TrixxieTriss Posted August 21, 2020 Share Posted August 21, 2020 I know I would like to know, Common Bioware, it’s time to spill the beans so the players can test if it is setup properly. Link to comment Share on other sites More sharing options...
Raansu Posted August 21, 2020 Share Posted August 21, 2020 That's a horrible idea. They need to fix it for sure, but how its designed doesn't need to be public knowledge, unless you enjoy the idea of people exploiting it. Link to comment Share on other sites More sharing options...
abhaxus Posted August 21, 2020 Share Posted August 21, 2020 I think it's pretty obvious how it works in terms of building the team comps, but I would like to know what factors into the skill of the players. As far as I know, I was the first person to suggest that a hidden ELO be used to determine reg matchmaking. (way back in 2014) http://www.swtor.com/community/showpost.php?p=7821669&postcount=25 Example: 16 players in queue. Two groups of 4, 1 group of 2, 6 solo players. For the purpose of this I will name the players A1-4, B1-4, C1-2, S1-6. Hypothetically I will assume they just use the same ELO metric from ranked queue rather than some combination of MVP votes, Valor rating, total warzones played, etc. Ratings for each player: A1: 1400 A2: 1900 A3: 2100 A4: 1400 A(mean): 1700 B1: 1200 B2: 1500 B3: 1700 B4: 1900 B(mean): 1575 C1: 1400 C2: 1100 C(mean): 1250 S1: 2200 S2: 1600 S3: 1400 S4: 1200 S5: 1100 S6: 1000 I am pretty sure the matchmaker calculates the grouped players average ELO, then fills the teams in order of highest to lowest ELOs on a round robin. Pick order: 1) S1 (team 1) 2) Group A (team 2) 6) S3 (team 1) 7) Group B (Team 2) 11-16) S3, S4, Group C, S5, S6 Team 1 (mean): 1375 S1: 2200 S2: 1600 S3: 1200 C1: 1400 S4: 1200 C2: 1100 S5: 1100 S6: 1000 Team 2 (mean): 1637.5 A3: 2100 A2: 1900 B4: 1900 B3: 1700 B2: 1500 A1: 1400 A4: 1400 B1: 1200 As you can see from this example, team 1 got the highest skilled player, but team 2 has the next 3 top rated players, and everyone on the team is higher rated or equal to the bottom 4 on team 1. I believe the matchmaker works this way because I have seen countless matches where two highly skilled groups are matched on the same team while the other team has no premade groups (or at most a 2-3 person group of lower skilled players). The only bulletproof explanation for this is the matchmaker doesn't actually prioritize group vs group first before filling the teams. Had the matchmaker prioritized group vs group, the draft order would look like this: 1) A1-4 (Team 1) 5) B1-4 (Team 2) 9) C1-2 (Team 1) 11) S1 (Team 2) 12) S2 (Team 1) 13) S3 (Team 2) 14) S4 (Team 1) 15-16) S5, S6 (Team 2) Team 1 (1337.5, heh): A3: 2100 A2: 1900 S2: 1600 A1: 1400 A4: 1400 C1: 1400 S4: 1200 C2: 1100 Team 2 (1500): S1: 2200 B4: 1900 B3: 1700 B2: 1500 S3: 1400 B1: 1200 S5: 1100 S6: 1000 I realize looking at the post now that I should have made the scenario with 8 players who were below average and 8 where above, but I'm not going back to redo it. My theory holds anyway. The end result is skilled players who are queued by themselves will almost always end up on heavily disadvantaged teams if there are groups in the queue, even if the groups aren't necessarily stacked with highly skilled players. Link to comment Share on other sites More sharing options...
phalczen Posted August 21, 2020 Share Posted August 21, 2020 I am pretty sure the matchmaker calculates the grouped players average ELO, then fills the teams in order of highest to lowest ELOs on a round robin. I suspect there are many factors that go into matchmaking, and many differences between the way matchmaking occurs in GSF versus ground. However, GSF does not use the average score, or rating, of the players in the group but rather the highest. Now, there isn't an ELO as far as GSF is concerned, at least not one we are aware of on any sort of measurable level. Instead it uses a score based on: 1) the requisition spent on each ship currently in the ship bar 2) The number of games played across the LEGACY and then uses that score to rank those in the queue. It also prioritizes pops for pilots waiting in the queue longer. Don't forget that it also prioritizes groups over solos, meaning that it will always pop a match for two groups of four and eight solo players, before it pops for 16 solo players, even if the groups haven't been waiting as long. So, assuming that ground calculates some sort of score or rating for everyone, even if it isn't an ELO, I'm willing to bet it chooses the group member with the highest score and uses that as the score for everyone in the group for matchmaking purposes, just like GSF. Link to comment Share on other sites More sharing options...
