Territory War Matchmaker

Replies

  • Brindle
    19 posts Member
    edited October 2018
    snarzenal wrote: »
    If you’re going to use an ELO based matchmaking system (which I would be very supportive of), do NOT tie rewards to ELO rating. If you do, you may find guilds disbanding and reforming at a rate that would put rats breeding habits to shame.

    “Oh well, we’ve lost three times in row now. Time to disband. Everyone on Discord/Line and ready to reset our ELO?”

    Definitely need to tie rewards to ELO.

    Reason for: ELO systems cause members to have about a 50% win rate once they reach an accurate ELO rating (excluding changes in their skill). If you DON'T tie rewards to ELO, then a guild with poor coordination, mods, teams, etc will get the same rewards as a highly functioning guild of the same GP. Guilds would have zero incentive to do well, because eventually their ELO would drop low enough to get 50% win rate.

    Counter to quoted reason against: The quote assumes a win at default ELO is better than the rewards for a loss at accurate ELO. The reward system should not be designed that way; for example, if default ELO is 1600 and a guild's ELO is 2000, then a loss at 2000 needs to be better than a win at 1600. The only type of guild that would benefit from this is one that is very much below average, and that's hard to do with all the near-dead guilds out there that struggle just to get enough TW members. A guild performing that poorly would not be organized enough to disband and reform. The default ELO's prize bracket can be calculated exactly by EA using hard data (that only they have access to) if they research & understand the ELO system, but I'd guess it should be around the current 5-10M prize bracket.

    Edit: To make it simple, could also just make the default bracket be the very bottom. No difference to established guilds, simplicity is almost always better, EA probably won't lose sleep over giving nearly empty guilds such a minor bump in rewards, and it erases any possibility of a guild thinking about disbanding for TW rewards.
  • @TheRudrick you can tell how many of the opposition signed up by how many spaces you need to fill in each zone. You had 48 sign up, which would result in 24 per zone. If it’s less than that, post in here and we can tell you how many signed up.
  • There are not that many guilds in the 175m+ range. If the algorithm is designed to avoid that they get set against each other too often it seems obvious that there has to be a Plan B type of rule that picks guilds from the 155-175m range.

    We were set against a 182m guild. We have 162m GP. We are 50, they are 47/48. No chance to win. Or this algorithm is extremely smart and we are very un-smart.
  • Oct 1st 2018 TW mismatch
    Wallet Warriors vs Mighty Chlorians
    226,674,622 vs 195,962,946 = 30,711,676 difference
    both guilds are 50/50cgouy5edg44n.png

  • Yeah, we’re 61m GP, they’re 95m GP, and it’s not even close.

    This has happened to us so many times I’m not really interested in telling you about it.
  • ... If it’s less than that, post in here and we can tell you how many signed up.

    @DarjeloSalas we had 18 spots per zone. it sounds like 48 / 24 * 18... = 36?

    such algorythm seems to be advantageous to a guild with less higher GP players that join TW over more players with less GP each. it would be so easy to abuse.
  • Catarax28
    184 posts Member
    edited October 2018
    x524qnlizqsy.png

    Date of TW: 3 october 2018
    Your Guild’s GP 73.4 M GP
    Opposing team’s GP: 115.4 M GP
    # of my guild members who joined: 45/50
    # of opposing guild members who joined: Not enough

    The reason I’m applying to this post is because TW is getting RIDICULOUS. Five times in a row (that’s all the TW’s since I joined my current guild) we were put against an opponent with MUCH higher GP then we. I’m not talking about 10 M GP differences here, but more like 30M - 40M. I don’t mind losing once or twice, I’m not a whiner, but I would like to play more FAIR matches!

    This is actually what happens in sports quite often, deliberately losing weight to fit into a lower weight class category and have a much better chance to win.

    Think about it, you’ll start winning 2 zeta’s in TW’s when you can put of an ACTIVE GP over 60M. You will always win 1 zeta if you lose and your active GP is 60M-119M GP and 2 zeta’s if you win in those same categories.
    A guild of 115 M GP put’s forward 25 members of 3.0M GP against a guild of 75 M GP with 50 members participating, averaging 1.5M GP, thus winning two zeta’s with two fingers up their nose, instead of fighting hard against an equal opponent and risk losing and getting just 1 zeta. They probably have some roulating system, to keep all players happy, by not participating every TW.

