+ Reply to Thread
Page 1 of 11 1 2 3 4 5 ... LastLast
Results 1 to 15 of 153
Like Tree55Likes

  Click here to go to the first Rift Team post in this thread.   Thread: How ELO Works in RIFT

  1.   Click here to go to the next Rift Team post in this thread.   #1
    Rift Team Ocho's Avatar
    Join Date
    Mar 2014
    Posts
    1,933

    Default How ELO Works in RIFT

    There have been a number of questions lately about how ELO works in RIFT, so here is a brief rundown.

    ELO is a constantly adjusting rating based on a player's win/loss record in Warfronts. Every time you are on the winning side in a Warfront, your ELO goes up; every time you lose it goes down. The amount it goes up and down depends on the challenge your team faced. If a team's total ELO score was less than their opponents and they won then those winning underdogs would have their ELO score go up more than usual, and the losers would go down more than usual. However if the compared total ELO ratings expected one team to dominate, and they did dominate, then the final scores would be adjusted only slightly (the winners getting very little, and the losers losing very little).

    When the matchmaker is creating two teams for matches, it takes several factors to balance the two teams:
    • team sizes
    • relative total ELO rating
    • attempts to get the same number of healers
    • attempts to get the same number of DPS
    • if possible, have the same number of Pre-mades

    After studying the logs of hundreds of thousands of warfront outcomes, the biggest indicator in who was likely to win a match isn’t which side has the most healers, or DPS, or pre-mades... The most accurate predictor of which side will win was which side had the highest total participating ELO rating.

    One interesting thing that we noted was that participation is the most important factor in which side wins. Although the match maker can try to put together two teams with a relatively equal chance of winning, if someone doesn't show up to a match, or bails out, then that side is going to be handicapped until the matchmaker can try to find someone to take the place of the person who didn't participate.

    We used to have a slightly more complex "player rating" computation rather than standard ELO, but a flaw was discovered in the calculations last September that managed to affect two players, and so we threw out the extra bits and stuck with traditional ELO (wikipedia has a good article on it if you like math).

  2. #2
    RIFT Fan Site Operator Riane's Avatar
    Join Date
    Apr 2011
    Posts
    5,612

    Default

    So is win/loss really the only contributing factor to the ELO? Nothing else at all like damage/healing done or objectives?

  3. #3
    Shadowlander
    Join Date
    Mar 2015
    Posts
    33

    Default

    Quote Originally Posted by Ocho View Post
    When the matchmaker is creating two teams for matches, it takes several factors to balance the two teams:
    • team sizes
    • relative total ELO rating
    • attempts to get the same number of healers
    • attempts to get the same number of DPS
    • if possible, have the same number of Pre-mades

    IF you have a chance can you please elaborate on how the matchmaker determines who is healing and who is dps'ing (or tanking for that matter, although tanking was never mentioned). Does the matchmaker check what soul you are currently using when you hit the queue button or something? Curious because very often I will queue up, enter the match, then switch to healer if we don't have enough healers.

    Mostly just curious here, since the WF queue button doesn't have role based queueing and players can freely change to healer/dps during a match thus it eludes me how an algorithm could discriminate between who is going to play healer during the match and who will play DPS (or tank).

  4. #4
    Plane Touched Sinsalot's Avatar
    Join Date
    Mar 2011
    Posts
    257

    Default

    Tell us how the teams are matched in the queue.

    Do you put the two highest on opposing teams, then the next two on opposing teams? Do you just start adding people to a team until their ELO is higher and then just start adding people to the other team? How do you determine who is on whos team?

  5. #5
    Plane Walker
    Join Date
    Feb 2014
    Posts
    403

    Default

    Heh so this is why my ELO is rubbish
    Strikes@Typhiria Impossibruu@Typhiria Huhuhuh@Laeyths Strikez@Deepwood Primes@Deepwood

  6. #6
    Shadowlander
    Join Date
    Mar 2013
    Posts
    26

    Default

    So based on discussion in HelloKitty chat channel earlier, these were the questions that came up...
    Are they averaging out the ELO when you queue in a Premade or in a duo?

    How far back will your ELO be calculated?

    Is the ELO ever going to be reset?

  7. #7
    Rift Chaser ShayolGhul88's Avatar
    Join Date
    Jul 2011
    Posts
    312

    Default

    Nice explanation but taking this into consideration I que solo 99% of the time and yesterday I literally was on the losing team about 10 WFs and mostly being dominated haha.
    Pls respond

  8. #8
    Ascendant spaceboots's Avatar
    Join Date
    Mar 2011
    Posts
    5,760

    Default

    I figured out the elo foruma

    (enemies rating total)+400*(number healers h(n times right button clicked) - loses)/games

    Also im tried of being in queues with your pvers just wanting to snipe a win.


    ELO doesn't work in rift, because ELO in rift WFs is a team game, and the teams you put with negatively impact my score and my match making.
    Scrap your broken system, or make it applicable to the premade only queue where it will work right.

