Forums / Community / Matchmaking Feedback & Discussion

[Locked] Matchmaking Feedback Update – April 16

OP ZaedynFel

  1. 1
  2. ...
  3. 18
  4. 19
  5. 20
  6. ...
  7. 21
Hey Josh, can you take a look at my Elimination history again? I placed Platinum 5, which feels really ridiculous to me... I think I should be ranked much lower... I know past seasons don't mean much in the new system, but I've always historically ranked lower in Breakout than most other lists.

Especially since 9 of my games, we either lost or I was carried. The other game, I can't remember if we won or lost, but I feel like I definitely pulled my weight in that match.

Thanks again for all you do Josh, and the extreme patience you have while dealing with the angry mobs.
I am in the same boat here. I finished my placement matches last night and was placed Plat 6.

I just don't see this being a Plat 6 performance. I know that it's partially about who I was matched up against, and my whole game history but still. I am kind of getting the impression that I am being placed higher than I should.
Some observations from the Trueskill 2 paper:
  • Despite the claim in this thread that TS2 is 75% accurate, vastly superior, etc.- most of those gains are for squads larger than the max in any ranked playlist (think Warzone, BTB). Squads less than or equal to 4 show minimal predictive improvement over TS1.
  • The incorporation of kill rate correlations means you have to stat, even in those frequent games where its 4v1 and the guy won't quit.
Just take note that when you are being told you aren't as good as you thought, you are being given global evidence to support that argument. It is possible the relevant sub-population data also matches those trends, but I do not see that in the paper and that data isn't otherwise available to us. What is clear in the TS2 paper is the biggest gains in outcome determination are from squads greater than 4, which does not apply to any of the ranked playlists.

I would also be interested to see the effect of gametype correlations, specifically the HWC playlist, where I presume high ranks in other modes does not track very well.
ZaedynFel wrote:
If all people do is slay the entire game they will lose and their skill will go down. The K/D side only informs the skill, it's still all about the win.

We also have analyzed the impact of assists, damage, and objectives already as inputs into the model. They don't improve its ability to predict the winner, so wouldn't add or improve the skill estimates.

They would improve our ability to predict your assists. But we don't define skill as "lots of assists" the ultimate definition of skill in the model is still being able to win, regardless of assists, kills, damage, etc.
When I was playing last night I was thinking about this post from you, and I started wondering about something. If it's still about the win am I right in understanding that things like our accuracy %, flag grabs, flag caps, and such don't also play a part?

Edit: I was looking back at some of what you have been saying over the past couple of days, and I came across this:

ZaedynFel wrote:
Wins and losses are irrelevant, it's about WHO you beat and WHO you lose to, and how well your perform in those matches.
This seems to contradict each other. Unless these are in reference to two different aspects. The "...it's all about the win." is concerning the skill prediction, and the "Wins and losses are irrelevant..." is about the CSR.
DaxSeven09 wrote:
ZaedynFel wrote:
Wins and losses are irrelevant, it's about WHO you beat and WHO you lose to, and how well your perform in those matches.
This seems to contradict each other. Unless these are in reference to two different aspects. The "...it's all about the win." is concerning the skill prediction, and the "Wins and losses are irrelevant..." is about the CSR.
You hit the nail on the head. Josh has pointed out games earlier in the thread where people's MMR have gone up even though they lost because they performed better than Trueskill2 predicted. So for skill determination (MMR), it is about who you kill, who kills you, and performance. CSR, your public ranking, is still following the win/loss system (if I understand what he's been saying).

This sounds like it could result in disparities between CSR and MMR, but according to Josh over time the two values should balance out.

Edit: I guess you hit the nail on the sharp end. The reverse of what you said seems to be the case (win/loss affects CSR, kills/deaths affects MMR).
Ever since this new TrueSkill2, feels the social went down the drain even more.

Here is an example of pretty much every game in Triple Teams I've been playing a lot lately.

- one top guy on each of the team. These get like 20+ KD

- the rest get rekd by these top guys, with KD anywhere between -6 to -20. Hardly can get any kill at all, get totally destroyed by those top guys.

- and yet, it's claimed "each team had 50% chance of winning", and "noobs get to win more often than before".

Now let me ask you this: do poor noobs who go down 1:20 actually care that their team won (because there was that top guy who was 36:5)?? Do you think the noobs now enjoy these social games more when they die a lot every time but "their team wins more often"?

I personally don't care for the win in social, but I do want games where I don't go down 1:20 every time, which seems to happen now a lot, but was much less so before TrueSkill2.

As for Triple Teams: it's basically now the battle between top guys in each team, with the rest being cannon fodder and filler.

