Jump to content

- - - - -

Matchmaking Phase 4 Follow-Up - Feedback


277 replies to this topic

#201 C E Dwyer

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 9,274 posts
  • LocationHiding in the periphery, from Bounty Hunters

Posted 30 June 2013 - 07:09 AM

View PostHayashi, on 19 April 2013 - 08:15 PM, said:

My 8 man has never gotten to drop even once. Just set the ELO tolerance to be good for 4man to PuG only, then completely eliminate ELO for 8 man games. When we're talking coordinated groups, ELO's irrelevant anyway - the results will be far more determined by the way the group works together, over individual skill - so forcing it on 8mans is pointless.



I agree with what Hayashi is saying i'd also like to further the opinion that 8 man sync drops should never face pure pug groups, at least a 4 man should be involved, and it would be better for all parties if 8 man only went up against other 8 man or two 4 man, unless of course the only reason for the 8 man was a pure CB crawl which in the long term would have a negative effect on total number of players as the pugs were rolled over

#202 C E Dwyer

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 9,274 posts
  • LocationHiding in the periphery, from Bounty Hunters

Posted 30 June 2013 - 07:17 AM

View PostAP514, on 30 June 2013 - 06:43 AM, said:

Dropped with mostly Med and a 2 Hvy meks......faced 5 ATLAS......GREAT MATCHING there...



I think it was good when this was first posted, 2 weeks ago there were some hideously unbalanced games 150 tonne mismatch was the norm, and it could top 200 and get close to 250, this weekend and since thursday the battles I've been in are evening out

Games on small maps this is an issue, larger maps not so as the weight sometimes balances with speed, for caps, its going to be 'intresting' to say the least, when 12v12 happens on small maps.

#203 Jonnara

    Member

  • PipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 184 posts
  • LocationAustralia

Posted 01 July 2013 - 09:29 PM

Not a fan of this ELO system at all.

Things will be better once a Ladder system like Starcraft's, where players with same skill levels are matched up together.

As the way things stand atm, the winning team is the one where the fails/newbie's in it, die slower than other teams fails/newbie's.

Edited by Jonnara, 01 July 2013 - 09:30 PM.


#204 Hauser

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 976 posts

Posted 04 July 2013 - 03:13 AM

View PostJonnara, on 01 July 2013 - 09:29 PM, said:

Not a fan of this ELO system at all.

Things will be better once a Ladder system like Starcraft's, where players with same skill levels are matched up together.

As the way things stand atm, the winning team is the one where the fails/newbie's in it, die slower than other teams fails/newbie's.


From what I can find about Starcrafts ladder and ranking system, it is based on a variation of Elo

http://www.teamliqui...topic_id=195273

#205 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 04 July 2013 - 06:44 AM

View PostHauser, on 04 July 2013 - 03:13 AM, said:


From what I can find about Starcrafts ladder and ranking system, it is based on a variation of Elo

http://www.teamliqui...topic_id=195273


More or less, but again, SC is 1 on 1 game just as chess. MWO isn't.

#206 Hauser

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 976 posts

Posted 04 July 2013 - 02:45 PM

View PostPhoenixFire55, on 04 July 2013 - 06:44 AM, said:

More or less, but again, SC is 1 on 1 game just as chess. MWO isn't.


SC also does 4vs4.

It doesn't matter though. As long as you're not playing with a consistent group the only thing your team mates will do is make you converge slower and less fluidly towards your actual rank. In the long run the influences from other team mates and your opponents (either positive or negative) cancel each other out. Your contribution to the team is the only constant. So if you Elo is too low, you'll win more on average and go up on average.

I did a few numerical examples of this earlier in the thread.

edit: The reason I keep up this line of argument is because you keep questioning the maths. The maths is fine. Implementation doesn't seem to be. And as we've been through often enough, there isn't enough data (not annecdotes) to discuss this at any serious level.

Edited by Hauser, 04 July 2013 - 03:26 PM.


#207 Znail

    Member

  • PipPipPipPipPipPip
  • 313 posts

Posted 04 July 2013 - 10:47 PM

View PostHauser, on 04 July 2013 - 02:45 PM, said:


SC also does 4vs4.

It doesn't matter though. As long as you're not playing with a consistent group the only thing your team mates will do is make you converge slower and less fluidly towards your actual rank. In the long run the influences from other team mates and your opponents (either positive or negative) cancel each other out. Your contribution to the team is the only constant. So if you Elo is too low, you'll win more on average and go up on average.

