Jump to content

Elo Rating Will Affect Team Players More Than Solo.


118 replies to this topic

#101 Havyek

    Member

  • PipPipPipPipPipPipPipPip
  • Mercenary Rank 3
  • 1,349 posts
  • LocationBarrie, ON

Posted 12 February 2013 - 08:32 AM

Posted Image

#102 MischiefSC

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • The Benefactor
  • The Benefactor
  • 16,697 posts

Posted 12 February 2013 - 08:42 AM

There are times I'm tempted to thwap myself on the forehead reading these threads.

Elo is very simple.

It doesn't matter how you got your Elo score. The whole point of your Elo score is that you have approximately 50/50 odds of beating another team with the same score. They could be pugs with a lot of natural talent. They could be a team of average players who work well together. The result is that they play at a certain level. The premade team has absolutely and in no conceivable way in any form or fashion better odds of winning against the pugs with the same Elo score. If they did, their score would be higher and they would play against whoever they have a 50/50 odds of winning against.

Please. Understand this. It's not that complicated. Truly it's not. You will either win or lose more than 50% of your matches until you settle in among players against whom you will win about 50% of the time. They might be playing at that skill level because they are really good pugs with a well tuned mech or a premade team who works well together but individually is only average.

It will continue to be 12 people on each side. Dropping pug will not somehow make you out-numbered. At any given skill level pugs will probably be better individually than premades because they're compensating for the absence of VOIP-assisted teamwork. The average of your whole 12 people will equate to the average of the other teams 12 people.

At the end of the match 12 people on the winning team will raise in Elo, 12 people on the losing team will drop in Elo. Their rate of rise or fall is not the same - if on the winning team 2 people had high Elo and 10 people had moderate Elo, the 2 high rankers will gain less from the win than the other 10. On the other team if all 12 had slightly above moderate Elo they will all lose a bit.

Does this make sense? This idea that somehow being on a premade team gives you some crazy magic overmind super-power to inexplicably and in defiance of mathematics win more than 50% of their matches (because they are sometimes playing against pugs I guess?) without gaining in Elo until they no longer win more than 50% is irrational.

Elo balances pug and premade.

Your Elo score in a premade team is some sort of average of your whole premade teams Elo score. What you gain or lose as part of a premade though is based on your personal skill level so that if it's 1 skilled person with a high Elo carrying 3 low Elo people, wins will benefit the 3 of them a lot and the 1 guy very little while conversely losses will hurt the 3 very little and the 1 a great deal until they are all at an average level anyway.

Elo eliminates any advantage playing as a premade gives. Any and all advantage. Totally and completely. Elo will drop you in games with and against people to give you a 50% chance of winning. Winning or losing more than 50% will raise or lower your score until you win or lose 50%. Pug, premade, it doesn't matter.

I'm going to say this again. Just because it seems to be hard for some people to grok.

Elo eliminates completely and totally any advantage playing in a premade gives you.

Again.

Elo eliminates completely and totally any advantage playing in a premade gives you.

Totally and completely.

One more time, with feeling.

Elo eliminates completely and totally any advantage playing in a premade gives you.

You will play against opponents and matched with allies that will give you 50/50 odds of winning. Premade, pug, it doesn't matter. If you have better than 50/50 odds then you'll be playing against someone else. You will not be playing (after your Elo score settles) with other than 50/50 odds. Shifts in Elo are slow and gradual once you settle. As you slowly get better you'll be dropped with and against slightly better people to keep you right around 50%.

Just in case anyone missed this.

Elo eliminates completely and totally any advantage playing in a premade gives you.

Oh, and one more thing...





.
.
.
.
.
.

Elo eliminates completely and totally any advantage playing in a premade gives you.

#103 NRP

    Member

  • PipPipPipPipPipPipPipPipPip
  • Fire
  • Fire
  • 3,949 posts
  • LocationCalifornia

Posted 12 February 2013 - 09:04 AM

I can't imagine who could think getting matched against equally skilled opponents is not a good thing, except maybe those whose skills are being masked by their team. Is that why a lot of "pro-premade" people are worried about ELO matchmaking?

Christ, anything is better than what we have now.

#104 Orzorn

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,327 posts
  • LocationComanche, Texas

Posted 12 February 2013 - 09:07 AM

View PostMischiefSC, on 12 February 2013 - 08:42 AM, said:

Elo eliminates completely and totally any advantage playing in a premade gives you.

