We have a sneaking suspicion that the match making code is different for pre-mades than solo pugs, and that this is most noticeable in non-primetime hours where the population at the extreme ends of the Elo bell curve will be minimal. It's very possible that pre-mades are assumed to have a communication advantage in the match making code, and are purposefully matched with team-mates of a lower Elo in order to compensate.

If you're incredulous at this claim, ask yourself how many modifications to the game haven't been in the patch notes... There are entire threads devoted to discussing these easter eggs.



I don't think it's a secret conspiracy against groups.


Let's assume you and your brother are both in the 1800's. If it's in the off-peak hours, the game searches and searches for 2 other 1800 players who launched within the last few seconds, but can't find any. So it widens the search bracket to look in the 1700's and 1900's. Then the 1600's and the 2000's.
Now, you are extremely unlikely to find anyone the higher you go. But the lower you go, there's also more people playing. So even though there might be some people queuing up in the 1600's, the game sees 2 or 3 people in the 1600's launched around the same time, and they're a much closer match, so they get paired with each other instead of you.

Now it's getting close to the 2 minute limit so the MM goes into panic mode and slaps you two in against two guys in the 1400's.

So now the MM needs to fill in the rest of the players. And it needs to keep the average elo the same. You and your brother add to 3600. Your opponents add to 2400.

Oh look, there's 2 guys in the 1100's that have been waiting 2 minutes for a game too because there's no one else down in the underhive to match them with. You get those guys.

Now your total is 5800. The other guys get 2 dudes in the 1300's. Now they're at 5400.

Some dude in the 900's accidentally launched while chewing on his Logitech G27. He's on your team.

The average joes pick up another 1400 player.

Now this game is already in the 'taking too long, no time for brackets' mode so the rest of both teams is filled with 1300 peoples who are pretty much unranked and random (with a huge population skew towards new players). GLHF.


Oh, and remember it works in the other direction too. If some dude in the 900's has been waiting 2 minutes for it to pair him up with someone, the game will pull you down to help average his score against the 1300's it needs to pair him with.


Posted Image

#73 Sean von Steinike

    Member

  • PipPipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 2,880 posts
  • LocationUSA

Posted 13 March 2013 - 08:41 PM

View PostFlyby215, on 13 March 2013 - 08:10 AM, said:

Perhaps your friend's problem is with his expectations? Being a multiplayer game, you can't win every match because someone always has to lose. If elo is doing it's job, you should have about a 50/50 split. For former pugstompers this is going to be hard to accept, for PUGers this is a massive improvement.

On a personal experience note, I've already put up a post under feedback regarding matches with my brother, who was brand new at the game. When I grouped with him, for the first time ever since elo, I had several people still learning the controls of the game on my team, but none on the opposing team; and we got owned, repeatedly. It felt as though my teammates were on par with my brother's low elo, and our opponents were of my usual elo bracket. As soon as I went back to solo play, skilled teammates returned. When he returned to solo play, newbie enemies returned as well. Wondering if maybe this is the same problem you/your friend is encountering?

I've had similar experiences. When I am alone, the ELO match up works pretty well. When I play with my friend, who is much better than I, it still seems OK. But when we add a 3rd friend who is somewhere between the two of us in skill level, we get destroyed constantly. I even joke about it to the other about how they must really stink to be matched against me because I am at best, average. But in truth, the ELO match maker seems wonky when the three of us team instead of just two.

#74 Adrian Steel

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 545 posts

Posted 13 March 2013 - 08:41 PM

View PostTargetloc, on 13 March 2013 - 08:33 PM, said:

Excellent reply


Indeed, this could be what's happening. As Elo scores are hidden, it's difficult to affirm anything. But your theory fact based argument is likely more relevant than mine.

#75 Adrian Steel

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 545 posts

Posted 13 March 2013 - 08:47 PM

View PostTargetloc, on 13 March 2013 - 08:33 PM, said:

Some dude in the 900's accidentally launched while chewing on his Logitech G27. He's on your team.


:P

#76 Deathlike

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 29,240 posts
  • Location#NOToTaterBalance #BadBalanceOverlordIsBad

Posted 13 March 2013 - 08:51 PM

View PostAdrian Steel, on 13 March 2013 - 08:12 PM, said:

I can corroborate what the OP is saying.