This is actually not true. While it is somewhat true for the really good or bad players so is it not true at all for the main 90% of the players. The Elo system doesn't suddenly lock your stats when you reach your 'true' value, you will still go up and down depending on your current luck. The problem is that there is so much static in the system. Are your team mates actual skill currently plus or minus 500 elo points? Did you get a total noob in your team, that for some reason gets an Elo in the middle of the veteran player base? So many random factors makes it mostly based on randoim luck if you win or lose and thus gain or lose Elo.

SC 2 while using a system based on Elo, so is it diffrent enough that it gets rid of most of the issues that the current MWO Elo system has. I guess that Blizzard has access to better mathematicians then PGI.

#208 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 05 July 2013 - 03:13 AM

View PostHauser, on 04 July 2013 - 02:45 PM, said:

edit: The reason I keep up this line of argument is because you keep questioning the maths. The maths is fine. Implementation doesn't seem to be. And as we've been through often enough, there isn't enough data (not annecdotes) to discuss this at any serious level.


I question both maths and implementation. The whole point of discussing this stuff on the forum is to let PGI know that its broken. I can perfectly see your point just as I'm sure you can perfectly see mine.

What worries me is that they aren't improving the matchmaker anymore. Lots of people are saying that its broken posting a ton of screenshots as proof, its not just me. It even fails to balance tonnages let alone do any real skill balancing. Games were unbalanced and still are just as unbalanced. The only thing that current matchmaker succeeds in achieveing is screwing game for groups. I know a couple dozen people who just left the game because they can't play with friends anymore.

Right now I have two options. 1st is to play in a group of 4 which means the rest of my team will be most likely filled with random people who has no clue of whats going on at all. I can hardly call playing with such people fun. 2nd option is to try and do 8 man group and then wait forever to get into game. And even if you are lucky and there is actually another 8 man doing it chances you are gonna be playing against same 8 man are really high. I can hardly call that fun as well. I'm sure I don't have to mention that 8 mans aren't being balanced at all, so if you drop a well balaned team you are still bound to run into 8 Atlases, 8 Ravens, 8 PPC Stalkers or whatever else the flavour of the month is. No fun in there as well...

The whole phase 1-4 bullcrap ruined the game completely for groups while at the same time it didn't improve game for PUGs. Elo matchmaker is useless, either improve it or get rid of it entirely.

Edited by PhoenixFire55, 05 July 2013 - 03:14 AM.


#209 Hauser

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 976 posts

Posted 05 July 2013 - 04:39 AM

View PostZnail, on 04 July 2013 - 10:47 PM, said:

This is actually not true. While it is somewhat true for the really good or bad players so is it not true at all for the main 90% of the players. The Elo system doesn't suddenly lock your stats when you reach your 'true' value, you will still go up and down depending on your current luck. The problem is that there is so much static in the system. Are your team mates actual skill currently plus or minus 500 elo points? Did you get a total noob in your team, that for some reason gets an Elo in the middle of the veteran player base? So many random factors makes it mostly based on randoim luck if you win or lose and thus gain or lose Elo.

SC 2 while using a system based on Elo, so is it diffrent enough that it gets rid of most of the issues that the current MWO Elo system has. I guess that Blizzard has access to better mathematicians then PGI.


Well obviously you'll move around. But you'll be near the rank you should be at all the same which is the whole point. All the random factors that influence your team also apply with equal chance to other red team. In the long run it evens out. I did some numerical examples of that earlier in the thread.

Now I'm not saying there is no room for improvement here, but the influence of your team mates generally cancels out.

View PostPhoenixFire55, on 05 July 2013 - 03:13 AM, said:

I question both maths and implementation. The whole point of discussing this stuff on the forum is to let PGI know that its broken. I can perfectly see your point just as I'm sure you can perfectly see mine.

What worries me is that they aren't improving the matchmaker anymore. Lots of people are saying that its broken posting a ton of screenshots as proof, its not just me. It even fails to balance tonnages let alone do any real skill balancing. Games were unbalanced and still are just as unbalanced.


Anecdotes aren't data, even if you have allot them. The only thing I ever see players complain about are lost games. Lost games in which they did well, but their team ****** up badly. I never see someone complain about a win in which they did 25 damage. For the games people complain about we don't know what the server population was, what was in the queue and for how long. We also don't know how frequently these lopsided games occur or even what the matchmaker expected the outcome to be. As such the games people complain about are useless for evaluating the matchmaker.

But to counter it with an anecdote of of my own. Before the matchmaker was implemented I used to have 5.0 win loss ratio and 3.0 k/d in a HBK-4P. These days Its at 1.30 w/l and 1.6 k/d because now I'm generally playing with people who know where to aim. I would say this means the match maker is doing something right.

Beyond that I'm just going to ignore every thing you say about groups. Even without a matchmaker limiting groups to 4 players would have been a good move. Groups would generally stomp over everything.

Edited by Hauser, 05 July 2013 - 04:42 AM.


#210 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 05 July 2013 - 05:37 AM

View PostHauser, on 05 July 2013 - 04:39 AM, said:

Anecdotes aren't data, even if you have allot them. The only thing I ever see players complain about are lost games. Lost games in which they did well, but their team ****** up badly. I never see someone complain about a win in which they did 25 damage. For the games people complain about we don't know what the server population was, what was in the queue and for how long. We also don't know how frequently these lopsided games occur or even what the matchmaker expected the outcome to be. As such the games people complain about are useless for evaluating the matchmaker.


Oh I complain a lot about 8:0, 8:1 wins myself. Or 0 damage wins when I'm in a Jenner on Alpine conquest just capping all 5 bases when my entire team is dead and I never see/take fire from any enemy mechs at all. When a matchmaker gives one team 6-7 assault mechs and places 4-5 mediums on the other team in assault mode or when one team has 4 lights while other team fastest mech is Hunchback in conquest mode it only means its broken, no matter if I win or lose really.

View PostHauser, on 05 July 2013 - 04:39 AM, said:

But to counter it with an anecdote of of my own. Before the matchmaker was implemented I used to have 5.0 win loss ratio and 3.0 k/d in a HBK-4P. These days Its at 1.30 w/l and 1.6 k/d because now I'm generally playing with people who know where to aim. I would say this means the match maker is doing something right.


Nothing to do with matchmaker if you ask me. People are using 'cheese' builds and PPCs on pretty much everything, which means that live expectancy of your Hunch is less then it used to be. Any Hunch is a brawler and brawlers die first and they die fast. Best brawlers in tournaments end up around 12th-15th places while top of the list is always some sort of 'sniper' build.

View PostHauser, on 05 July 2013 - 04:39 AM, said:

Beyond that I'm just going to ignore every thing you say about groups. Even without a matchmaker limiting groups to 4 players would have been a good move. Groups would generally stomp over everything.


You ignore it because just like PGI you have nothing to say to justify this crap. Poor puggies were getting stomped by groups, but surprise surpise ... its a friggin team game and a better team will win. Thing is ... poor puggies are still getting stomped by other puggies and they still cry about 'premades' and such. Its fine to limit group sizes in PUG games if you give groups an alternative way to play. But really, its such a nice FU gesture from PGI to groups, to take everything away from us and give us nothing in return. Make a bloody separate queue for PUG-only mode and another one that allows you to play with whatever number of people you have in your group. Right now if your in a group your pretty much screwed, they force you to PUG.

I shouldn't be surprised tho, PGI said that they don't give a damn about groups. They want a stupid PUG game for stupid PUGs who want 'fair' matchmaker that puts bads vs bads so even bads have 50% win rate yaay!

I only have one more thing to say. When CW comes out and we can finally do private matches ALL groups will leave PUG queue forever. I'm pretty sure that all players from units are gonna have a lot of fun without PUGs. As for PUGs, well, GL waiting 10 mins for every match and GL getting any sort of balanced matches out of Elo system. I wonder how many people in PUG queue will be keeping up with this crap after CW starts.

#211 Tsenado

    Member

  • PipPipPip
  • Overlord
  • Overlord
  • 90 posts

Posted 05 July 2013 - 05:38 AM

How about giving PUGers an option to not join part of a premade? if a PUGer wants to be in a premade they would have done it with their friends, not some random premade groups. 90% of the time when I PUG I am stuck in some sort of premade.

Think of a new player joins MWO, they just want to try, and all the games they joined are premades and instantly get slaughtered, and more slaughtering, will they return and play a game that's just keep themselves slaughtered?

Some people might think this game is all about teams, but how do these new players know about teaming when they are instantly turned off by getting slaughtered by the oh-so-godly premade teams?

Edited by Tsenado, 05 July 2013 - 05:52 AM.


#212 Hauser

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 976 posts

Posted 05 July 2013 - 06:09 AM

View PostPhoenixFire55, on 05 July 2013 - 05:37 AM, said:

Oh I complain a lot about 8:0, 8:1 wins myself. Or 0 damage wins when I'm in a Jenner on Alpine conquest just capping all 5 bases when my entire team is dead and I never see/take fire from any enemy mechs at all. When a matchmaker gives one team 6-7 assault mechs and places 4-5 mediums on the other team in assault mode or when one team has 4 lights while other team fastest mech is Hunchback in conquest mode it only means its broken, no matter if I win or lose really.


Again anecdotes are not data. And to make it worse you're only describing parts of both teams. Last time we did a systematic check of what people were dropping in the weights weren't that far off and the extra tonnage did not increase the chance to win.

http://mwomercs.com/...19#entry2220919

And lastly, as much as you see complaints on the forums about the matchmaker, the weights are roughly even (the maximum was 135 tonnage difference), and the win rate of the heavier team is roughly 50%, which seems to actually suggest it's doing something right.


Now the data is all there feel free to run your own tests and see if specific combinations win assault more often then others.

View PostPhoenixFire55, on 05 July 2013 - 05:37 AM, said:

Nothing to do with matchmaker if you ask me. People are using 'cheese' builds and PPCs on pretty much everything, which means that live expectancy of your Hunch is less then it used to be. Any Hunch is a brawler and brawlers die first and they die fast. Best brawlers in tournaments end up around 12th-15th places while top of the list is always some sort of 'sniper' build.


Not at all. Koreanse won the first 6 days in the Spring Tournament. I won the 7th. Both of us were in a 9mlas 4P.

And for a while now I've been telling you to make a new account. Just to see what it is like playing at the bottom of the Elo barrel. Play the champion trial mech and you find to be ridiculously easy.

Edited by Hauser, 05 July 2013 - 06:14 AM.


#213 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 05 July 2013 - 07:02 AM

View PostHauser, on 05 July 2013 - 06:09 AM, said:

Again anecdotes are not data. And to make it worse you're only describing parts of both teams. Last time we did a systematic check of what people were dropping in the weights weren't that far off and the extra tonnage did not increase the chance to win.


Of course it doesn't.

In assault you want more tonnage in conquest you want more speed. But, any assault game can turn into a cap race and then you need more speed, just like any conquest can turn into big fight in the middle of the map and you need more tonnage. Thing is, when two teams are evenly matched it doesn't matter what game mode and what situation they are in, its even chance for either of them to win. But if they aren't evenly matched then depending on a situation the team that needs tonnage and has tonnage or team that needs speed and has speed will win, other team won't have a chance. Now the probability of those situations occuring are about same, which only means that weight disbalance does not affect your overall win %. What it does affect is the outcome of each given match. If your team is in a situation when it needs speed and it doesn't have it then you will lose, there is NO way to overcome that (with every other aspect like skill balanced).

Also, there can be no tonnage difference between teams (and I've already described this to you before). One team has 4 mediums and 4 heavies and another team has 4 assaults and 4 lights. Well surprise surpise ... team.1 is screwed. With other factors being even they have no chance, because they have no speed nor firepower to catch lights and kill assaults. Perfect tonnage matching and yet games will always end up same way...

Balanced match means it can go either way. Skill can be balanced (miracles do happen) and tonnage can be balanced (which as I said means nothing) but its not a balanced match, because depending on what mechs each team has the outcome of the game is already decided before it starts.

I've played enough games and I know a lot of players. I can tell you the outcome of each match I play with 90% success rate before it starts depending on people I see on both teams and depending on mech types I see in my team.

View PostHauser, on 05 July 2013 - 06:09 AM, said:

Not at all. Koreanse won the first 6 days in the Spring Tournament. I won the 7th. Both of us were in a 9mlas 4P.


Substitute my 'sniper' build for 'alpha' build and there you go. Your win/loss and KD ratios suffering only means that PPC cheese is much better then your ML cheese nowadays. And don't forget the fact that like 80% people drop in assault mechs nowadays. Much harder to drop 7 assaults in that Hunch then it is to drop 7 mediums eh?