In fact, ESPECIALLY noticeable in Dota 2, I'd argue it makes the game harder.

Look at it like this. One bad player in a game like Dota 2 makes it a 4v5. That game is both about skill and about numbers. Entering a 3v5 battle is very likely for you to lose. Entering a 2v1 is much the same. Having 1 absolutely awful player can really mess your team up, because it makes them a body and not a player. They're there to soak as much damage as their (awful and poorly farmed) hero can take, and nothing more. When you play with, for instance, 3 players in a premade, you have a REALLY high chance of ELO using 2 rather low ELO players to balance it out.

I would much, much, much rather be playing with (in addition to the 3 man premade) 2 other average players than 1 other high ELO player and a 1 really low ELO player who is absolutely garbage, because then the game effectively becomes 4v5, which is never fun. Its worse in Dota than it would be in MWO because you can feed enemies gold and experience in Dota 2 if you're a poor player who dies a lot, which then creates a snowball effect. Sure, having one guy die in MWO with 0 damage is bad, but it isn't as bad as if him dying made the enemy deal more damage, have more HP, or other such things.

So, depending on the system and the game, playing in a premade can make the game harder.

Edited by Orzorn, 12 February 2013 - 09:08 AM.


#105 Zero Neutral

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,107 posts
  • LocationEast Coast USA

Posted 12 February 2013 - 09:11 AM

View PostMischiefSC, on 12 February 2013 - 08:42 AM, said:

There are times I'm tempted to thwap myself on the forehead reading these threads.

Elo is very simple.

It doesn't matter how you got your Elo score. The whole point of your Elo score is that you have approximately 50/50 odds of beating another team with the same score. They could be pugs with a lot of natural talent. They could be a team of average players who work well together. The result is that they play at a certain level. The premade team has absolutely and in no conceivable way in any form or fashion better odds of winning against the pugs with the same Elo score. If they did, their score would be higher and they would play against whoever they have a 50/50 odds of winning against.

Please. Understand this. It's not that complicated. Truly it's not. You will either win or lose more than 50% of your matches until you settle in among players against whom you will win about 50% of the time. They might be playing at that skill level because they are really good pugs with a well tuned mech or a premade team who works well together but individually is only average.

It will continue to be 12 people on each side. Dropping pug will not somehow make you out-numbered. At any given skill level pugs will probably be better individually than premades because they're compensating for the absence of VOIP-assisted teamwork. The average of your whole 12 people will equate to the average of the other teams 12 people.

At the end of the match 12 people on the winning team will raise in Elo, 12 people on the losing team will drop in Elo. Their rate of rise or fall is not the same - if on the winning team 2 people had high Elo and 10 people had moderate Elo, the 2 high rankers will gain less from the win than the other 10. On the other team if all 12 had slightly above moderate Elo they will all lose a bit.

Does this make sense? This idea that somehow being on a premade team gives you some crazy magic overmind super-power to inexplicably and in defiance of mathematics win more than 50% of their matches (because they are sometimes playing against pugs I guess?) without gaining in Elo until they no longer win more than 50% is irrational.

Elo balances pug and premade.

Your Elo score in a premade team is some sort of average of your whole premade teams Elo score. What you gain or lose as part of a premade though is based on your personal skill level so that if it's 1 skilled person with a high Elo carrying 3 low Elo people, wins will benefit the 3 of them a lot and the 1 guy very little while conversely losses will hurt the 3 very little and the 1 a great deal until they are all at an average level anyway.

Elo eliminates any advantage playing as a premade gives. Any and all advantage. Totally and completely. Elo will drop you in games with and against people to give you a 50% chance of winning. Winning or losing more than 50% will raise or lower your score until you win or lose 50%. Pug, premade, it doesn't matter.

I'm going to say this again. Just because it seems to be hard for some people to grok.

Elo eliminates completely and totally any advantage playing in a premade gives you.

Again.

Elo eliminates completely and totally any advantage playing in a premade gives you.

Totally and completely.

One more time, with feeling.

Elo eliminates completely and totally any advantage playing in a premade gives you.