    I can’t blame them either, zeta’s are a sparsity, and this technique will get you 4 more zeta’s per month. Please CHANGE the SYSTEM EA! And make the TW’s fun again for us too!

    Thanx!
    Catarax
  • StrangeTalent
    8 posts Member
    edited October 2018
    Date of TW: 10/3/18
    Your Guild’s GP: 123m
    Opposing team’s GP: 142m
    # of my guild members who joined: 50
    # of opposing guild members who joined: 48

    300 more zetas.

    wntz3isxedj1.png
    qx59h1ugu7nw.png
  • Date of TW-10/2/18
    My Guild GP: 141 million GP
    Opponent’s GP: 172 million GP
    # of our players joined: 48
    # of opponent players: 44

    Up until this match up we had an 11 TW win streak going. Then we came across Italy Stikes Back, a guild trying to cheat and game the system. One of our guild members quit the game after the Territory War.

    a3nt4xcfppw8.png



    g6oyi5do9c0a.png


    w86couu1afqc.png
    07hmmdo2p42i.png
  • yohann269 wrote: »
    Oct 1st 2018 TW mismatch
    Wallet Warriors vs Mighty Chlorians
    226,674,622 vs 195,962,946 = 30,711,676 difference
    both guilds are 50/50cgouy5edg44n.png

    You are joking, right? If not: build a stronger guild. Make Chlorians great again!
  • CG_SBCrumb
    685 posts EA Community Manager
    edited October 2018
    Thanks for all the data points and suggestions everyone. I've sent this over to the team to review and we will try to determine why there are consistently reported mismatched battles.

    I've also asked the matchmaker designer to write a brief breakdown of how the system works and what maybe causing these issues.
  • CG_SBCrumb wrote: »
    Thanks for all the data points and suggestions everyone. I've sent this over to the team to review and we will try to determine why there are consistently reported mismatched battles.

    With all due respect, the reason is obvious to everyone who plays and almost certainly everyone at EA - Straight GP comparisons do not lead to competitive matches. I imagine it is still "under investigation" because changing it is not trivial?

    Straight GP claims that two 1.75M accounts are worth one 3.5M account. That is grossly inaccurate in TW; the 3.5M account is probably worth more like a dozen 1.75M accounts.
  • Brindle wrote: »
    CG_SBCrumb wrote: »
    Thanks for all the data points and suggestions everyone. I've sent this over to the team to review and we will try to determine why there are consistently reported mismatched battles.

    With all due respect, the reason is obvious to everyone who plays and almost certainly everyone at EA - Straight GP comparisons do not lead to competitive matches. I imagine it is still "under investigation" because changing it is not trivial?

    Straight GP claims that two 1.75M accounts are worth one 3.5M account. That is grossly inaccurate in TW; the 3.5M account is probably worth more like a dozen 1.75M accounts.

    Huh? As far as CG_Devs said the matchmaking is not GP based. And yet there are plenty of mismatches.
  • origDeathfly
    127 posts Member
    edited October 2018
    Brindle wrote: »
    snarzenal wrote: »
    If you’re going to use an ELO based matchmaking system (which I would be very supportive of), do NOT tie rewards to ELO rating. If you do, you may find guilds disbanding and reforming at a rate that would put rats breeding habits to shame.

    “Oh well, we’ve lost three times in row now. Time to disband. Everyone on Discord/Line and ready to reset our ELO?”

    Definitely need to tie rewards to ELO.

    Reason for: ELO systems cause members to have about a 50% win rate once they reach an accurate ELO rating (excluding changes in their skill). If you DON'T tie rewards to ELO, then a guild with poor coordination, mods, teams, etc will get the same rewards as a highly functioning guild of the same GP. Guilds would have zero incentive to do well, because eventually their ELO would drop low enough to get 50% win rate.

    Counter to quoted reason against: The quote assumes a win at default ELO is better than the rewards for a loss at accurate ELO. The reward system should not be designed that way; for example, if default ELO is 1600 and a guild's ELO is 2000, then a loss at 2000 needs to be better than a win at 1600. The only type of guild that would benefit from this is one that is very much below average, and that's hard to do with all the near-dead guilds out there that struggle just to get enough TW members. A guild performing that poorly would not be organized enough to disband and reform. The default ELO's prize bracket can be calculated exactly by EA using hard data (that only they have access to) if they research & understand the ELO system, but I'd guess it should be around the current 5-10M prize bracket.