    The game isn't fun and it's because this broken thing.
    Last edited by spaceboots; 04-06-2015 at 06:28 PM.

  9. #9
    Shadowlander
    Join Date
    Jan 2015
    Posts
    41

    Default

    Quote Originally Posted by Ocho View Post
    .
    When the matchmaker is creating two teams for matches, it takes several factors to balance the two teams:
    • team sizes
    • relative total ELO rating
    • attempts to get the same number of healers
    • attempts to get the same number of DPS
    • if possible, have the same number of Pre-mades

    After studying the logs of hundreds of thousands of warfront outcomes, the biggest indicator in who was likely to win a match isn’t which side has the most healers, or DPS, or pre-mades... The most accurate predictor of which side will win was which side had the highest total participating ELO rating.

    One interesting thing that we noted was that participation is the most important factor in which side wins. Although the match maker can try to put together two teams with a relatively equal chance of winning, if someone doesn't show up to a match, or bails out, then that side is going to be handicapped until the matchmaker can try to find someone to take the place of the person who didn't participate.
    This part is so contradicting. The biggest indicator of winning is participating elo rating. But there is no factor of this in the matchmaker's criteria. If there are not enough players with high participating elo to accept and join the match, the match should be canceled (or else I will be waiting here for more people crying about broken matchmaking)

    Maybe this is not be related (but still is a good read i hope) if someone is not familiar with matchmaking rating and doesnt like math. Dota2 matchmaking concept and method (http://de.dota2.com/2013/12/matchmaking-2/) - Brotip: After one year of improving this matchmaker, people still say matchmaker is broken in dota2. And I will wait here a couple months for this exact same drama

  10. #10
    Shadowlander
    Join Date
    Oct 2013
    Posts
    25

    Default

    Quote Originally Posted by Ocho View Post
    There have been a number of questions lately about how ELO works in RIFT, so here is a brief rundown.

    ELO is a constantly adjusting rating based on a player's win/loss record in Warfronts. Every time you are on the winning side in a Warfront, your ELO goes up; every time you lose it goes down. The amount it goes up and down depends on the challenge your team faced. If a team's total ELO score was less than their opponents and they won then those winning underdogs would have their ELO score go up more than usual, and the losers would go down more than usual. However if the compared total ELO ratings expected one team to dominate, and they did dominate, then the final scores would be adjusted only slightly (the winners getting very little, and the losers losing very little).

    When the matchmaker is creating two teams for matches, it takes several factors to balance the two teams:
    • team sizes
    • relative total ELO rating
    • attempts to get the same number of healers
    • attempts to get the same number of DPS
    • if possible, have the same number of Pre-mades

    After studying the logs of hundreds of thousands of warfront outcomes, the biggest indicator in who was likely to win a match isn’t which side has the most healers, or DPS, or pre-mades... The most accurate predictor of which side will win was which side had the highest total participating ELO rating.

    One interesting thing that we noted was that participation is the most important factor in which side wins. Although the match maker can try to put together two teams with a relatively equal chance of winning, if someone doesn't show up to a match, or bails out, then that side is going to be handicapped until the matchmaker can try to find someone to take the place of the person who didn't participate.

    We used to have a slightly more complex "player rating" computation rather than standard ELO, but a flaw was discovered in the calculations last September that managed to affect two players, and so we threw out the extra bits and stuck with traditional ELO (wikipedia has a good article on it if you like math).
    Stop grouping scrubs with high rated players and try to get two teams of similar rating. it's pretty dumb when I'm on a team of obviously new 65s who haven't pvpd much and the team loses badly. Also your heal matching is beyond broken lol
    Last edited by Definitleynottechfall; 04-06-2015 at 07:19 PM.

  11. #11
    Plane Touched BigTerj's Avatar
    Join Date
    Aug 2012
    Posts
    286

    Default

    Ocho,

    Thanks for the clarification of how the system currently works, it will be nice to end all the speculation threads.

    I am a little bummed that my ELO score is solely based on the performance of the team instead of individual skill. I do however understand that over time the good players will filter to the top of the ELO score chart and the bad players will filter down.

    What I do not like though is that the AFKers and quitters (people who sit at respawn because they think the match is over) still get their ELO boosted for victories that they did not help win, which is why I would rather see more personal metrics taken into account. Like you said, people who don't participate ruin the balance of the match. There needs to be a better way to deal with AFK/quitters.
    Bigterj - <Grievance> - Wolfsbane

  12. #12
    Ascendant Violacea's Avatar
    Join Date
    May 2011
    Posts
    12,128

    Default

    Quote Originally Posted by Definitleynottechfall View Post
    Stop grouping scrubs with high rated players and try to get two teams of similar rating. it's pretty dumb when I'm on a team of obviously new 65s who haven't pvpd much and the team loses badly. Also your heal matching is beyond broken lol
    The system would work but what wrenches it is that a more critical class on the other team could be the higher rated player among that team. Maybe all the low players on one team are the healers. Opposed to a basic FPS where that would sort of work because everyone contributes in an equal fashion to the win more so, same roles.

