+ Reply to Thread
Results 1 to 8 of 8
Like Tree3Likes
  • 1 Post By Nefelia
  • 1 Post By Faeruna
  • 1 Post By Nefelia

Thread: Improvement to the Matchmaking System...

  1. #1
    Ascendant
    Join Date
    Dec 2011
    Posts
    2,357

    Default Improvement to the Matchmaking System...

    ...since mixing seasoned veterans with green amateurs leads to frustration on both sides. I propose Trion use the current tools to improve the quality of the matches.

    Ocho mentioned that weighted ELO (ELO based on individual metrics from the players' in-match performance) still exists, even though it is not being used. I suggest that matches are not started until there are enough players queued up to form 10 man teams for two warfront matches (40 players in all). Once 40 people have queued up, they would be separated by weighted ELO: the 20 players with the higher weighted ELO scores would be placed into one match and sorted into teams by raw ELO, while those 20 with the lower weighted ELO would be placed into another match and sorted into teams.

    This would naturally result in slightly longer queue times, but should also result in a narrower range of skill/experience (competence) within individual matches. In my opinion - as well as the expressed opinion of many other PvP veterans - this would also result in much better matches and less friction between players at opposing ends of the competence spectrum.

    I am aware that Trion is not fond of long queue times, but I honestly feel this would not be an issue: with the introduction of minions, carnage quests, onslaughts (once the NT onslaughts get working properly), etc, even pure PvP players can find productive things to do during the minute it takes for the WF queue to pop.

    I feel this simple alteration to the current matchmaking system would result in a significant improvement of warfront matches with relatively little effort. I hope Trion takes the time to give this (or similar solutions) due consideration.

  2. #2
    Shadowlander Faeruna's Avatar
    Join Date
    Nov 2014
    Posts
    43

    Default

    Quote Originally Posted by Nefelia View Post
    ...since mixing seasoned veterans with green amateurs leads to frustration on both sides. I propose Trion use the current tools to improve the quality of the matches.

    Ocho mentioned that weighted ELO (ELO based on individual metrics from the players' in-match performance) still exists, even though it is not being used. I suggest that matches are not started until there are enough players queued up to form 10 man teams for two warfront matches (40 players in all). Once 40 people have queued up, they would be separated by weighted ELO: the 20 players with the higher weighted ELO scores would be placed into one match and sorted into teams by raw ELO, while those 20 with the lower weighted ELO would be placed into another match and sorted into teams.

    This would naturally result in slightly longer queue times, but should also result in a narrower range of skill/experience (competence) within individual matches. In my opinion - as well as the expressed opinion of many other PvP veterans - this would also result in much better matches and less friction between players at opposing ends of the competence spectrum.

    I am aware that Trion is not fond of long queue times, but I honestly feel this would not be an issue: with the introduction of minions, carnage quests, onslaughts (once the NT onslaughts get working properly), etc, even pure PvP players can find productive things to do during the minute it takes for the WF queue to pop.

    I feel this simple alteration to the current matchmaking system would result in a significant improvement of warfront matches with relatively little effort. I hope Trion takes the time to give this (or similar solutions) due consideration.
    I would be totally fine with longer queue times. It's a worthwhile price to pay for more balanced matchups - assuming ELO is a valid measure. I don't know what to say about your idea of using weighted and raw ELO nor do I have any idea about waiting for 40ppl. It would need testing imho. All I know is that I would gladly queue for 5-9 min a pop if the matches were mostly even and enjoyable.

  3. #3
    Ascendant
    Join Date
    Dec 2011
    Posts
    2,357

    Default

    Quote Originally Posted by Faeruna View Post
    I would be totally fine with longer queue times. It's a worthwhile price to pay for more balanced matchups - assuming ELO is a valid measure. I don't know what to say about your idea of using weighted and raw ELO nor do I have any idea about waiting for 40ppl. It would need testing imho. All I know is that I would gladly queue for 5-9 min a pop if the matches were mostly even and enjoyable.
    Weighted ELO were high for those who had lots of damage/healing done, and low for those who did not. Problems arose when high weighted-ELO players were placed with low weighted-ELO players to 'balance' the teams. I don't suggest a return to balancing teams by weighted-ELO, as that apparently did not work for various reasons. I'd be happy to settle for roughly grouping active/competent players together and inactive/incompetent players together. Balancing the respective matches can be done with raw ELO just as it is now.

    The competent players would have more competitive matches, while the less competent players would be able to increase their skills in a more newbie-friendly match (less incidents of them getting blown up in 2 seconds by well-oiled assist trains as they back-peddle or stand still)

    I set the benchmark at 40 people in order to have teams of 10 players on each team before the match begins. I've seen matches begin with a 3 to 5 disadvantage a time or three. Those tend to start off poorly, obviously, and eventual reinforcements are rarely thrilled to be placed in a team significantly behind in points/objectives. Starting matches with 10 each would make for better balance, and would also provide a buffer in case one or two players is lagging, momentarily AFK, or a complete potato that managed to slip through the cracks: one of ten players being useless is somewhat manageable, one in five or four is much less manageable.