    Edit: To make it simple, could also just make the default bracket be the very bottom. No difference to established guilds, simplicity is almost always better, EA probably won't lose sleep over giving nearly empty guilds such a minor bump in rewards, and it erases any possibility of a guild thinking about disbanding for TW rewards.

    Moving to an ELO system is probably the best (fairest) way to fix this system.... It would also create a persistent TW rank.

    The downside that I see: ELO has a long memory. This is ok as long as the "teams" don't change much and no one tries to game the system. A weak guild that gets an influx of stronger players would have an advantage corresponding to the amount of power gained. A strong guild that loses a good group of members could be at a disadvantage for multiple TWs. Higher level guilds could game the system by rotating through an alt guild or 2. Tying rewards to ELO rating would help alleviate some of this problem.

    Maybe the answer is to merge the current system with an ELO-based system. Mix the current match making system with an ELO system. This would provide stability from where we are but allow for a guild's record help define the next opponent (and possibly rewards)

    (This assumes a guild-based ELO system. There are also member-based ELO systems [see 538's CARMELO system] that may be the best way to do this)
  • Member5973 wrote: »
    yohann269 wrote: »
    Oct 1st 2018 TW mismatch
    Wallet Warriors vs Mighty Chlorians
    226,674,622 vs 195,962,946 = 30,711,676 difference
    both guilds are 50/50

    You are joking, right? If not: build a stronger guild. Make Chlorians great again!

    Why do you think we might be joking? We do have a stronger guild at 215M GP. They got an appropriate match. While some may consider 30M GP deficit "OK" at 190M level, let me assure you, it isn't. It was an extremely challenging war and the disparity was enough to warrant the post. It is not much different then facing the same deficit at a lower GP level.
  • DJBoboFett
    64 posts Member
    edited October 2018
    Hey @CG_SBCrumb

    I'm one of the officers of NSFTW Black (formerly NSC Black). We are the current global leader in terms of average arena ranking which I believe could be one of the reasons as to why we have to fight above our weight in most TWs.(https://swgoh.gg/g/) I can say on behalf of our whole guild that TW is our favourite game mode and the main reason some of us keep going with this game. We have always endorsed the challenge and taken pride in our ability to challenge stronger GP'd guilds. Hence our recent name change to NSFTW (Not Safe For Territory War).

    Somewhat ironically we have lost quite a few TWs since changing our name as we find ourselves increasingly put up against guilds where it becomes highly difficult to overcome the GP gap. We feel that the GP gap with our opponent has been growing as of recent and many members of our Tw-focused guild are quite disheartened by that fact.

    We have detailed information about every single TW fought by us to date and can provide this information upon request. Your awareness of these match-making issues, your hard work and commitment to rectify this situation is highly appreciated.
    Until then we can only keep our fingers crossed that you'll find a way to make TW great again. :)

    Please see below details outlining our last 6 match-ups. We have had 50/50 members registered for every single TW, not sure about our opposition.

    39th Territory War
    Opponent : StarForge Jedha
    Guild GP : 171,180,709 GP - 183,897,447 GP

    40th Territory War
    Opponent : {BHG} AAA Stars
    Guild GP : 175,665,030 GP - 185,061,382 GP

    41st Territory War
    Opponent : Order 66 501st Division
    Guild GP : 177,099,237 GP - 188,764,269 GP

    42nd Territory War
    Opponent : ANZGC Ataru
    Guild GP : 180,158,457 GP - 184,888,392 GP

    43rd Territory War
    Opponent : RedForce VAAPAD
    Guild GP : 177,495,697 GP - 186,155,164 GP

    44th Territory War
    Opponent : Ł1313 UniCoRns
    Guild GP : 183,911,467 GP - 198,476,747 GP


    Post edited by DJBoboFett on
  • The downside that I see: ELO has a long memory. This is ok as long as the "teams" don't change much and no one tries to game the system. A weak guild that gets an influx of stronger players would have an advantage corresponding to the amount of power gained. A strong guild that loses a good group of members could be at a disadvantage for multiple TWs. Higher level guilds could game the system by rotating through an alt guild or 2. Tying rewards to ELO rating would help alleviate some of this problem.