And I haven't seen any difference TS2 made in Ranked. I still play against exactly same people in SWAT as I played against last half a year. And we got all same ranks as last season etc.

Disappointing.
Ever since this new TrueSkill2, feels the social went down the drain even more.

Here is an example of pretty much every game in Triple Teams I've been playing a lot lately.

- one top guy on each of the team. These get like 20+ KD

- the rest get rekd by these top guys, with KD anywhere between -6 to -20. Hardly can get any kill at all, get totally destroyed by those top guys.

- and yet, it's claimed "each team had 50% chance of winning", and "noobs get to win more often than before".

Now let me ask you this: do poor noobs who go down 1:20 actually care that their team won (because there was that top guy who was 36:5)?? Do you think the noobs now enjoy these social games more when they die a lot every time but "their team wins more often"?

I personally don't care for the win in social, but I do want games where I don't go down 1:20 every time, which seems to happen now a lot, but was much less so before TrueSkill2.

As for Triple Teams: it's basically now the battle between top guys in each team, with the rest being cannon fodder and filler.

And I haven't seen any difference TS2 made in Ranked. I still play against exactly same people in SWAT as I played against last half a year. And we got all same ranks as last season etc.

Disappointing.
If you want more evenly ranked teams you should play ranked. While social takes MMO ranks into consideration it's a wider window to form games. It's just soical. Its going to makes games based on the pool of available players so not every game is going to have evenly matched players.
I just went through something similar in my last 3 games of arena where i played diamond, platinum and onyx. How is that fair at all when i'm just gold and i wasn't even ranked. Needless to say i got crushed by those *bleeps* since i know most of them have no life. I have been saying it since this game was released that I want even and fair games that are close and no one dominating. IF the skill level of the other team is too good then they can wait! What's worse is many of those players have secondary accounts so they can play and get matched. I am so irked at this game...seriously makes me question whether I will want to buy 6 when it releases. I say focus on the community and making sure games are fair and don't focus on the 1% that are top level. Forget them for a second and focus on the rest.

That's all.
I just went through something similar in my last 3 games of arena where i played diamond, platinum and onyx. How is that fair at all when i'm just gold and i wasn't even ranked. Needless to say i got crushed by those *bleeps* since i know most of them have no life. I have been saying it since this game was released that I want even and fair games that are close and no one dominating. IF the skill level of the other team is too good then they can wait! What's worse is many of those players have secondary accounts so they can play and get matched. I am so irked at this game...seriously makes me question whether I will want to buy 6 when it releases. I say focus on the community and making sure games are fair and don't focus on the 1% that are top level. Forget them for a second and focus on the rest.

That's all.
You only played 3 games, and if I am not mistaken did you team up with a friend who was a Diamond? If so, then that would be the reason you were matched that way.

And even if that was not the case 3 games is not enough to give the new TS2 system a chance to give you balance matches. I have been having the opposite experience. I have been matching up with some good teams. Yeah there is the occasional -Yoink- stomp, but that's gonna happen no matter what.
This post has been edited by a moderator. Please do not post spam.
*Original post. Click at your own discretion.
Spoiler:
Show
DaxSeven09 wrote:
You only played 3 games, and if I am not mistaken did you team up with a friend who was a Diamond? If so, then that would be the reason you were matched that way.

And even if that was not the case 3 games is not enough to give the new TS2 system a chance to give you balance matches. I have been having the opposite experience. I have been matching up with some good teams. Yeah there is the occasional -Yoink- stomp, but that's gonna happen no matter what.
Also, we don't if any players quit during matching which can cause an imbalance in the teams.
hi,

will trueskill2 be applied to ranked matches?

I wrote a couple weeks ago about being constantly matched in unfair ranked games and not much has changed. I took your advice in changing my search preferences to focused but it feels the same as balanced.

my last match:

https://www.halowaypoint.com/en-ca/games/halo-5-guardians/xbox-one/mode/arena/matches/e5f86be8-727e-4c76-9808-53fdf1721bd0/players/pablo%20says%20hi?gameHistoryMatchIndex=0&gameHistoryGameModeFilter=All

red team had two champions and tow onyx players while we (blue team) had a platinum, two low level diamonds and an unranked.

could you please explain to me why the current system allows games like these to be played? Why doesnt the current system recognize this as unfair?

thanks,

Pablo
The same for me. The new system doesnt work well 👎
fix it!!! all my recent games were like
4 unranked vs 4 diamond
4 unranked vs 3 diamond 1 onix
4 unranked vs 2 diamond 1 onix 1 champion