I often run a 2 man pre-made with my brother, Ryan Steel. We're of comparable skill. We coordinate on comms. We can be matched with bads on our team that do everything wrong, while the other team appears to be chock full of decent players.

When we each solo pug, there never seems to be the same disparity. Every bad on our team correlates to a bad on the other team.

We have a sneaking suspicion that the match making code is different for pre-mades than solo pugs, and that this is most noticeable in non-primetime hours where the population at the extreme ends of the Elo bell curve will be minimal. It's very possible that pre-mades are assumed to have a communication advantage in the match making code, and are purposefully matched with team-mates of a lower Elo in order to compensate.

If you're incredulous at this claim, ask yourself how many modifications to the game haven't been in the patch notes... There are entire threads devoted to discussing these easter eggs.


This actually would make sense, in a really sad way. One would hope it's not the rest of the PUGs that are that much lower than the premade's average ELO...


View Postjay35, on 13 March 2013 - 10:15 AM, said:

Question: If someone frequently runs into well-known folks here like V. Rain, Wispsy, and Broceratops in their PUG matches, is that a sign of a high ELO?


Seen them all... I think it may be in error. /shrug

#77 Hedonism Robot

    Member

  • PipPipPipPipPipPip
  • 421 posts
  • LocationSpace Pirate

Posted 13 March 2013 - 09:29 PM

View PostZnail, on 13 March 2013 - 08:40 AM, said:

That is correct, it got be a large game DESPITE using a flawed Elo system. They also changed the system recently into a better one. The old LoL system had the same problems the current MWO have and that is why they changed it.


This is fales, League of Legends still uses the ELO system expect they run it in the background now and made a slower moving ladder progression visible to users. This was entirely for the psychological effect of players getting too worked up about increasing their ELO number that they could read.

Also if you actually played League of L:egends after playing MOBAs that didn't incorporate ELO match making then you could easily see the reason LoL is popular is thanks to ELO. As stated earlier if you just want to stomp easy opponents play a game with bots and god mode. People are more likely to play a game long term if they win half their games.

#78 CheezPanther

    Member

  • PipPipPipPipPip
  • 125 posts

Posted 13 March 2013 - 09:34 PM

I wanna wipe my ELO every day so i dont have to fight f4ggot 3L's anymore..

#79 Hedonism Robot

    Member

  • PipPipPipPipPipPip
  • 421 posts
  • LocationSpace Pirate

Posted 13 March 2013 - 09:35 PM

View PostCheezPanther, on 13 March 2013 - 09:34 PM, said:

I wanna wipe my ELO every day so i dont have to fight f4ggot 3L's anymore..


you will still have to fight them, but they will be the ones with 200 std engines and without double heatsinks :(

#80 OneEyed Jack

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,500 posts

Posted 13 March 2013 - 10:45 PM

View PostZeh, on 13 March 2013 - 08:13 AM, said:

His score will go down from losing. Don't know wtf you're talking about BillyM.

Since you have to lose a match that was either ranked fairly close, or in which you were favored for your ranking to go down, and since people that always lose tend to quit, it actually gets rather difficult to continue dropping an Elo score by sand-bagging past a certain point.

View PostSears, on 13 March 2013 - 08:21 AM, said:

Also your friend should realise that he has four Elo scores, one for each weight class. Suiciding in Lighs and mediums will do nothing to his Assault Elo. Just sayin'

I could be wrong, but I was under the assumption that they hadn't started the class-based ranking, yet. Of course, it could explain a lot about some inconsistencies in my pairings....

View Postjay35, on 13 March 2013 - 09:28 AM, said:

The number one factor that impacts which team will win, is teamwork. And ELO cannot account for, or balance for, teamwork. That is why ELO is essentially worthless beyond dropweight matching.

Actually, it can. Elo is not a measure of skill, but of general success. It doesn't care why you won or lost, only that you did. Generally, a given player will be more successful with a group than without. His Elo will reflect this and he and his team will tend to be dropped against players with a similar success rate, whether they be grouped or skilled soloists. If he chooses to play solo, he will have a tough time, as the people he's matched with will either be grouped or individually better than he. But the beauty is that his Elo will only go that high if he pretty much always plays that way. If he only groups occasionally it won't go up much, and if he splits his time, it will be somewhere in the middle.

What the score would be if he played in a different way doesn't matter. It based on the way he does play.





18 user(s) are reading this topic

0 members, 18 guests, 0 anonymous users