    Maybe the answer is to merge the current system with an ELO-based system. Mix the current match making system with an ELO system. This would provide stability from where we are but allow for a guild's record help define the next opponent (and possibly rewards)

    (This assumes a guild-based ELO system. There are also member-based ELO systems [see 538's CARMELO system] that may be the best way to do this)

    The "weak guild that gets an influx of stronger players" wouldn't be at an advantage from the rewards standpoint; they would probably win, but would be in a lower reward bracket. The real loser in this scenario is the team that gets steam-rolled by the previously weak guild. Agree that a mixed system is best for this; let rating have a (heavy) effect on matching, but still use whatever the current garbage system is.

    Also agree that tying rating to members is better than tying to guilds. Mitigates the 192,384,761,209,347,189,234 variants of guild hopping that could game the system.
  • 3lv1s
    7 posts Member
    edited October 2018
    Previous TW
    my guild : https://swgoh.gg/g/33839/hexagone-paddawins/ (56M)

    30/09
    opponent guild : https://swgoh.gg/g/4545/galactic-kights/ (116M)

    21/09
    opponent guild : https://swgoh.gg/g/12646/thegalacticorder/ (115M)

    18/09
    opponent guild : https://swgoh.gg/g/28017/jabbashutts2/ (70M)

    10/09
    opponent guild : https://swgoh.gg/g/16960/la-resistance/ (84M)


    it starts being annoying.
  • Gannon
    1619 posts Member
    I don't know if anyone has asked this yet, but can we just eliminate the "Preview Phase"?
    Wars seem to take too long, can't we combine the join period with the defense setup phase? Just automatically enter everyone like TB does? Anyone else think this would be better, faster, and fix some of the matchmaking issues?
  • We're a 126M PG Guild, and from the moment we passed the 100M we have never, Never, NEVER faced a guild with a lower PG than ours, being honest we have more wins than defeats, but is really demotivating that we always face someone with superior GP, every time the difference increases, started with 3-5M, then 10M and for the past couple of months we are always facing guilds with at least 20M GP difference, so obviously we are on a defeat streak which is totally demotivating our members, and last one got ridiculous with a delta of 30M, we faced "Unionize the Jawas" and we're "SW Halcones Galacticos"
    We haven't been able to start farming Traya, which will make it worst, with this delta on PG we will start facing Traya walls soon and will make it impossible to win.
    I hope someone is really looking at this and fix it soon! I can understand that sometimes the formula can go wrong, but I'm not exaggerating when I'm saying that from the 100M PG we haven't faced a lower PG Guild ever and the delta between our rivals and ours keeps getting bigger everytime, so something is really wrong here!
    Lets see what today's TW brings...
  • @CG_SBCrumb instead of thinking about overall GP why not look at arena ranks, and teams (like you guys say you are doing ) Please get this fixed, we have been talking about this long enough.

    GP is not even close to an accurate measure of a guilds strength.

    Date of TW: 10/5/18
    Guild GP: 156M (avg arena rank 95)
    Opposing 165 GP: (avg arena rank 8 !!!!!!)
    # of G12: 1,238
    # of opposing G12: 1,796 (500 advantage which is a 100 team advantage...c'mon)
    #of Traya: 6
    # of opposing traya: 50 (44 advantage)
    # of chewbacca above G9: 12
    # of opposing chewbacca above g9: 40 (31 advantage)

    Sucking the fun right out of the best content in the game (when the matchmaking works).
  • jkray622
    1636 posts Member
    Date of TW: 10/5-10/6
    Your Guild’s GP: 143M https://swgoh.gg/g/3263/the-republic/
    Opposing team’s GP: 155M https://swgoh.gg/g/27804/irrationally-confident/
    # of my guild members who joined: 50
    # of opposing guild members who joined: 48

    Other fun facts:
    Zetas: Our 1,096 to their 1,400
    G12 Toons: Our 964 vs their 1,429
    G11+ Toons: Our 1,806 vs their 2,768
    g11+ Trayas: Our 7 vs their 30

    All the other stats we reviewed - # of Jango, Chewie, Bastila, Qi'ra, etc were all pretty flat.

    However, they have 962 more G11+ toons than we do - that's enough to fill more than 8 complete zones with the extra toons. I'm really "excited" about this one...
  • jkray622 wrote: »
    Date of TW: 10/5-10/6
    Your Guild’s GP: 143M https://swgoh.gg/g/3263/the-republic/
    Opposing team’s GP: 155M https://swgoh.gg/g/27804/irrationally-confident/
    # of my guild members who joined: 50
    # of opposing guild members who joined: 48

    Other fun facts:
    Zetas: Our 1,096 to their 1,400
    G12 Toons: Our 964 vs their 1,429
    G11+ Toons: Our 1,806 vs their 2,768
    g11+ Trayas: Our 7 vs their 30

    All the other stats we reviewed - # of Jango, Chewie, Bastila, Qi'ra, etc were all pretty flat.

    However, they have 962 more G11+ toons than we do - that's enough to fill more than 8 complete zones with the extra toons. I'm really "excited" about this one...

    This is a proper way of looking at two guilds..imo @CG_SBCrumb At least if the two guilds have similar zetas, G11/12 characters, and arena rank then mods, setup, strategy, and zeta choices all matter.

    But when a guild is down hundreds of G11/12 characters, they don't stand a chance. The GP comparison is not even close to being an accurate way of measuring a guilds strength. It's more of a coincidence.



  • Hope this gets changed soon. Maybe some compensation Zeta’s would be in order 😉



    0acl9vf2ha6c.png
    a9oldvo01tkq.png
  • Oberonpm
    1 posts Member
    edited October 2018
    i0lyckgp8e4i.png
    vr1pzaoss6sh.png
    5i9n1nmnt4tf.png
    We are live long shoot 1st 124mil gp vs a German guild(I guess with 160mil gp) they are 2 reward teirs above us. Ever since you guys "fixed" the matchmaking system this has happened every. Single. Time. When it's reportes our guild leader gets the go isn't a fair estimate of guild strength.We have one only a handful of the wars since the change which makes it really hard to recruit new members to the guild. @CG_SBCrumb please fix this. It has soured the vast majority of our guild to tw seeing a 30 to 40 mil gp gap every war. 49 or 50 joined this time it puts us at the bottom of our reward teirs.
  • Date of TW-10/5/18
    My Guild GP: 119 million GP
    Opponent’s GP: 151 million GP
    # of our players joined: 48
    # of opponent players: 48?

    Their top player that is visible to us is 4.7m GP. 156 G12.

    https://swgoh.gg/g/9569/rabusutatoretsuku/

    ax5h9m2bfd73.png
  • aperingo wrote: »
    We're a 126M PG Guild, and from the moment we passed the 100M we have never, Never, NEVER faced a guild with a lower PG than ours, being honest we have more wins than defeats, but is really demotivating that we always face someone with superior GP, every time the difference increases, started with 3-5M, then 10M and for the past couple of months we are always facing guilds with at least 20M GP difference, so obviously we are on a defeat streak which is totally demotivating our members, and last one got ridiculous with a delta of 30M, we faced "Unionize the Jawas" and we're "SW Halcones Galacticos"
    We haven't been able to start farming Traya, which will make it worst, with this delta on PG we will start facing Traya walls soon and will make it impossible to win.
    I hope someone is really looking at this and fix it soon! I can understand that sometimes the formula can go wrong, but I'm not exaggerating when I'm saying that from the 100M PG we haven't faced a lower PG Guild ever and the delta between our rivals and ours keeps getting bigger everytime, so something is really wrong here!
    Lets see what today's TW brings...

    What a surprise, on today's TW, we're facing "Reborn Revanites" 146M guild vs our 127M guild.... (was 126M last week)
  • Current TW we are a 104M guild with 97.5M active (44/50 joined) vs a 128M guild (40/50). I can't remember the last time we weren't the lower GP guild. It is VERY frustrating. Basically these are the same stats for 99% of our TW in the last 6 months (Bet you can't find more than 1 in the last 12 TW where we had the higher GP). The closest match I can remember having is still about a 10M gap where we were the underdog. It is very difficult to match the quality of an opponent's roster when 20% of their guild have higher GPs (500k-1M more) than our highest GP member.

    Guild: Wampa Noms
  • We are a 55M PG guild and we face a 92M guild AGAIN IT IS BORING
Sign In or Register to comment.