  4. #4
    Shadowlander Faeruna's Avatar
    Join Date
    Nov 2014
    Posts
    43

    Default

    Quote Originally Posted by Nefelia View Post
    Weighted ELO were high for those who had lots of damage/healing done, and low for those who did not. Problems arose when high weighted-ELO players were placed with low weighted-ELO players to 'balance' the teams. I don't suggest a return to balancing teams by weighted-ELO, as that apparently did not work for various reasons. I'd be happy to settle for roughly grouping active/competent players together and inactive/incompetent players together. Balancing the respective matches can be done with raw ELO just as it is now.

    The competent players would have more competitive matches, while the less competent players would be able to increase their skills in a more newbie-friendly match (less incidents of them getting blown up in 2 seconds by well-oiled assist trains as they back-peddle or stand still)

    I set the benchmark at 40 people in order to have teams of 10 players on each team before the match begins. I've seen matches begin with a 3 to 5 disadvantage a time or three. Those tend to start off poorly, obviously, and eventual reinforcements are rarely thrilled to be placed in a team significantly behind in points/objectives. Starting matches with 10 each would make for better balance, and would also provide a buffer in case one or two players is lagging, momentarily AFK, or a complete potato that managed to slip through the cracks: one of ten players being useless is somewhat manageable, one in five or four is much less manageable.
    Ok nice explication.

    After considering your initial post and taking note of your emphasis on the following:
    1. Your 'fix' can be achieved using tools and mechanisms that Trion currently has in place.
    2. Your 'fix', if working as you feel it may (as expressed in your follow up post) would facilitate:
    -better balanced matchups
    -alleviate poor starts
    -all at the cost of an increased queue time (a price You and I are willing to pay)

    I feel that it is worth a shot as a temporary measure viz. temporary until such time as a valid metric is developed that actually correlates with player skill AND such a metric is incorporated with a matchmaking engine that has an excellent algorithm for creating balanced matches in a minimum queue time.

    Furthermore there are other issues wrt matchmaking that can also be addressed e.g. Pre-made vs solo queueing, team role composition etc etc.

    I wonder if a possible tweak to team composition could be to incorporate restrictions on role switching during a match? Perhaps such restrictions could hinge on the role one queued up with e.g.

    Queue as DPS: may not switch to Heal, may switch to Tank or Support and back again.
    Queue as Heal: may not switch to DPS, may switch to Tank or Support and back again.
    Queue as Support: may not switch to Heal, may switch to Tank or DPS and back again.
    Queue as Tank: may not switch to DPS, may switch to Heal or Support and back again.

    The idea here is to limit crossheals somewhat. Also to make it easier for the matchmaking algorithm to balance each team's composition. This could also lend itself to Role/Soul based ELO metrics giving tailored measures of player skill.

    Anyway, i like your idea as a hassle free way that Trion could improve matchmaking right now with the tools they have right now. I feel that it should be temporary until a better way is developed. I also feel that it should be tested to see how much longer the queue times would actually be.
    Last edited by Faeruna; 04-20-2015 at 01:43 AM.

  5. #5
    Plane Touched Fusscle's Avatar
    Join Date
    Sep 2014
    Posts
    273

    Default

    The Elo system needs to be completely redone. When it actually balances anything it's only by chance.
    Last edited by Fusscle; 04-20-2015 at 07:07 AM.
    Fusscle of Cheat Engine

  6. #6
    Ascendant
    Join Date
    Dec 2011
    Posts
    2,357

    Default

    Quote Originally Posted by Fusscle View Post
    The Elo system needs to be completely redone. When it actually balances anything it's only by chance.
    I don't disagree. As Faeruna states, there is likely a better long term fix, but that would require time and resources that Trion seems uninterested in devoting to PvP. The entire system can do with a serious re-examination with a focus on providing good, competitive matches. The current priorities seems to be making sure everyone gets a win, which is rather silly: in a random, well-balanced system, players would be getting close to a 50% win unless they were terrible to an unrealistic degree.

    In the mean time, I would like to nudge Trion towards using their tools to provide more competitive matches. I'm suggesting a fix that seems to me to be the easiest and least costly/timely solution to the current problem of pros being matched up with complete newbies in warfronts.

  7. #7
    Shadowlander Faeruna's Avatar
    Join Date
    Nov 2014
    Posts
    43

    Default

    Quote Originally Posted by Nefelia View Post
    In the mean time, I would like to nudge Trion towards using their tools to provide more competitive matches. I'm suggesting a fix that seems to me to be the easiest and least costly/timely solution to the current problem of pros being matched up with complete newbies in warfronts.
    This, exactly this.

  8. #8
    Telaran
    Join Date
    Feb 2015
    Posts
    69

    Default

    Yep we need something like this.

+ Reply to Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts