Matchmaker And Elo, Epic Fail
#61
Posted 23 December 2013 - 05:10 PM
If your ELO hits a certain level you can only field Locusts with a single MG and half armor and the smallest engine available until you lose enough games to make it "fair"
#62
Posted 25 December 2013 - 04:40 PM
wwiiogre, on 23 December 2013 - 10:43 AM, said:
The point of this discussion is not winning and losing. The point is the matchmaker is not attempting to make a balanced game. It is not using some of the most important things that should be included when setting up a match. When one side has a 300 ton advantage, that is the same as having three extra Atlas mechs. Now, instead of the weight advantage, there would be no argument if in fact one side did have 3 xtra Atlas mechs drop instead of the tonnage advantage. \
Tonnage does not need to be exact, but should be less than 100 tons total in every match. Equipment value and values of mechs used should be close in every match, not equal but close. Say within 5 million cbills. That way you do not have trial mechs facing elite mechs. You have near equal tonnage and near equal mechs and equipment including modules. So far the above two would go a long way to leveling the playing field.
Then all you have left is player experience and skill and the ability to build an efficient killing machine in the mech lab. There is nothing to whine over, because the game started balanced. If you lose when this occurs you have nothing to blame but yourself and your teamwork. Which is what I am hoping PGI is trying to do.
But currently the matchmaker does not do this, it does not even attempt to do this because the above two things are not part of ELO. They are not even a small part of it. PGI has decided to go the route of everyone gets a ribbon for participating by manipulating win/loss records so nobody can brag about their amazing skills on the forum. Instead you get this horrible matchmaker that literally stacks most games for oneside to win. Giving them more 4mans, giving one side better mechs, giving one side more tonnage. All based on win/loss ELO of pilots in that mech.
Sorry, win/loss k/d is not how to rate a pilot especially when there is no tonnage matchmaking or mech/equipment/module matchmaking. How much skill do you have when you are in a 12 man with all assaults/heavies vs another 12 man in lights and mediums? None, sorry no skill involved when you beat up 5 year olds. Or literally come to the fist fight with a machinegun.
The entire point of this post is to point out PGI has utterly failed in matchmaking. The games and drops are not balanced, and therefore not fun whether winning or losing. a 12-0 is not fun. Its easy and stat padding. Its ok for people that like beating up on 5 year olds but for serious competitive players the current matchmaker is a joke and just like getting a ribbon at school for participating its worthless and really doesn't feel like anything important was accomplished when you win vs a team that had 4 trial mechs and you out weighed by 300 tons.
PGI needs to either completely scrap the current matchmaker or add the things I have been saying as well as many others. BV or Battle Value from tabletop works for equipment and mechs, but would need to be reworked since PGI has changed so many values. Plus BV has a rating for pilots based on table top skill levels which would not work in MWO. So a system for evaluating the skill of a pilot would be needed.
I do not think win/loss or even k/d work as a value for pilot skill. Why? Because currently those marks were earned under a system that does not evenly match mech vs mech on a ton or equipment basis. If every game a pilot had played up till now had been against other teams evenly tonned and evenly equipped then win/loss and k/d would be a valid part of an equation to rate a pilot. But that is not the case at the moment. The only thing I think that could be used at this point is to actually count total games played.
Meaning you rate pilots by experience, with experience being how many missions/drops they have made. You would then have rookie (still on first 25 missions); green (26-100), regular (100-500); veteran (500-1000) and elite (1000+) drops. This gives you an idea of how many times a player has played. You can do the same with mech based on type of equipment used. A mech with standard equipment, meaning no double heat sinks, no endo, no FF, no artemis, no tag, no ecm, no bap, no command module, no ams, no XL engine, no ER weapons, no pulse weapons, no streak, no gauss.
Give each piece of upgraded equipment a value, add that to the tonnage of the mech, given a number, do this with every mech in a 12 man and then attempt to find equal value opponents for them. So tonnage and equipment are easily figured, just some math and a discussion of PGI devs to decide what equals what. Then a simple rating system for pilots based on drops and not win/loss or k/d. Because if mechs are near equal and tonnage is near equal. I am ok facing a far superior pilot and eating a loss because his skill was better than I am facing a near certain loss because my team is out tonned and out gunned from the beginning.