View PostHauser, on 05 July 2013 - 06:09 AM, said:

And for a while now I've been telling you to make a new account. Just to see what it is like playing at the bottom of the Elo barrel. Play the champion trial mech and you find to be ridiculously easy.


And I've told you I'm not gonna do that. Am I supposed to create a new account each time my previous one gets into high elo or whatever zone I am in? Is it only horribads that can have fun in this game now? I am NOT gonna become a horribad, I am NOT gonna drop my elo by whatever way including creating a new account. I am not gonna have 'FUN' accounts and 'SKILL' accounts.

Instead of being fun for everyone PGI insists on making it fun for 'not-so-good' PUG-only players. Because the higher you go with your elo score the more cheese there is, the more frustration there is. Because there is nothing but frustration when one team has all good players and other team has 4 people who fail to do anything useful at all. And there is nothing but frustration when one team has good balanced mechs and other team has 6-7 mechs with 4 PPCs each or 6-7 mechs with dual AC20s each. Because THAT is how your matchmaker works.

PGI can post whatever data they want. They'll even tell you that Sun revolves around Earth to cover their fails. They have made a tremendous amount of fail decisions and they are yet to man up and admit even one. Only they have the data so they always have an excuse like 'it works as intended' or 'we see no disbalance'. But as long as I see that my experience is different I am gonna keep saying the same thing ... its broken.

And as I said, for a whole year PGI did nothing but ruin game for organized players. A team game without teams is just another fail. Your average PUG player will play for a year and leave forever, but if you make group players leave the game then in a year you'll have a game without players at all.

#214 Hauser

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 976 posts

Posted 05 July 2013 - 07:51 AM

View PostPhoenixFire55, on 05 July 2013 - 07:02 AM, said:

Of course it doesn't.

In assault you want more tonnage in conquest you want more speed. But, any assault game can turn into a cap race and then you need more speed, just like any conquest can turn into big fight in the middle of the map and you need more tonnage. Thing is, when two teams are evenly matched it doesn't matter what game mode and what situation they are in, its even chance for either of them to win. But if they aren't evenly matched then depending on a situation the team that needs tonnage and has tonnage or team that needs speed and has speed will win, other team won't have a chance. Now the probability of those situations occuring are about same, which only means that weight disbalance does not affect your overall win %. What it does affect is the outcome of each given match. If your team is in a situation when it needs speed and it doesn't have it then you will lose, there is NO way to overcome that (with every other aspect like skill balanced).

Also, there can be no tonnage difference between teams (and I've already described this to you before). One team has 4 mediums and 4 heavies and another team has 4 assaults and 4 lights. Well surprise surpise ... team.1 is screwed. With other factors being even they have no chance, because they have no speed nor firepower to catch lights and kill assaults. Perfect tonnage matching and yet games will always end up same way...

Balanced match means it can go either way. Skill can be balanced (miracles do happen) and tonnage can be balanced (which as I said means nothing) but its not a balanced match, because depending on what mechs each team has the outcome of the game is already decided before it starts.

I've played enough games and I know a lot of players. I can tell you the outcome of each match I play with 90% success rate before it starts depending on people I see on both teams and depending on mech types I see in my team.


If that is what you think, go through the data then. Look for matches were you have teams with a large standard deviation on tonnage playing against teams with a small std. deviation on tonnage. Shouldn't be too hard to prove your point.

edit: For your convenience, you may need to send hammer a PM for the spread sheets.

http://mwomercs.com/...19#entry2220919

View PostPhoenixFire55, on 05 July 2013 - 07:02 AM, said:

Substitute my 'sniper' build for 'alpha' build and there you go. Your win/loss and KD ratios suffering only means that PPC cheese is much better then your ML cheese nowadays. And don't forget the fact that like 80% people drop in assault mechs nowadays. Much harder to drop 7 assaults in that Hunch then it is to drop 7 mediums eh?


Hardly. Against assaults and heavies it is much easier to put the damage on target. It's the mediums and lights that tend to ruin the day. Have you ever even played a 4P?

View PostPhoenixFire55, on 05 July 2013 - 07:02 AM, said:

And I've told you I'm not gonna do that. Am I supposed to create a new account each time my previous one gets into high elo or whatever zone I am in? Is it only horribads that can have fun in this game now? I am NOT gonna become a horribad, I am NOT gonna drop my elo by whatever way including creating a new account. I am not gonna have 'FUN' accounts and 'SKILL' accounts.