Raansu Posted August 21, 2020 Share Posted August 21, 2020 I am pretty sure the matchmaker calculates the grouped players average ELO, then fills the teams in order of highest to lowest ELOs on a round robin. It does, and thats kind of the problem. Team composition is overall more important than average ELO. Even overwatch eventually had to accept this fact and they began to force team comp meta into their solo queue. I'm not saying swtor do it to that extreme because, to be frank, the pop is too small (though I'd absolutely love consistent proper match every time with a tank/healer/dps comp as those are way more fun than the random chaos of all dps), but I feel like they 100% should consider composition first, then rating. One team with an off healer and a off guard against a team with no off heals or off guard for example? The team with those options win rate just shot through the roof. Obviously off guard could be solved by doing the proper thing and just removing guard from dps already, but things like class stacking should never happen. Shouldn't be running into a team with 3 PT's on it for example. Link to comment Share on other sites More sharing options...
Pallais Posted August 21, 2020 Share Posted August 21, 2020 I know I would like to know, Common Bioware, it’s time to spill the beans so the players can test if it is setup properly. No. You would just be giving people ways to exploit the matchmaking. Link to comment Share on other sites More sharing options...
TrixxieTriss Posted August 22, 2020 Author Share Posted August 22, 2020 That's a horrible idea. They need to fix it for sure, but how its designed doesn't need to be public knowledge, unless you enjoy the idea of people exploiting it. I disagree. I think we need some transparency so we can test it because it’s never worked since they added it and I have zero confidence in a fix we cannot evaluate and test rigorously to make sure all the wrinkles are ironed out. Link to comment Share on other sites More sharing options...
TrixxieTriss Posted August 22, 2020 Author Share Posted August 22, 2020 (edited) No. You would just be giving people ways to exploit the matchmaking. How can you exploit matchmaking when you get locked out for 15 mins? And honestly, who is going to spend the time doing that in regs? Edited August 22, 2020 by TrixxieTriss Link to comment Share on other sites More sharing options...
TrixxieTriss Posted August 22, 2020 Author Share Posted August 22, 2020 It does, and thats kind of the problem. Team composition is overall more important than average ELO. Even overwatch eventually had to accept this fact and they began to force team comp meta into their solo queue. I'm not saying swtor do it to that extreme because, to be frank, the pop is too small (though I'd absolutely love consistent proper match every time with a tank/healer/dps comp as those are way more fun than the random chaos of all dps), but I feel like they 100% should consider composition first, then rating. One team with an off healer and a off guard against a team with no off heals or off guard for example? The team with those options win rate just shot through the roof. Obviously off guard could be solved by doing the proper thing and just removing guard from dps already, but things like class stacking should never happen. Shouldn't be running into a team with 3 PT's on it for example. Everything you said makes sense. I totally agree with every point. Link to comment Share on other sites More sharing options...
AocaVII Posted August 31, 2020 Share Posted August 31, 2020 (edited) Now this is comical: I q for ONE wz last night, I'm not sure I ever seen such a discrepancy and imbalance like this, maybe I have idk or someone has at some point: 75 kills on my team (qd solo dps) 4 kills for the other team. That is an élite premier matchmaking system, average death per player just under 10 (9.35) LOOOOOOOOOL Important question here is: who or which players accounted for 4 deaths on my team LMAO Edited August 31, 2020 by AocaVII Link to comment Share on other sites More sharing options...
TrixxieTriss Posted September 12, 2020 Author Share Posted September 12, 2020 A simple question to the devs if they don’t want to reveal the whole system. Is it based per character or per legacy? Or per class? Link to comment Share on other sites More sharing options...
AocaVII Posted September 12, 2020 Share Posted September 12, 2020 (edited) A simple question to the devs if they don’t want to reveal the whole system. Is it based per character or per legacy? Or per class? From what I can tell it has to be based on, wins, votes, medals PER character because I have a sage where I was in a exceptional big PvP guild. We never lost practically so my win to loss ratio and number of wins/numbers on that toon are: 3,394 wins 112k kills 78k medals 7 k mvp votes in 3 years. It was my main. I don't play it anymore. LOL SOOOO thing is.....when ever I got on this toon after they implemented it, I got the worst of the worst players possible, and would lose almost every match. My other toons, (probably near 20 I have pvped on all the classes in the game) I see slight discrepancies on toons I have won a lot on but no where near the numbers on that sage. I could be wrong but IT WAS very noticeable on that particular sage to the point where I didn't even want to play it. If that holds true, we have a system that rewards people farming dps and not winning LMAO; I'm pretty sure it's much like ranked where wins give you ello etc Edited September 12, 2020 by AocaVII Link to comment Share on other sites More sharing options...
TrixxieTriss Posted September 13, 2020 Author Share Posted September 13, 2020 From what I can tell it has to be based on, wins, votes, medals PER character because I have a sage where I was in a exceptional big PvP guild. We never lost practically so my win to loss ratio and number of wins/numbers on that toon are: 3,394 wins 112k kills 78k medals 7 k mvp votes in 3 years. It was my main. I don't play it anymore. LOL SOOOO thing is.....when ever I got on this toon after they implemented it, I got the worst of the worst players possible, and would lose almost every match. My other toons, (probably near 20 I have pvped on all the classes in the game) I see slight discrepancies on toons I have won a lot on but no where near the numbers on that sage. I could be wrong but IT WAS very noticeable on that particular sage to the point where I didn't even want to play it. If that holds true, we have a system that rewards people farming dps and not winning LMAO; I'm pretty sure it's much like ranked where wins give you ello etc That’s been my experience on a lot of mine too. But I’ve been lvling up some new Alts in lowbies and Mids and I’m starting to wonder if there is a legacy element to it as well because I can get on an Alt I’ve not pvpd with yet and it will still put me on a team of the worst vs good players. I could be pvping with the same small queue all night on my other Alt and winning (you get to know who is who). And the more you win, the harder it gets because the system starts to stack the odds more and more against you till you cant win because it’s like playing 1v4 because your team gets so bad. It feels like you basically have to lose as many as you’ve won till you start to get one other player that’s not the worst. I’m sure you know what mean. Anyway, I jump on a brand new Alt (one that hasn’t pvpd) and I will still get put with the bad guys as if there is a legacy element there that is still trying to make you carry the 3 worst players in the queue, This really becomes a problem when I jump on a class that isn’t my best or I’m trying to get better at or it’s a totally different spec to normal. Just because I’ve been skilful on “x” class, doesn’t automatically translate that I’m as good on “y” class. Now I could be completely wrong and this could just be coincidence or perception bias or what ever. I’d just like to know one way or the other if the stats are stacked against me before I start. Link to comment Share on other sites More sharing options...
RedDeadFord Posted September 13, 2020 Share Posted September 13, 2020 I don't care how it works. I just want it better. Link to comment Share on other sites More sharing options...
Sappharan Posted September 13, 2020 Share Posted September 13, 2020 I disagree. I think we need some transparency so we can test it because it’s never worked since they added it and I have zero confidence in a fix we cannot evaluate and test rigorously to make sure all the wrinkles are ironed out. It makes sense to know how the matchmaking works to me also. There are people who would attempt to exploit it as Raansu suggests, but people do that now. Knowing the matchmaking design would at least reveal how much is players' fault and how much is the poor matchmaking. Obviously it's never going to be completely one or the other, but the insight could be valuable. Link to comment Share on other sites More sharing options...
TrixxieTriss Posted September 14, 2020 Author Share Posted September 14, 2020 It makes sense to know how the matchmaking works to me also. There are people who would attempt to exploit it as Raansu suggests, but people do that now. Knowing the matchmaking design would at least reveal how much is players' fault and how much is the poor matchmaking. Obviously it's never going to be completely one or the other, but the insight could be valuable. And we don’t need to know all of it. Just enough to be able to understand how it is supposed to work so we know wether it is working or not. Link to comment Share on other sites More sharing options...
-Bob- Posted September 15, 2020 Share Posted September 15, 2020 And we don’t need to know all of it. Just enough to be able to understand how it is supposed to work so we know wether it is working or not. Would we still know for a fact whether or not it is working? I doubt it...if things are going the way we want then it's working, if they aren't going our way, then something must be broken. Everyones opinion will be confirmation biased. But that's just my opinion. Link to comment Share on other sites More sharing options...
Recommended Posts