DO YOU CALL THIS EQUILIBRATED?
The last system was better than this one
👎👎👎👎👎👎👎👎👎👎👎👎👎
I started my placement yesterday and finished today in SWAT. I played 3 games against the same team. 1 champion and 3 onyx. Then I play 1 game against 3 champions and an onyx. The games were not close. All 4 games ended 50-30. I was on the loosing side of the score. My placement was a diamond 3. The last two season I've ended in onyx. I don't mind playing against high rank players but when I'm experiencing a significant connection drop against the repeating team I feel this system favors a group of players over others. I am solo playing. As a solo player primarily I felt the previous rank system was more balanced.
https://abload.de/img/5-5-2018_11-38-04_pmxgo18.png

Because Matchmaking is fixed now and balanced. Not to mention that it was like 10 games in a row like this, and its not the first game, its 140something. Two of us even played against champions but we somehow won.
fix it!!! all my recent games were like
4 unranked vs 4 diamond
4 unranked vs 3 diamond 1 onix
4 unranked vs 2 diamond 1 onix 1 champion

DO YOU CALL THIS EQUILIBRATED?
Links to the games in question.
fix it!!! all my recent games were like
4 unranked vs 4 diamond
4 unranked vs 3 diamond 1 onix
4 unranked vs 2 diamond 1 onix 1 champion

DO YOU CALL THIS EQUILIBRATED?
The last system was better than this one
👎👎👎👎👎👎👎👎👎👎👎👎👎
Unranked people have ranks as well. They have their MMR present at all times (actual skill level), and they have their CSR rank present at all times (Bronze-Champion), which is simply hidden until the completion of 10 placement matches, but it is there regardless. Having unranked vs ranked matchups is perfectly normal at the start of a new season. Not everyone gets their ranks during the first week, some get them as late as a month into the season.

Matchups can be unfair because the population isn't that great, and can especially seem unfair during the placement matches, since you are being tested more during them versus after getting your rank. 'Tested more' means that you are matched with worse opponents (minority of the 10 games) and better opponents (majority of the 10 games) to see where your limits and placing in MMR and CSR are precisely located. Don't worry, you could lose all of your 10 placement games to Onyxes, but the system will find your skill by looking at how well you did against those Onyxes. I can personally vouch for the system since it found me the same rank in HCS playlist for my main account (SR152) and a brand new smurf account (SR1) with 10 games each.
So far in this new season I'm 22 wins & 42 losses. Across all ranked playlists. This isn't close to 50%. And I've been playing against players WAY higher in skill than me many times and getting dominated. (Champ 44 against a low plat/high gold player in FFA?) With TS 2.0 I've ranked the highest I've ever been in many playlists... but I've lost a higher percentage of games than ever before, and MANY by blowouts (looking at you Elimination playlist). I don't care if my rank is higher if I'm losing all the time. I just wanted competitive games more regularly than not! And, so far I unfortunately feel like the old TS system was better in that regard.
So far in this new season I'm 22 wins & 42 losses. Across all ranked playlists. This isn't close to 50%. And I've been playing against players WAY higher in skill than me many times and getting dominated. (Champ 44 against a low plat/high gold player in FFA?) With TS 2.0 I've ranked the highest I've ever been in many playlists... but I've lost a higher percentage of games than ever before, and MANY by blowouts (looking at you Elimination playlist). I don't care if my rank is higher if I'm losing all the time. I just wanted competitive games more regularly than not! And, so far I unfortunately feel like the old TS system was better in that regard.
I am in the exact same boat as you. But from what I can understand and how the numbers work out. Its a 50% win rate across the whole population. So while player A wins say 70 out of 100 games. Unfortunatley player B will win 30/100. Now say they play another 100 games, I believe the system should work itself out and those win % should be flipped to equal 50/50. But I doubt thats how it will work. Look at all the statistics you want. When you are only winning 10-20%, why would you want to keep playing, its very discouraging.
Your matchmaking is terrible. My team is either getting beat by a large margin or winning by a large margin. This does not seem very even to me. Would this apply to quick play as well?
Few days later, here we are. Deja Vu. I finish my elimination placement matches and to my surprise, I get ranked identical to my swat. To put this into context, I feel like I can compete at the higher end in swat, or at least hold my own against (high diamond/onyx), as far as breakout goes, it almost feels foreign to me. I am just awful at breakout, no idea what I am doing. Yet I yielded identical results. To me this make zero sense, regardless of what the statistics shoot out. They are wrong (in my opinion). I also doubt I am the only one getting wrongfully placed.

After 10 placement matches (50% win rate no where in site.. Hmmm)
Swat 0.973 K/D 20%
Breakout 0.905 30%

How does someone with that kind of winning % belong with the top 43% of players? (assuming each rank would host a 14% population roughly) Unless the whole halo community someone became better and no one can obtain the low ranks anymore..