And again I'm not telling you to play it permanently. I'm telling you to play it to see what it's like at the bottom of the barrel. Seriously, play 3 games. If you're any good you'll win them all without even trying. You'll win simply by the fact that you can pilot your mech and shoot at the same time.

View PostPhoenixFire55, on 05 July 2013 - 07:02 AM, said:

Instead of being fun for everyone PGI insists on making it fun for 'not-so-good' PUG-only players. Because the higher you go with your elo score the more cheese there is, the more frustration there is. Because there is nothing but frustration when one team has all good players and other team has 4 people who fail to do anything useful at all. And there is nothing but frustration when one team has good balanced mechs and other team has 6-7 mechs with 4 PPCs each or 6-7 mechs with dual AC20s each. Because THAT is how your matchmaker works.

PGI can post whatever data they want. They'll even tell you that Sun revolves around Earth to cover their fails. They have made a tremendous amount of fail decisions and they are yet to man up and admit even one. Only they have the data so they always have an excuse like 'it works as intended' or 'we see no disbalance'. But as long as I see that my experience is different I am gonna keep saying the same thing ... its broken.


PGI did not post that data. What are you talking about? Did you even read the data collection thread?

I also get the nagging suspicion that you only think the game is fun if you're stomping over other people. I might understand how this system upsets you, it's meant to prevent exactly that.

Edited by Hauser, 05 July 2013 - 08:00 AM.


#215 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 05 July 2013 - 11:12 AM

View PostHauser, on 05 July 2013 - 07:51 AM, said:

If that is what you think, go through the data then. Look for matches were you have teams with a large standard deviation on tonnage playing against teams with a small std. deviation on tonnage. Shouldn't be too hard to prove your point.


With all due respect for his work (and work done by people helping him) I think it is irrelevant. When any player does a data gathering like that he does it in his elo zone. When a group of people does it its that groups combined elo score zone. If I'm gonna do that its gonna by my zone. Only PGI has the data and they don't show it... why? If its working you have nothing to hide, if its not working, well ...

View PostHauser, on 05 July 2013 - 07:51 AM, said:

Hardly. Against assaults and heavies it is much easier to put the damage on target. It's the mediums and lights that tend to ruin the day. Have you ever even played a 4P?


I find it funny. Yeah sure any mech does better % of its damage into bigger targets but seriousely how long will you live vs smth like a Highlander? 732 (in the basic PPC-Gauss build) has 45 damage alpha. It takes 2 shots to destroy your hunch and he is more heat efficient then you. Then what are you gonna do with those 2 MLs you'll have left? And with auto convergence it takes next to zero skill to destroy that huge hunch.

And yes, my 4P is mastered.

View PostHauser, on 05 July 2013 - 07:51 AM, said:

And again I'm not telling you to play it permanently. I'm telling you to play it to see what it's like at the bottom of the barrel. Seriously, play 3 games. If you're any good you'll win them all without even trying. You'll win simply by the fact that you can pilot your mech and shoot at the same time.


Maybe its different in different elo zone but I don't care! I care about me and my zone. As I said, if this game is only fun for not-so-good players then its broken.

View PostHauser, on 05 July 2013 - 07:51 AM, said:

PGI did not post that data. What are you talking about? Did you even read the data collection thread?


I never said they posted the data. All I said is they gave that ... 'matchmaker is working as intended' and 'we see no serious disbalance'. Now as I said I don't care if 99% of games that does not include me are balanced, mine aren't.

View PostHauser, on 05 July 2013 - 07:51 AM, said:

I also get the nagging suspicion that you only think the game is fun if you're stomping over other people. I might understand how this system upsets you, it's meant to prevent exactly that.


I told you long ago that I liked the 'random' matchmaker we had back in closed beta that was matching mech types. It wasn't perfect but at least it didn't lie to me. Elo matchmaker claims it gives me balanced matches and it is a lie. Back in CB I was playing solo and was stomping and was getting stomped, I was playing in a group with all friends I had availiable and we were stomping and we were getting stomped. And it was fun, and it was far better balanced. If your team is better you win, if you team is worse you lose. Shame that kids who play this game can't live with it and want 'fair' matchmaker. Well there is nothing fair in elo based matchmaker. It is broken and it is wrong.