You will play against opponents and matched with allies that will give you 50/50 odds of winning. Premade, pug, it doesn't matter. If you have better than 50/50 odds then you'll be playing against someone else. You will not be playing (after your Elo score settles) with other than 50/50 odds. Shifts in Elo are slow and gradual once you settle. As you slowly get better you'll be dropped with and against slightly better people to keep you right around 50%.

Just in case anyone missed this.

Elo eliminates completely and totally any advantage playing in a premade gives you.

Oh, and one more thing...





.
.
.
.
.
.

Elo eliminates completely and totally any advantage playing in a premade gives you.


Ideally, that is how it should work, but not always in practice, as mentioned above, with the example of DOTA.

View PostMercules, on 12 February 2013 - 08:28 AM, said:


I would also argue with someone who's opinion was that the sky was blue because it reflected the color of the oceans. I would then explain Rayleigh scattering. If they still insisted it was because of the oceans I would approach it from a different direction and explain that oceans are blue because water absorbs the red end of the color spectrum and scatters blue so we see the blue reflecting back or passing through. In fact I could go into a whole lecture on color theory and light because I am a pedantic by nature.

There is a point though to continuing to argue with someone who has an opinion based off of false conclusions or invalid process.


Whatever bruh.

#106 Joseph Mallan

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 35,216 posts
  • Google+: Link
  • Facebook: Link
  • LocationMallanhold, Furillo

Posted 12 February 2013 - 09:24 AM

View PostMischiefSC, on 12 February 2013 - 08:42 AM, said:

You will play against opponents and matched with allies that will give you 50/50 odds of winning. Premade, pug, it doesn't matter. If you have better than 50/50 odds then you'll be playing against someone else.
+3%
No system is fool proof.

#107 MischiefSC

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • The Benefactor
  • The Benefactor
  • 16,697 posts

Posted 12 February 2013 - 09:38 AM

View PostOrzorn, on 12 February 2013 - 09:07 AM, said:

In fact, ESPECIALLY noticeable in Dota 2, I'd argue it makes the game harder.

Look at it like this. One bad player in a game like Dota 2 makes it a 4v5. That game is both about skill and about numbers. Entering a 3v5 battle is very likely for you to lose. Entering a 2v1 is much the same. Having 1 absolutely awful player can really mess your team up, because it makes them a body and not a player. They're there to soak as much damage as their (awful and poorly farmed) hero can take, and nothing more. When you play with, for instance, 3 players in a premade, you have a REALLY high chance of ELO using 2 rather low ELO players to balance it out.

I would much, much, much rather be playing with (in addition to the 3 man premade) 2 other average players than 1 other high ELO player and a 1 really low ELO player who is absolutely garbage, because then the game effectively becomes 4v5, which is never fun. Its worse in Dota than it would be in MWO because you can feed enemies gold and experience in Dota 2 if you're a poor player who dies a lot, which then creates a snowball effect. Sure, having one guy die in MWO with 0 damage is bad, but it isn't as bad as if him dying made the enemy deal more damage, have more HP, or other such things.

So, depending on the system and the game, playing in a premade can make the game harder.


Possibly, but remember it's going to be 12 v 12 so the balance will actually be spread out pretty well.

View PostZero Neutral, on 12 February 2013 - 09:11 AM, said:


Ideally, that is how it should work, but not always in practice, as mentioned above, with the example of DOTA.


I've gone over variances in detail elsewhere but variations will be a statistical anomaly, rare to say the least. With 12v12 matches and 8v8 matches there's enough people on each side to smooth things out for everyone...

except those poor fellas in T2, who are going to get crushed relentlessly and forever as punishment for their months of pug-stomping. If they'd been playing in 8v8 more often and having a closer to 50/50 win rate they'd either be good enough to be in T1 or they'd be kickin' it in T3. As it is they will serve as perpetual sacrifices to the T1 queue until they abandon all the stats, mechs and everything to start over from scratch.

View PostJoseph Mallan, on 12 February 2013 - 09:24 AM, said:

+3%
No system is fool proof.


Absolutely, but I think most of us will be okay with winning between 47 and 53% of games. Currently we've got folks swinging between 20% (before they quit) and 80% (oh, what a terrible fate awaits most of them).

By all accounts Murphys Law sounds like it's going to be a T1 group. Are you up for your duties as both the uber-elite, honing your skills against other top-tier teams as well as serving as infernal imps, administering perpetual punishment to the old pugstompers now confined to T2 and roasting for eternity for the crime of sync-dropping?