I feel like K/D is almost irrelevant in ranked unless its just purely slayer. W/L should be the only thing the matters. Period. But thats for another day.

I will attach my last 4 breakout games, feel free to take a look. I am quite curious about a few things. Firstly, I matched the exact same 3 guys (1 onyx, 2 diamond), 4 games in a row (they were on my team once). They man handled me every game... Safe to assume they were in a party (maybe they weren't), which is fine. Id like to ask why I got matched up against a party but I feel like I will just get the same answer over & over again, so I won't.

https://www.halowaypoint.com/en-ca/games/halo-5-guardians/xbox-one/mode/arena/matches/f95d37cf-07b1-43dd-8bc8-c0a346f183b4/players/xxstrangesxx?gameHistoryMatchIndex=4&gameHistoryGameModeFilter=All

https://www.halowaypoint.com/en-ca/games/halo-5-guardians/xbox-one/mode/arena/matches/a284d54a-5c3d-458e-b176-9b521ec2a652/players/xxstrangesxx?gameHistoryMatchIndex=3&gameHistoryGameModeFilter=All

https://www.halowaypoint.com/en-ca/games/halo-5-guardians/xbox-one/mode/arena/matches/7bcc7ea8-616f-4ea9-a2f4-33815ef798cc/players/xxstrangesxx?gameHistoryMatchIndex=2&gameHistoryGameModeFilter=All

https://www.halowaypoint.com/en-ca/games/halo-5-guardians/xbox-one/mode/arena/matches/000ba021-e2d7-4e84-9c49-3a05cc047877/players/xxstrangesxx?gameHistoryMatchIndex=1&gameHistoryGameModeFilter=All

For context in the games above I had no one on my team that was in the previous ones, so the quote below is somewhat relevant to that of the above. I play on balanced on EST, so I feel finding other players shouldn't be a problem. Unless there were only 8 of us on playing breakout? (Obviously this is an exaggeration). If this is the case where all I face is teams and I get wrong ranking placements, then I will probably avoid rank playlists like the plague.

"If you were solo queuing yet matched against the same 8 players as before and had 4 of the same people on your team from the previous game, then that just sounds to me like those were the only people able to match with you at the time. Maybe due to regional settings and/or time of day; were you playing on Focused, Balanced, or Expanded?In general, the matchmaker will consider a full party team to be about 1 tier better than a team of the same average skill composed of solo queuers. This is to offset the advantage of playing in a full party. What this means is that the matchmaker can sometime match teams against solo players, but only if the average skill of the solo players is above the average skill of the team. According to the data collected by the matchmaker, this usually results in the team of solo players winning over the full party about 50% of the time, which is the desired result. This may not mean that you, personally, will win against teams 50% of the time, but rather the population at large.Is everyones rank "

In the games above I feel like skill levels couldn't be more opposite.

Moving on, is there a reason so many ranks are being inflated? If mine are, I'm sure others are too. I feel like this may have been answered so I apologize if this is a repetitive question. One can assume ranks are being inflated since the number of people in a single playlist isn't enough to support the multiple different ranks and this would create extended wait times or very uneven games. (this is a guess since the population seems to be the biggest kept secret in the world). So therefore matching players would become more difficult since the skill gap would be greater (only if there is population issues, I have no idea). But bump everyone up between plat & champ, essentially creating a bottleneck effect, and this creates a smaller gap in ranks and makes it easier to match people since everyones will be ranking much closer to each other. Even though the skill gap still remains the same......Those that get their ranks inflated don't stand a chance.

(completely off topic, sorry) Lastly, I know my internet and connection is great so I have no idea why recently there have been so many problems with lag. I always searched balanced. Im not too tech savvy but I have heard there have been server issues? (or just have seen the odd posts about server complaints) Or maybe there hasn't who knows. Maybe its just who I was playing at the time. ill just leave some links below. I feel like this is unacceptable for a game of Halo 5s standards. Its unfortunate since my game experience is ruined and those of my teammates since I become useless in the game when I can't even move around.

http://xboxdvr.com/gamer/xxstrangesxx/video/49431783
http://xboxdvr.com/gamer/xxstrangesxx/video/49431663
http://xboxdvr.com/gamer/xxstrangesxx/video/49431721

Apologies for the long post, but I feel if I don't share my honest opinion & feedback of what is going on, then these problems will never get better and history will just repeat itself.

Thanks
  1. 1
  2. ...
  3. 18
  4. 19
  5. 20
  6. ...
  7. 21