    And you can't make a heal match making system without role locking. I am pretty sure it just looks at frequency the person has healed or been a healer in WF's, and makes a gamble hoping they play heals. Everyone has various roles on their hotbar that they can change with a click. I am not sure what you expectation would be to induce forced circumstances in lieu with free will. A higher frequency is no promise at all. But it is a better gamble to get someone on the 'heal' spot on your team who has actually demonstrated that they heal perhaps 40% of the time in WF's compared to someone who maybe has healed 5% of the time in WFs.
    Last edited by Violacea; 04-06-2015 at 07:59 PM.
    New round up of some high rank matches
    Chun-Li*E.Honda*Evil Ryu
    Abel*Ibuki*Dhalsim
    Gouken*Zangief*Ryu

  13. #13
    Ascendant
    Join Date
    Dec 2011
    Posts
    2,357

    Default

    Quote Originally Posted by Ocho View Post
    ELO is a constantly adjusting rating based on a player's win/loss record in Warfronts. Every time you are on the winning side in a Warfront, your ELO goes up; every time you lose it goes down. The amount it goes up and down depends on the challenge your team faced. If a team's total ELO score was less than their opponents and they won then those winning underdogs would have their ELO score go up more than usual, and the losers would go down more than usual. However if the compared total ELO ratings expected one team to dominate, and they did dominate, then the final scores would be adjusted only slightly (the winners getting very little, and the losers losing very little).
    That certainly explains why the matchmaking is so unreliable. ELO is not a reflection of player skill. Instead, it is a measure of luck, time spent on a good PM, and a very small measure of a player's ability to pull of a win despite being teamed with potatoes.

    I am beyond disappointed with this bare-bones ELO system.

    It does certainly explain why those who spend a great deal of time in premeds have such a miserable time when they solo-queue: their ELO is artificially inflated by their easy PM wins so they get extremely bad teams when they solo-queue. Trion has basically trapped these people into their PMs, unless they wish to go on a losing spree long enough to correct their ELO.

    Thank you for the information, Ocho. It gives us an indication of what we should be discussing in order to improve the ELO system and match-making in general. We could use some further clarification though: what was the flaw spotted in September that led to the scrapping of the more complex ELO rating?

  14. #14
    Shadowlander
    Join Date
    Oct 2013
    Posts
    25

    Default

    Quote Originally Posted by Violacea View Post
    The system would work but what wrenches it is that a more critical class on the other team could be the higher rated player among that team. Maybe all the low players on one team are the healers. Opposed to a basic FPS where that would sort of work because everyone contributes in an equal fashion to the win more so, same roles.

    And you can't make a heal match making system without role locking. I am pretty sure it just looks at frequency the person has healed or been a healer in WF's, and makes a gamble hoping they play heals. Everyone has various roles on their hotbar that they can change with a click. I am not sure what you expectation would be to induce forced circumstances in lieu with free will. A higher frequency is no promise at all. But it is a better gamble to get someone on the 'heal' spot on your team who has actually demonstrated that they heal perhaps 40% of the time in WF's compared to someone who maybe has healed 5% of the time in WFs.
    nah system doesnt work in any way. not separating premades and pugs you automatically have easy way to increase ELO by running in PM and facing pugs majority of the time. siNow, a bad player can raise there ELO to high levels since there is no separation between premade and solo ELO. Now when that player solo queues they screw it up for the rest of the team because they are high rated but bad.

    as long as the system does not increase the strength of your opponent as your ELO increases it will forever be useless, since basically ever match is just an average of crappy to high ELOs on both teams so it's basically just the same average rank more or less every game, and all reliant on luck for most part with some skill making a factor to a certain point. GW2 is a good example. in solo queue when you start out you play vs people who are bad and you can win easily then as you win you can clearly see better opponents both on your team and against you. eventually you start seeing people as good as you or better and it levels out. if there are 100 people in the random queue the top 20-30 ELOs should be put into one map and so on. instead it sounds like it just takes the first 20 splits them up to try to make an equal average on both sides, which is dumb.

    also your ELO should not go up if you are premade and beat a team with no equal 5 man premade

    also one thing they could do to make this better is make smaller maps, 5v5, 10v10 max where player skill is more of a factor
    Last edited by Definitleynottechfall; 04-06-2015 at 09:18 PM.

  15. #15
    Ascendant Ianto Jones's Avatar
    Join Date
    Feb 2011
    Posts
    6,018

    Default

    Quote Originally Posted by Definitleynottechfall View Post
    Stop grouping scrubs with high rated players and try to get two teams of similar rating. it's pretty dumb when I'm on a team of obviously new 65s who haven't pvpd much and the team loses badly. Also your heal matching is beyond broken lol
    This is what I was trying to tell you in the other thread: I think the system is theoretically supposed to work this way, but it can't because there aren't enough players with high-ish rating to fill teams. The vast majority are so bad and people who aren't bad are so few that it's impossible for it to make matches like this.

+ Reply to Thread
Page 1 of 11 1 2 3 4 5 ... LastLast

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