#108 Zero Neutral

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,107 posts
  • LocationEast Coast USA

Posted 12 February 2013 - 09:44 AM

View PostMischiefSC, on 12 February 2013 - 09:38 AM, said:


except those poor fellas in T2, who are going to get crushed relentlessly and forever as punishment for their months of pug-stomping. If they'd been playing in 8v8 more often and having a closer to 50/50 win rate they'd either be good enough to be in T1 or they'd be kickin' it in T3. As it is they will serve as perpetual sacrifices to the T1 queue until they abandon all the stats, mechs and everything to start over from scratch.


Pretty sure that every one starts at the same ELO and moves up or down from there.

#109 Greyfyl

    Member

  • PipPipPipPipPipPipPip
  • 983 posts

Posted 12 February 2013 - 09:45 AM

View PostMercules, on 12 February 2013 - 08:28 AM, said:


I would also argue with someone who's opinion was that the sky was blue because it reflected the color of the oceans. I would then explain Rayleigh scattering. If they still insisted it was because of the oceans I would approach it from a different direction and explain that oceans are blue because water absorbs the red end of the color spectrum and scatters blue so we see the blue reflecting back or passing through. In fact I could go into a whole lecture on color theory and light because I am a pedantic by nature.

There is a point though to continuing to argue with someone who has an opinion based off of false conclusions or invalid process.


I am pretty sure the sky is grey today.

#110 Havyek

    Member

  • PipPipPipPipPipPipPipPip
  • Mercenary Rank 3
  • 1,349 posts
  • LocationBarrie, ON

Posted 12 February 2013 - 09:50 AM

Personally, I'm just waiting for ELO to go with full implementation, and then the tears to start that "ELO isn't working right", "ELO dropped me against 8 man pre-mades" etc etc.


ELO may improve the experience, and it could work with 100% effectiveness from a numbers stand point. That still won't stop the tears.

#111 MischiefSC

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • The Benefactor
  • The Benefactor
  • 16,697 posts

Posted 12 February 2013 - 09:57 AM

View PostZero Neutral, on 12 February 2013 - 09:44 AM, said:


Pretty sure that every one starts at the same ELO and moves up or down from there.


They've actually been gathering Elo data for a while. Shifts are slow, if they started 60k people out at once at the same level it would take as much as 400 or 500 games for tiers to separate out.

#112 Joseph Mallan

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 35,216 posts
  • Google+: Link
  • Facebook: Link
  • LocationMallanhold, Furillo

Posted 12 February 2013 - 10:00 AM

View PostMischiefSC, on 12 February 2013 - 09:38 AM, said:

By all accounts Murphys Law sounds like it's going to be a T1 group. Are you up for your duties as both the uber-elite, honing your skills against other top-tier teams as well as serving as infernal imps, administering perpetual punishment to the old pugstompers now confined to T2 and roasting for eternity for the crime of sync-dropping?
Good lord I hope we're not among the pompous! But if we are there, then I have to tighten my belt and check my big boy pants are on the right way! :P :blink:

#113 Mercules

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Nimble
  • The Nimble
  • 5,136 posts
  • LocationPlymouth, MN

Posted 12 February 2013 - 10:05 AM

View PostGreyfyl, on 12 February 2013 - 09:45 AM, said:


I am pretty sure the sky is grey today.


I am fairly certain that is only from a localized perspective. Wait... *looks out window* yep not true here and probably not true for the majority of the sky. :P

#114 StalaggtIKE

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Raider
  • The Raider
  • 2,304 posts
  • LocationGeorgia, USA

Posted 12 February 2013 - 10:25 AM

View PostRiffleman, on 11 February 2013 - 06:29 PM, said:

Keep seeing this arguement that ELO will be a nightmare for solo players. This is pretty much backwards. Solo elo players with a high score earned that by being decent at the team, without others to rely on. This can not be said for everyone on a team. If you have 8 people on a team that play regularly, odds are one of them isnt good. Now when the teams elo rating slowly but surely rises, the bad player will be more of a hinderance as the games get tougher, in effect, an all star team is dragging dead weight with them. And when the games really matter, that person wont be able to deliver. If a 7 man team is in the same game, and they get a pickup person to fill it with the same elo rating, chances are that solo player earned that rating.


There is no way to confirm if this true or false. Because it all depends on how the ELO score is calculated for a team.
  • Does it use use their average?
  • Does it calculate a separate ELO score for the team?
  • Does it simply uses the ELO score of the team creator?
Nothing can be proven until its formula is revealed or it has been calculated through practice.

#115 Commander Kobold

    Member

  • PipPipPipPipPipPipPipPip
  • The Territorial
  • 1,428 posts
  • LocationCanada

Posted 12 February 2013 - 10:47 AM

as for the finding someone with a better ELO to fill the spot, this is assuming you can see your/other people's ELO.

#116 StalaggtIKE

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Raider
  • The Raider
  • 2,304 posts
  • LocationGeorgia, USA

Posted 12 February 2013 - 10:50 AM

View PostThontor, on 12 February 2013 - 10:36 AM, said:


Awesome. Thanks.

So why is this even being discussed? Having better players, would not make a difference in the long run, because all player's abilities are being accounted for.

Edited by StalaggtIKE, 12 February 2013 - 10:53 AM.


#117 MischiefSC

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • The Benefactor
  • The Benefactor
  • 16,697 posts

Posted 12 February 2013 - 10:51 AM

View PostJoseph Mallan, on 12 February 2013 - 10:00 AM, said:

Good lord I hope we're not among the pompous! But if we are there, then I have to tighten my belt and check my big boy pants are on the right way! B) :P


Nah, T1 isn't about being pompous. It's the pro queue. It's for people who take teamwork and winning pretty seriously. Folks who do alright in the 8v8 queues and don't avoid them because 'it's too hard' or 'it's all ECM and 3ls!' and the like but just can't find people there to play with.

See, people who work at improving will continue to improve. Teamwork sees one of its biggest strengths here, as you guys are all active or ex military you know this better than anyone. Elo functions by rewarding steady improvement. At the top of the tier you won't see any improvement by beating inferior teams so it's going to do what natural selection does best and really start to trim the wheat from the chaff. Like all challenges it'll also bring out the best in people whos response to a challenge is to rise from it.

Pompous folks don't tend to last well in that.

The best T2 teams will rise to T1 via their own aptitude, more power to them. The teams however that were only successful because they sought exploits to find and farm inferior opponents and avoided other skilled teams.... well, those people are going to be filler for T1 matches when there are not enough people around and it's going to hurt.

I genuinely think that this sort of design is where the MWO business model really shines. The really competitive teams are going to get to play against really competitive teams and when CW comes out you guys will play front line, helping shift battle lines and taking part in the higher level tactical pieces of the game.

Pugs are teams just here to play for fun with friends and don't worry too much about the competitive part of it are going to buzz around this core of elite players like Zeros around B51s, doing our thing, hoping to score a kill on a Spitfire and do enough damage to divert you off course. Indirectly involved in the larger metagame but not driving it.

This gives everyone a chance to play with other people of their own skill level while wrapped in a larger overall tactical game.

Don't think about it as 'putting on your big boy pants', think about it as the environment changing to one that really challenges you and encourages your whole team to play at their best all the time without having to feel bad about demolishing newbies or making the game less fun for anyone else....

except those people who are probably due some payback.

Good clean family fun. Everyone wins.

#118 Garth Erlam

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,756 posts
  • Facebook: Link
  • YouTube: Link
  • LocationVancouver, BC

Posted 12 February 2013 - 12:34 PM

Two things:
1) It's tough to compare this to League of Legends, as in all DoTA-esque games, having a bad player is worse than having one less player. IE. You can't 'feed' players in MWO.
2) The ELO system will help PUG vs. Premade, as Premades tend to win more, and thus, will be pitted against other premades. What you'll see a lot more of is Four person premade + 4 Lone Wolves vs. Four person premade vs. 4 Lone Wolves, simply due to the fact that the premades will perform 'higher' than someone on their own.

Also, we will tweak/change the system as time goes on and we get more metrics off of matches. As a theoretical system, though, ELO works quite well at balancing matches.

#119 BertyBargo

    Member

  • PipPipPipPipPip
  • Lucky Seven
  • Lucky Seven
  • 175 posts
  • LocationBurgerland

Posted 12 February 2013 - 01:37 PM

Garth,
Was there ever gonig to be a little icon or a number next to peoples' names that are rolling around in premades?





3 user(s) are reading this topic

0 members, 3 guests, 0 anonymous users