I am not really asking for much, just a system that means matchmaking actually tries to make even matches. Not to manipulate win/loss records so every body has better self esteem. I know Canadians are nice and polite, but this game is about killing the other guy. Not feeling good because you participated. Give us near even equipment and let the bodies fall where they may.
Chris
Chris,
I agree with your points on matchmaker, but focusing solely on tonnage is not the way to go. Battle Value summations equate to the fairest matches. I have been playing and GMing Tabletop (CBT) for over two decades and am always rewarded with fair matches when the BV is within a couple hundred points. Please see my earlier post as a roadmap for the implementation of such a system.
Kyle "TAZ" Kerns
#63
Posted 26 December 2013 - 08:33 AM
...he loves to watch the continued imbalance so long as it means his premade time can be pugstomps due to tonnage and skill mismatching with no assistance from him (because he's mediocre at best).
But you're completely right, the matchmaker currently barfs any time a premade group runs heavy or light in tonnage/ELO, and tonnage limits will do much to resolve this resulting in closer matches and less roflstomps. I look forward to it...
-billyM
Edited by BillyM, 26 December 2013 - 08:33 AM.
#64
Posted 26 December 2013 - 10:38 AM
Unfortunately, PGI is not acting like a GM, they are acting like Politcally Correct school guidance experts trying to give everyone good self esteem by manipulating overall win/loss records by punishing anyone with a positive winning record by stacking the game against them. Instead of having a combat game that gives each side an even and balanced playing field where equipment, mech tonnage and pilot experience are near equal, PGI goes so far as to stack a game with upto 450 tons more on one side, more 4 mans and better equipment vs all solo pugs, with 4 trial mechs and new players and less tonnage. How is that even possible? It shouldn't be, yet I see it time and again every single day. Usually in the 250-350 ton advantage.
So I give PGI a failing grade for its matchmaker which really isn't even a matchmaker. It is nothing more than a bad attempt to give every person a participation ribbon by grossly manipulating sides in a match so everyone has a near average record. It is horrible, it is not fun and it is getting very old, very fast. And it is why so many of the veteran players are getting burned out.
It is a sad state of affairs when the only way to get a fair and balanced game is to participate in a fan run league like the Marik Civil war event. Where tonnage and equipment were regulated and the matches were lots of fun and very challenging.
Yet, PGI has continually decided that building everyone's self esteem comes before fair and balanced matches.
With a good matchmaker, new players should be playing against other new players if they solo drop. Trial mechs should be playing against other less elite equipped mechs if they solo pug drop. Now when you 2, 3 or 4 man drop you should be facing other groups not all solo pugs.
I am sorry the above is basic programming and should be very easy to setup and would go a long way to making the matches more balanced and a lot more fun.
Chris
#65
Posted 28 December 2013 - 08:04 PM
If you have a better solution, please present it so we can help improve this game.
Kyle
#66
Posted 15 January 2014 - 05:23 PM
Currently there is no way to judge pilot skill since matchmaker is broken, I propose using total number of drops in any mech as a base. Then total mech value as another base and total tonnage as another base. Thus you have three main ways to balance a match. Tonnage should never be more than 95 tons difference. Ever! Period! Total mech value including modules and upgraded weapons, etc should have no more than 20 million cbill advantage to one side or a nicely outfitted Atlas. So no team gains what is the equivalent of more than a single assault mech as an advantage. Number of drops could be used as well with people with less than 50 drops always in games with other players with 50 or less drops.
Note these ideas for matchmaker are based on solo drop or 2man drops. 3man and 4man drops should always face each other only. Meaning unless a solo player chooses to be placed as a solo in the 3man/4man que he would never face other teams. This means you get less rofl stomps because of tech/tonnage/premades. You will still get some rofl stomps but they will happen because of skill and teamwork and not because of having a huge advantage that teamwork and skill cannot usually overcome.