#216 Rebas Kradd

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,969 posts

Posted 05 July 2013 - 12:06 PM

View PostPhoenixFire55, on 05 July 2013 - 11:12 AM, said:

Well there is nothing fair in elo based matchmaker. It is broken and it is wrong.

What was your suggested alternative, again? I looked for it in your previous posts and couldn't find it...

#217 Hauser

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 976 posts

Posted 05 July 2013 - 12:07 PM

View PostPhoenixFire55, on 05 July 2013 - 11:12 AM, said:

With all due respect for his work (and work done by people helping him) I think it is irrelevant. When any player does a data gathering like that he does it in his elo zone. When a group of people does it its that groups combined elo score zone. If I'm gonna do that its gonna by my zone. Only PGI has the data and they don't show it... why? If its working you have nothing to hide, if its not working, well ...


Okay, just let me get this straight. Even before you've tried to look through the data you're going to assume the data doesn't include what you're looking for, and therefore don't bother at all?

View PostPhoenixFire55, on 05 July 2013 - 11:12 AM, said:

I never said they posted the data. All I said is they gave that ... 'matchmaker is working as intended' and 'we see no serious disbalance'. Now as I said I don't care if 99% of games that does not include me are balanced, mine aren't.

(....)

Maybe its different in different elo zone but I don't care! I care about me and my zone. As I said, if this game is only fun for not-so-good players then its broken.


If things are different in a different elo zone, that means the match maker is working.

If you think it is unfair that you have to play your equals I think you need to take a damn good hard look at yourself. If you think your equals are poor players, guess what you are? Either try to improve yourself, or stop playing to win and just have fun.

Edited by Hauser, 05 July 2013 - 12:14 PM.


#218 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 06 July 2013 - 12:55 AM

View PostRebas Kradd, on 05 July 2013 - 12:06 PM, said:

What was your suggested alternative, again? I looked for it in your previous posts and couldn't find it...


Because suggestions are in different much older threads.

For starters get rid of elo balancing thats not working and bring back mech type balance. Implement real 'skill' based balancing if you want, as a start you can make it based off game score instead of wins and losses, then improve the game score that it includes much needed bonuses for actions others then doing damage, real scouting, keeping locks and such. Then get rid of group limitations and make two separate queue for solo-only players and groups of any size up to 8 and up to 12 when 12 v 12 comes out. Remove stupid must have 8 players to start a game limit and allow 7 v 7 and 6 v 6 drops in groups queue and possibly PUG gueue. As a bonus you can make matchmaker always place at least 1-2 mechs of each type into any team. In solo queue light pilots will find games instantly (because they'll be needed everywhere) while assault pilots will wait longer to get into games (as there are more then enough assaults waiting for game). It'll naturally balance out mech types among players and won't allow matches like 8 assaults on 8 assaults happen. In groups queue do the same, it'll make players who group up build balanced groups instead of bringing all-PPC-Stalkers or all-AC40-Jagers or light-****-squads.

As I said I'm not a big fan of any form of skill balancing in matches because I don't think it is necessary or even possible. Too many factors determine real skill, you can't transform them into single numerical value, thus any skill balancing in matches isn't possible. Most importantly BT was never about 'balanced', better units will always stomp worse units. I've been happy with the CB matchmaker, just make two separate player queues if you can't live with OMG premades stomping you.

My biggest concerns with the thory of a matchmaker we have now are the facts that even if you hit 1000+ dmg and kill 6 mechs and your team still loses you lose elo score, even tho you obviousely played well. Also, matching good players vs good players and bads vs bads makes them earn same amount of c-bills and exp in their games even tho the effort and skill required to do so are very different. Apart from that its just the fact that matchmaker isn't working, things I mentioned about getting 3-4 trial mechs on one side and 2 premade groups on the other side etc. or impossible to win situations haveing a bunch of assaults vs a group of mediums in assault games or smth like 4 lights vs none on conquest.

#219 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 06 July 2013 - 01:20 AM

View PostHauser, on 05 July 2013 - 12:07 PM, said:

Okay, just let me get this straight. Even before you've tried to look through the data you're going to assume the data doesn't include what you're looking for, and therefore don't bother at all?


I did look at all data sheets at that post. I can't make an analysis I wanted out of them because as I said to do it you need to look into team compositions in every given game rather then in multiple games on average. I can make this myself in say a week, but again ... it will only be relevant to my elo zone and as I have no idea what zone it is I can't judge if its any relevant for any significant number of people apart from myself.

View PostHauser, on 05 July 2013 - 12:07 PM, said:

If things are different in a different elo zone, that means the match maker is working.

If you think it is unfair that you have to play your equals I think you need to take a damn good hard look at yourself. If you think your equals are poor players, guess what you are? Either try to improve yourself, or stop playing to win and just have fun.


If its different as in 'noob zone is fun and no-noob zone is no-fun' then its working wrong.

Playing equals isn't the case. I'm either playing with a bunch of noobs vs a team full of equals or with a team full of equals vs a bunch of noobs. It is supposed to balance average elo scores for teams, well maybe it does on paper but it does not lead to balanced games. I don't care why it happens, either maths or the code is broken, all I care about is the fact that its not working.

I do think thats its unfair that I always need to bring cheese mechs, because if I don't I'm not competitive at all. I do think thats its unfair that I always have to stay on tip of my toes in every single game otherwise I'll be killed easily and will let my team down. I do think its unfair that one team gets 4 trial mechs while other team gets two 4-man premades. I do think its unfair that one team gets 7 assault mechs due to the fact that they have a group of 4 assaults and a group of 3 assaults while other team is loaded with mediums and heavies all solo players. There is no fun whatsoever when this happens regardless if you win or lose, thus matchmaker is a fail.

And before you tell me that its all anecdotes I'll tell you that one or more of these anecdotes happen in nearly every single game I play. All games we've met and played together over the last two months were either extremely one-sided or disbalanced so you should have seen it for yourself.

Edited by PhoenixFire55, 06 July 2013 - 01:21 AM.


#220 Hauser

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 976 posts

Posted 06 July 2013 - 03:30 AM

View PostPhoenixFire55, on 06 July 2013 - 01:20 AM, said:

I did look at all data sheets at that post. I can't make an analysis I wanted out of them because as I said to do it you need to look into team compositions in every given game rather then in multiple games on average. I can make this myself in say a week, but again ... it will only be relevant to my elo zone and as I have no idea what zone it is I can't judge if its any relevant for any significant number of people apart from myself.


You can use the raw data sets for that. Mechs per team per match were recorded. But you're better of if you send a mail to hammer for the raw data sets in a spreadsheet.

View PostPhoenixFire55, on 06 July 2013 - 01:20 AM, said:

Playing equals isn't the case. I'm either playing with a bunch of noobs vs a team full of equals or with a team full of equals vs a bunch of noobs. It is supposed to balance average elo scores for teams, well maybe it does on paper but it does not lead to balanced games. I don't care why it happens, either maths or the code is broken, all I care about is the fact that its not working.

I do think thats its unfair that I always need to bring cheese mechs, because if I don't I'm not competitive at all. I do think thats its unfair that I always have to stay on tip of my toes in every single game otherwise I'll be killed easily and will let my team down. I do think its unfair that one team gets 4 trial mechs while other team gets two 4-man premades. I do think its unfair that one team gets 7 assault mechs due to the fact that they have a group of 4 assaults and a group of 3 assaults while other team is loaded with mediums and heavies all solo players. There is no fun whatsoever when this happens regardless if you win or lose, thus matchmaker is a fail.

And before you tell me that its all anecdotes I'll tell you that one or more of these anecdotes happen in nearly every single game I play. All games we've met and played together over the last two months were either extremely one-sided or disbalanced so you should have seen it for yourself.


I didn't wanna say this, but the ones where I'm not playing with you tend to go allot better. The feeling is probably mutual. :)

Anyway, I agree with you that always playing on the tip of your toes isn't fun. That's a reasonable argument to make and irrelevant of the matchmaker used. When trying to setup matches anything else is unfair for someone else though so there is no way around. I hope it soon will be possible to setup private matches and just play for fun against friends.

I also agree that playing against premades isn't always fun and it's kinda obvious that when they all pick assaults or lights the matchmaker has a hell of a time trying to find suitable opponents. This needs to be solved.

Though I don't agree with you when you say the match maker is a fail, that is simply overstating the problem, the matchmaker has a few very specific failure cases. Nobody is helped by damning the whole system for a few specific flaws. The whole point of this thread is to articulate these failure cases.





5 user(s) are reading this topic

0 members, 5 guests, 0 anonymous users