Just saying.
Chris
#67
Posted 15 January 2014 - 10:53 PM
Still there is some people who thinks stomping is fun. Newsflash: There is game mode for that, it's training grounds.
There is no simple way to balance drops, because game is complex in many ways.
We are really missing user options of game making/ finding.
How about sync drop of 2 lances. If you have 2 good teams, I bet new player in third lance, gets more out of this game.
If we have few good players and all others are new comers, its not even if other team is medium( skill level) players.
So simple balancing equation is not balancing. Also tonnage is effecting.
For me this is not winning and losing, I want to have game play. At the moment game play is only content. There is no other point, until clan and team matches are ongoing.
In development point, new mechs, and modules are content, but these aren't solving ongoing issue.
I'm bit worried, because PGI is not giving any comments of match making development. There is UI 2.0, what is resolving everything, or is it?
#68
Posted 16 January 2014 - 05:11 AM
#70
Posted 17 January 2014 - 07:20 AM
BillyM, on 26 December 2013 - 08:33 AM, said:
...he loves to watch the continued imbalance so long as it means his premade time can be pugstomps due to tonnage and skill mismatching with no assistance from him (because he's mediocre at best).
But you're completely right, the matchmaker currently barfs any time a premade group runs heavy or light in tonnage/ELO, and tonnage limits will do much to resolve this resulting in closer matches and less roflstomps. I look forward to it...
-billyM
That might be true... If the Law dropped more often together. I have more PUG drops than Lawman drops by far over the last 8 months.
You really need to work harder if you want to try to discredit me sir.
#71
Posted 17 January 2014 - 07:31 AM
wwiiogre, on 23 December 2013 - 10:43 AM, said:
The point of this discussion is not winning and losing. The point is the matchmaker is not attempting to make a balanced game. It is not using some of the most important things that should be included when setting up a match. When one side has a 300 ton advantage, that is the same as having three extra Atlas mechs. Now, instead of the weight advantage, there would be no argument if in fact one side did have 3 xtra Atlas mechs drop instead of the tonnage advantage. \
Tonnage does not need to be exact, but should be less than 100 tons total in every match. Equipment value and values of mechs used should be close in every match, not equal but close. Say within 5 million cbills. That way you do not have trial mechs facing elite mechs. You have near equal tonnage and near equal mechs and equipment including modules. So far the above two would go a long way to leveling the playing field.
Then all you have left is player experience and skill and the ability to build an efficient killing machine in the mech lab. There is nothing to whine over, because the game started balanced. If you lose when this occurs you have nothing to blame but yourself and your teamwork. Which is what I am hoping PGI is trying to do.
But currently the matchmaker does not do this, it does not even attempt to do this because the above two things are not part of ELO. They are not even a small part of it. PGI has decided to go the route of everyone gets a ribbon for participating by manipulating win/loss records so nobody can brag about their amazing skills on the forum. Instead you get this horrible matchmaker that literally stacks most games for oneside to win. Giving them more 4mans, giving one side better mechs, giving one side more tonnage. All based on win/loss ELO of pilots in that mech.
Sorry, win/loss k/d is not how to rate a pilot especially when there is no tonnage matchmaking or mech/equipment/module matchmaking. How much skill do you have when you are in a 12 man with all assaults/heavies vs another 12 man in lights and mediums? None, sorry no skill involved when you beat up 5 year olds. Or literally come to the fist fight with a machinegun.
The entire point of this post is to point out PGI has utterly failed in matchmaking. The games and drops are not balanced, and therefore not fun whether winning or losing. a 12-0 is not fun. Its easy and stat padding. Its ok for people that like beating up on 5 year olds but for serious competitive players the current matchmaker is a joke and just like getting a ribbon at school for participating its worthless and really doesn't feel like anything important was accomplished when you win vs a team that had 4 trial mechs and you out weighed by 300 tons.
PGI needs to either completely scrap the current matchmaker or add the things I have been saying as well as many others. BV or Battle Value from tabletop works for equipment and mechs, but would need to be reworked since PGI has changed so many values. Plus BV has a rating for pilots based on table top skill levels which would not work in MWO. So a system for evaluating the skill of a pilot would be needed.
I do not think win/loss or even k/d work as a value for pilot skill. Why? Because currently those marks were earned under a system that does not evenly match mech vs mech on a ton or equipment basis. If every game a pilot had played up till now had been against other teams evenly tonned and evenly equipped then win/loss and k/d would be a valid part of an equation to rate a pilot. But that is not the case at the moment. The only thing I think that could be used at this point is to actually count total games played.
Meaning you rate pilots by experience, with experience being how many missions/drops they have made. You would then have rookie (still on first 25 missions); green (26-100), regular (100-500); veteran (500-1000) and elite (1000+) drops. This gives you an idea of how many times a player has played. You can do the same with mech based on type of equipment used. A mech with standard equipment, meaning no double heat sinks, no endo, no FF, no artemis, no tag, no ecm, no bap, no command module, no ams, no XL engine, no ER weapons, no pulse weapons, no streak, no gauss.
Give each piece of upgraded equipment a value, add that to the tonnage of the mech, given a number, do this with every mech in a 12 man and then attempt to find equal value opponents for them. So tonnage and equipment are easily figured, just some math and a discussion of PGI devs to decide what equals what. Then a simple rating system for pilots based on drops and not win/loss or k/d. Because if mechs are near equal and tonnage is near equal. I am ok facing a far superior pilot and eating a loss because his skill was better than I am facing a near certain loss because my team is out tonned and out gunned from the beginning.
I am not really asking for much, just a system that means matchmaking actually tries to make even matches. Not to manipulate win/loss records so every body has better self esteem. I know Canadians are nice and polite, but this game is about killing the other guy. Not feeling good because you participated. Give us near even equipment and let the bodies fall where they may.
Chris
And I have a winning record both as a PUG and as a Lawman. You will not have a balanced game with random players on your team every drop. with random builds and no cooperation. You have to many variable to try to balance and balance using Elo is n W/L not Personal ability, so if o want things made more even we need to find a better balancing metric. Also this is a game of attrition. Once numeric superiority is established, losing becomes quite difficult as should be in a combat game.
I know Pilots in the Law that can crush me in an Atlas while they are in a Jenner. Mass is not the tell all either in other words. Plain and simple there is not many metrics that can tell how good a player is in a group or alone.
As to beating up 5 year olds... We have honored a young lady near that age who used to beat up grown men (I have her Mech). So if you are bringing fists to a gun fight shame on you sir.
#72
Posted 19 January 2014 - 02:50 PM
wwiiogre, on 15 January 2014 - 05:23 PM, said:
Currently there is no way to judge pilot skill since matchmaker is broken, I propose using total number of drops in any mech as a base. Then total mech value as another base and total tonnage as another base. Thus you have three main ways to balance a match. Tonnage should never be more than 95 tons difference. Ever! Period! Total mech value including modules and upgraded weapons, etc should have no more than 20 million cbill advantage to one side or a nicely outfitted Atlas. So no team gains what is the equivalent of more than a single assault mech as an advantage. Number of drops could be used as well with people with less than 50 drops always in games with other players with 50 or less drops.
Note these ideas for matchmaker are based on solo drop or 2man drops. 3man and 4man drops should always face each other only. Meaning unless a solo player chooses to be placed as a solo in the 3man/4man que he would never face other teams. This means you get less rofl stomps because of tech/tonnage/premades. You will still get some rofl stomps but they will happen because of skill and teamwork and not because of having a huge advantage that teamwork and skill cannot usually overcome.
Just saying.
Chris
Chris, I still believe that damage done, per weight class, over multiple matches is the best metric to grade pilot skill for a BV based system. This will be the only way to fairly balance matches. Upon further review, I would have to agree that a medium company dropping against an assault company is grossly unbalanced. Your proposal for averaging number of total drops and averaging 2 thru 4 man drops with a 95 ton max disparity seems to be a reasonable immediate fix. I would like to see that run first as an improvement over the current method in an effort to find a more permanent fix. It may be the BV system is difficult to implement, but your proposal will be an immediate improvement in the short term. It is my hope that they implement something sooner rather than later. Kyle
#73
Posted 20 January 2014 - 03:13 AM
wwiiogre, on 19 December 2013 - 03:54 PM, said:
It always was an epic fail in its base concept Chris. If you'll just start randomly picking 24 players and randomly dividing them into 2 teams you won't notice any difference with what we have now. Only searching times will be like 2 seconds not 20 minutes.
#74
Posted 21 January 2014 - 06:22 AM
Otherwise you get players knowing that damage and kills and assists are worth more than anything else and then all you see are assaults and heavies with meta damage builds. Just like what is happening right now. There is no point in playing most of the light or medium mech builds at the moment. When you do, your team because of the current matchmaker has a good chance of being out tonned because you brought a light mech to an assault mech match.
Saw another almost 500 ton mismatch this week and many 300 to 400 ton mismatches. I am seeing so many 4man groups now dropping all assaults just so they can roflstomp the next wave of pugs.
I reiterate again, 4mans and 3mans should only face others like themselves and solo pugs that agree to drop with them. All other groups should be dropping separately. This means less chance of 12 solo players with 4 new players in trial mechs facing 3 4man teams. Which I faced again yesterday. This is ridiculous and should never happen, but is becoming more and more near normal. With one team having either 2 4man teams or a 4man and a 3man vs 12 solo players and the majority of the new players with trial mechs are not on the team that has several groups and then no trial mechs and oh yeah they get a 200-450 ton advantage since both groups dropped all assaults or assaults and heavies.
It is just so broken at the moment that when I do finally get a game that actually comes down to the last few players on each side it is so rare I have to mention it in game chat. I get one of those games about every 20-30 games dropped. I get the unbalanced games in over 50% of all drops sometimes as much as 75%. That should not be happening. Ever.
Chris
#75
Posted 21 January 2014 - 06:35 AM
Edited by Joseph Mallan, 21 January 2014 - 09:32 AM.
#76
Posted 21 January 2014 - 06:57 AM
#77
Posted 21 January 2014 - 07:00 AM
#78
Posted 21 January 2014 - 07:54 AM
Edited by Albuquerque, 21 January 2014 - 08:33 AM.
#79
Posted 21 January 2014 - 08:39 AM
Don't whine and [Redacted] about how it isn't realistic or based on lore. No one wants to play a game where the premade of four high elo players running four assaults against a bunch of new guys riding trial atlas. THIS IS NOT A FAIR GAME. Replace those four newbies with 'low skill' players and the story stays the same. It promotes a bad experience for players and pads the high ELO players to increase their queue times.
Get off the high horse and telling newbies to play as well as high elo players if you want tonnage/equipment/weight limits only.
That being said.
Matchmaker just needs an ELO reset. Tonnage Limits. And the loosened restrictions. ELO is a terrible system that is overused but it manages to quantify a very rough skill estimate which is better than having the random chance scenarios where the brand new player in a jenner vs Edmeister in his.
A good example is the Warthunder Matchmaker. It has a quantified battle rating for planes and it matches you into an appropriate tier. It still manages to mess up and set jets against mid era prop planes. You also have scenarios where the entire team is nothing but the FOTM planes. You have cases where queues take ten minutes and you get placed into a game with 4 other people and they are not on your team. Even if you managed to get a balanced game the skillful pilots becoming overwhelmingly apparent when the Mustang, FW or N1K2 on the other team has 5 kills.
Edited by Tichorius Davion, 21 January 2014 - 08:45 AM.
#80
Posted 21 January 2014 - 09:22 AM
Tichorius Davion, on 21 January 2014 - 08:39 AM, said:
Loosened restrictions are already here, as we've seen, and Russ's interview last week mentioned an Elo reset, probably coinciding with the Launch Module when tonnage limits will be in effect. It's all coming down the pipe...soon.
1 user(s) are reading this topic
0 members, 1 guests, 0 anonymous users






















