Forums / Community / Matchmaking Feedback & Discussion

[Locked] MATCHMAKING FEEDBACK UPDATE – FEB 28

OP ZaedynFel

  1. 1
  2. ...
  3. 2
  4. 3
  5. 4
  6. ...
  7. 6
Most of the things I thought of here were mentioned by other posters, so i just want to say that its a real breath of fresh air to get this kind of communication and honesty from you. Thanks.
I DSG I wrote:
So does season reset tomorrow and will breakout be back in this season ( the real thing )
Season resets March 9th to coincide with the maintenance update.
Ahh ok, thanks
ZaedynFel wrote:
EBLspartan wrote:
I know it doesn't fit exactly with the points you raise every week, but I think it fits with the matchmaking argument in general: is there a single chance to see selectable data centers in the future (most likely H6 I suppose)?
There's always a chance, but it's not clear to me that selectable data centers is the way to go. What decision do you need to make that can't be detected automatically?

Showing the DC exposes a pretty low level decision all players, many of whom will choose that data center incorrectly, or feel like they should mess with a setting that's more complex than it looks.

I personally prefer a compromise more on the lines of, "automatically pick the best data center for the party leader"

That way, we can use all the great data center health tech we have on our side to pick a great data center, and still give players control over who in their party that data center is based on.

So if you play with a friend in a different country, you know how to pick who gets the better experience.
Thanks for the answer, really appreciated.

Of course there should be some compromises, but I really think that showing the actual connected data center and the ping towards it can be a real act of transparency.
I mean: there are people, having instruments like netduma routers or simply intercepting packets with their PCs, claiming that H5 in Europe is currently using only the 2 original azure datacenters (Dublin, Amsterdam) while the past summer another 4 (2 Germany and 2 England if I remember correctly) where "released".
One of them showed me recently a report of this from his netduma router, confirming this behavior.

For the automatic/manual data center selection matter: systems like titanfall's and gow's are already using an hybrid where the game itself chooses for the player the best server/region at first main menu access, allowing the player to see ping but also to change it manually in any moment.
In a party scenario your thoughts should be still valid even in with this kind of system.

Another matter is adjust all of this to work with a skilled based system without exploding search times, but games like gow are already doing this,at least in their ranked playlists, fairly well.
I DSG I wrote:
So does season reset tomorrow and will breakout be back in this season ( the real thing )
Season resets March 9th to coincide with the maintenance update.
where did u read it?
Please just bring back the 1-50 ranking system. Much more appealing than these tiers.
Agreed. Halo 2's ranking system is easily the best out of all the Halo games, and it isn't even close.
Awesome thread thanx for the info.
I feel like the ranking system is too focused on teaming up and grinding. I'm currently gold but platinum and low daimond players I match don't always feel deserving of their rank and aren't all that challenging.

With the hidden ranking system does it also get a reset I used to play WZA 9/10 and I feel I got quite good at it at one point but eventually it became unplayable because of too much latency and lag. I stopped playing it for months and it only slightly got better but is still unbearable to play.
ZaedynFel wrote:
RyInfinity wrote:
When is the soonest we could be seeing actual alterations to Halo 5's CSR/Matchmaking? Obviously a lot of this is theory crafting for future titles but we've had ~10 seasons with little to no changes so we are pretty hungry for some improvements.
We had one CSR change a couple of weeks ago, and we're strongly considering decreasing the amount of games you need to reach your target CSR for next season.
A risk I see with this is that if you have an outlier game (good or bad) it can have a greater impact on your rank than it should have. Remember there are some connection quality issues here too that can have a heavy impact on some games and thus also impacting rank too much if its based on a (too) limited set of games.
LITTLE TE wrote:
Please just bring back the 1-50 ranking system. Much more appealing than these tiers.
Agreed. Halo 2's ranking system is easily the best out of all the Halo games, and it isn't even close.
Word
ZaedynFel wrote:
EBLspartan wrote:
I know it doesn't fit exactly with the points you raise every week, but I think it fits with the matchmaking argument in general: is there a single chance to see selectable data centers in the future (most likely H6 I suppose)?
There's always a chance, but it's not clear to me that selectable data centers is the way to go. What decision do you need to make that can't be detected automatically?

Showing the DC exposes a pretty low level decision all players, many of whom will choose that data center incorrectly, or feel like they should mess with a setting that's more complex than it looks.

I personally prefer a compromise more on the lines of, "automatically pick the best data center for the party leader"

That way, we can use all the great data center health tech we have on our side to pick a great data center, and still give players control over who in their party that data center is based on.

So if you play with a friend in a different country, you know how to pick who gets the better experience.
Appreciate your openminded posts here Josh - a much needed breath of fresh air! :)

But I have to somewhat disagree with you here even though I see you have a point about at what level a player should be able to make his/her own selections/preferences. In all honesty my trust in these "highly advanced automated optimizing systems" is very limited. Why? Based on experience. Here is my suggestion if we cannot be allowed to pick the datacenters ourselves:

  • Implement more datacenters in more geographic regions to offer optimum ping. There are more azure servers available and other developers are using a mix of azure and other data centers to provide the best possible user experience.
  • If I chose "focused" search, I never ever want to play against someone using expanded (or balanced for that matter) search to get into the same game. It is a well known fact by now that a single player from far away (with high ping) will have an impact on that game for everyone else as well. Several forum threads on this related to "heavy aim". A focused game should be just that, focused search for all players in it.
  • Games played using focused search should also have a minimum requirement on the players internet connection. Several variables go into determining what is a quality internet connection but at least the most basic bandwidth requirements should be met or one should not be allowed into a focused search game. It ruins the experience for all others.
Focused search should guarantee a game with only local players with good connections and on a geographically close server. It should be the ping connoisseurs search preference!

Today matchmaking is unfair as some who are fortunate enough to live close to a data center most often get a ping advantage to those who live farther away who are always at a disadvantage (which was not the case with P2P as it would not be a constant). In a competitive fps game like H5 this becomes very obvious and very critical.

BR
Unsung Hero
ZaedynFel wrote:
EBLspartan wrote:
I know it doesn't fit exactly with the points you raise every week, but I think it fits with the matchmaking argument in general: is there a single chance to see selectable data centers in the future (most likely H6 I suppose)?
There's always a chance, but it's not clear to me that selectable data centers is the way to go. What decision do you need to make that can't be detected automatically?

Showing the DC exposes a pretty low level decision all players, many of whom will choose that data center incorrectly, or feel like they should mess with a setting that's more complex than it looks.

I personally prefer a compromise more on the lines of, "automatically pick the best data center for the party leader"

That way, we can use all the great data center health tech we have on our side to pick a great data center, and still give players control over who in their party that data center is based on.

So if you play with a friend in a different country, you know how to pick who gets the better experience.
Appreciate you openminded posts here Josh - a much needed breath of fresh air! :)

But I have to somewhat disagree with you here even though I see you have a point about at what level a player should be able to make his/her own selections/preferences. In all honesty my trust in these "highly advanced automated optimizing systems" is very limited. Why? Based on experience. Here is my suggestion if we cannot be allowed to pick the datacenters ourselves. Anyways, below are my suggestions:

  • Implement more datacenters in more geographic regions to offer optimum ping. There are more azure servers available and other developers are using a mix of azure and other data centers to provide the best possible user experience.
  • If I chose "focused" search, I never ever want to play against someone using expanded (or balanced for that matter) search to get into the same game. It is a well known fact by now that a single player from far away (with high ping) will have an impact on that game for everyone else as well. Several forum threads on this related to "heavy aim". A focused game should be just that, focused search for all players in it.
  • Games played using focused search should also have a minimum requirement on the players internet connection. Several variables go into determining what is a quality internet connection but at least the most basic bandwidth requirements should be met or one should not be allowed into a focused search game. It ruins the experience for all others.
Focused search should guarantee a game with only local players with good connections and on a geographically close server. It should be the ping connoisseurs search preference!

Today matchmaking is unfair as some who are fortunate enough to live close to a data center most often get a ping advantage to those who live farther away who are always at a disadvantage (which was not the case with P2P as it would not be a constant). In a competitive fps game like H5 this becomes very obvious and very critical.

BR
Unsung Hero
100% quote
ZaedynFel wrote:
That said, I do also see a future where instead of directly showing the MMR that made the match, we could show some summarized version of your opponents' strength to give you a feel for why you were matched with them. For example, we could tell you that even though your opponents all look Diamond, they just played like Gold players, which matches up nicely with you being Gold.
Matches are better towards the season and when CSR is closer to MMRCSR has no influence on matchmaking at all. The matches you get throughout the season are exactly the same as they are at the end, unless you have actually change[d] as a player, thus moving your MMR. Even though you sometimes see a different range of CSRs in your match, the actual skills are really close.
I don't understand what you are trying to say or exactly the system used to rank in halo 5 from what you have said, would appreciate if you could clear some stuff up for me that has me scratching my head.
If my opponent is Diamond and I am a Gold, surely I should either be matched with ranks one above or below (eg. silver/platinum) not someone who in theory should be a 'better' player?
If we are of the same skill, why would they be ranked Diamond and me Gold? They either need to be going against tougher opponents so there rank falls enough to their skill level again or people who are Gold should be ranked higher.

If CSR has no impact on matchmaking why does it exist? Surely CSR should be the matchmaking system not something that is invisible so it doesn't feels as if your matchmaking system is broke? It makes me wonder why the CSR exists, its only purpose is to inflate ego's.
I find it unfair for players whose MMR is the same as people with a higher CSR, just seems a bit bonkers how someone who is the same as a Diamond is not a Diamond.
In theory Diamonds should play against Diamonds, Platinum's against Platinum's, I know the player base may not be as diverse and is why earlier I suggested their should be rules where you can play one above or below your level. It is just very frustrating as a player going against those with a higher CSR who in my opinion if I'm playing against them and winning do not deserve their CSR to be as high as it is. If I feel I'm losing my matches to unfair matchmaking I'm less inclined to play.
This is especially true considering the ridiculous grind for players of lower levels to reach higher CSR. It isn't fair for people who are gold who have to grind for much longer to go against diamond despite MMR correctly matching them.
I think the fact that MMR and CSR do not match up very well together indicative of a bad ranking system, I am not saying you don't rank accurately with skilled players, but it baffles me how someone who has the same MMR is so much higher CSR than others.

Please explain the thought process behind this system and why there needs to be two different ranking systems, one which is invisible and one which has no effect in matchmaking? Why not just have one which you can see at all times? That does everything, that way you wouldn't have the issue where people match MMR but not CSR?I feel the system defies logic and does little to help the game, I sometimes wish Halo adopted a ranking system similar too Hearthstone? For example why does grinding from Gold 1-5 take the same time as grinding Diamond 1-5? surely the higher you get, the longer the grind should be? Just my thoughts on it. As right now ranking can be a lot of effort for little reward. Especially when the matchmaking system puts you up against players of the same skill but with them having the possibility to be double your rank.
I DSG I wrote:
So does season reset tomorrow and will breakout be back in this season ( the real thing )
Season resets March 9th to coincide with the maintenance update.
Wow I don't know that thanks
Collet005 wrote:
I DSG I wrote:
So does season reset tomorrow and will breakout be back in this season ( the real thing )
Season resets March 9th to coincide with the maintenance update.
where did u read it?
https://www.reddit.com/r/halo/comments/5win43/janfeb_season_ending/deafscr/
Hopefully I can snag your attention once more, Josh, but I'm really interested in your response.

How do you feel about 343's choice to introduce the ability to drop between divisions? I think you mentioned something about a buffer period for the 6/1 tiers in one of your original posts but I'm curious which one you would rather have by default (Locked Divisions vs Droppable Divisions)

Personally, as long as MMR is active in the background, I don't see the reason to include this method of ranking. Division locking provides more meaning to rating progression and having a player float between two divisions repeatedly is very jarring and odd for skill representation. Granted, there are a lot of factors that can play into this debate, but those factors aren't even present in Halo 5 (Seasonal rewards, promotion matches, etc.)

LITTLE TE wrote:
Please just bring back the 1-50 ranking system. Much more appealing than these tiers.
Agreed. Halo 2's ranking system is easily the best out of all the Halo games, and it isn't even close.
This is a very arbitrary statement, you can't really say that without presenting an argument as to why that is true.

1-50 systems are flawed (forced grind causes inaccurate matching in lower ranks, high-end skill has no endgame/differentiation, visual skill differential is much smaller, etc.). There is a reason every competitive game uses rating/division systems.
eLantern wrote:
For example, I think turning the (Social) Fiesta Slayer playlist into a mix-bag (Social) Team Slayer playlist that included all of these slayer-based modes:
  1. Slayer (medium high rotation)
  2. Super Fiesta (medium high rotation)
  3. Regicide (medium rotation)
  4. Shotty Snipers (medium low rotation)
  5. SWAT (medium low rotation)
...is an excellent method to maintain mode diversity and bring back a past mode or modes under the social banner while fitting it all under a particular playlist theme (aka a slayer-based objective).

Replace Fiesta with a Mix-Bag playlist is a half -Yoinked!- solution. You take players the possibility to play their favorite game modes and split the playerbase even further. It would be better to let people create their own Mix-Bag playlist with multiple game modes. This would kill two birds with one stone.

Titanfall 2 and Rocket League do something similar. In both games you can select multiple game modes at once and matchmaking will look for available games for all simultaneously.

The benefits are:
- You only play your favorite game modes
- It's possible to look for games in lifeless playlists without sitting in matchmaking forever. If you can't find a game in Fiesta for example it will put you in Action Sack. After that game you get another chance for a fiesta game
- More potential players for matchmaking. Most people back out of a game mode if they can't find a match within a minute. At off-peak hours it's even not worth to try because most players know the playlist is empty. This leads to an chicken-and-egg problem. If you can select multiple game modes at once, people will still try to find a game in a lifeless playlists even if the chance to find one is small. This leads to more potential players
- Shorter matchmaking times. It's faster to find games in a large player pool from multiple playlists instead of one that isn't very frequented
- If you want to play a specific playlist and are willing to wait for a while you are able to.

Even if you don't use this option you profit from it because playlists are more frequented. It is a godsend for Titanfall 2 on PC. It populated game modes like FFA or Last Titan standing. Before it was impossible to find games late at night or early in the morning in those. Best thing is every game mode is used with equal frequency, you don't play the most populars only.

Only drawback I see is the more computing power required on server side. And please give us a button to skip the 30 seconds between matches.
So do we know at all if there will be changes to breakout March 9th?..... I hold out hope every weekend update and new season that my hearts desire will come true. Please for the love of all that is good and holy make breakout great again. Pleaaaasssssse
eLantern wrote:
Replace Fiesta with a Mix-Bag playlist is a half -Yoinked!- solution. You take from players the possibility to play their favorite game modes and split the player base even further. It would be better to let people create their own Mix-Bag playlist with multiple game modes. This would kill two birds with one stone.

Titanfall 2 and Rocket League do something similar. In both games you can select multiple game modes at once and matchmaking will look for available games for all simultaneously.

The benefits are:
- You only play your favorite game modes
- It's possible to look for games in lifeless playlists without sitting in matchmaking forever. If you can't find a game in Fiesta for example it will put you in Action Sack. After that game you get another chance for a fiesta game
- More potential players for matchmaking. Most people back out of a game mode if they can't find a match within a minute. At off-peak hours it's even not worth to try because most players know the playlist is empty. This leads to an chicken-and-egg problem. If you can select multiple game modes at once, people will still try to find a game in a lifeless playlists even if the chance to find one is small. This leads to more potential players
- Shorter matchmaking times. It's faster to find games in a large player pool from multiple playlists instead of one that isn't very frequented
- If you want to play a specific playlist and are willing to wait for a while you are able to.

Even if you don't use this option you profit from it because playlists are more frequented. It is a godsend for Titanfall 2 on PC. It populated game modes like FFA or Last Titan standing. Before it was impossible to find games late at night or early in the morning in those. Best thing is every game mode is used with equal frequency, you don't play the most populars only.

Only drawback I see is the more computing power required on server side. And please give us a button to skip the 30 seconds between matches.
I did acknowledge the drawback of the developer created mix-bagged playlist in my post on the first page and I would agree that a custom mix-bagged approach using a system similar to what Titanfall 2 and Rocket League have would be fantastic (though, Halo's playlist offering would need to be expanded -- not condensed with that type of system), but the problem as I see it is that I don't know how practical implementing something like that would be for Halo 5 given the investment of resources needed to make it happen. I'm all for 343i looking into that type of system for future titles (and as an update to MCC), but in the absence of being able to provide that type of system within Halo 5 right now I think that a developer created mix-bagged playlist, and in particular the two playlists that I listed and outlined, would be ideal for Halo 5 under the Social banner.
ZaedynFel wrote:
Do I end up 200 CSR back from where my CSR ended last season?No, you end up 200 back from where your MMR moves to during placements. So if you go 10-0, that’s a different spot than where you were before placement, and you could actually start ahead of where you were last season. We aren’t just backing you up again and again.
Matches are better towards the season and when CSR is closer to MMRCSR has no influence on matchmaking at all. The matches you get throughout the season are exactly the same as they are at the end, unless you have actually change[d] as a player, thus moving your MMR. Even though you sometimes see a different range of CSRs in your match, the actual skills are really close.
What if I did not play enough games to get a rank last season but had a rank the season prior to that? Surely the game must sense what I did last time I played even if that's a couple of seasons ago and not merely based on last season. I had that scenario just happen where I really can´t make any sense of what you just described compared to the outcome after the 10 placement matches which I played with a friend of mine. I outlined the details of it in a post in the previous thread you made.

I get the feeling my MMR is almost always higher than my CSR. In theory my CSR should increase and it does over the course of the season but since I am almost always dropped down back to about where I began last season this becomes a grind with no end, a vicious cycle even, against higher ranks. I can tell MMR changes rapidly by who I am matched against after a couple of wins or losses but it feels like I never catch up. I think this is good in initial placement matches to quickly gauge a new players skill, less so during regular games or with a seasoned player.

This is especially true when you mostly play alone or with one or at most two friends. Not having a full fireteam always decreases your chances of winning even though on a strict skill level you can match your opponents. Made further unjust by the way some of us always have higher ping than some others to the servers. I think you get what I am trying to get across.
ZaedynFel wrote:
Hey guys, this will be a bit shorter this week, partly because I was traveling. But keep your feedback coming!

Why not show MMR all the time instead?The main reason we don’t splash MMR all over is we reserve the right to use whatever information we can to update your MMR. This can include information that could potentially be unfair to use to update your CSR. We don’t do a lot of this right now, but I could see a world where MMR can be affected by, e.g., how you played in a completely different playlist. In that case, we update your MMR to a more accurate reflection of your current skill, but it would be unfair to de-rank you magically outside of you playing on that playlist. Showing an MMR that is this volatile would be confusing.

CSR should be the most accurate measure of what you deserve to show as your Rank. MMR should be the most accurate measure of where we should matchmake you right now. Those aren’t always the same, and don't always have the same rules of fairness that govern them.

That said, I do also see a future where instead of directly showing the MMR that made the match, we could show some summarized version of your opponents' strength to give you a feel for why you were matched with them. For example, we could tell you that even though your opponents all look Diamond, they just played like Gold players, which matches up nicely with you being Gold.
Seasonal Resets I’ve seen a few posts discussing the pros and cons of seasonal resets. We don’t have any plans on changing the core mechanics of how season’s currently work. Some nice things about seasons:

    • They give you something to shoot for
    • They filter out inactive players so we don’t need decay or a way to mark activity
    • Can be used as a platform for more competitive constructs
That said, we are evaluating:

    • The number of placement matches
    • How many games it takes to “prove” your rank. This is around 30-50 now.
No news yet on changes, but we’re thinking.
Do I end up 200 CSR back from where my CSR ended last season?No, you end up 200 back from where your MMR moves to during placements. So if you go 10-0, that’s a different spot than where you were before placement, and you could actually start ahead of where you were last season. We aren’t just backing you up again and again.
Matches are better towards the season and when CSR is closer to MMRCSR has no influence on matchmaking at all. The matches you get throughout the season are exactly the same as they are at the end, unless you have actually change[d] as a player, thus moving your MMR. Even though you sometimes see a different range of CSRs in your match, the actual skills are really close.

Are we looking at dealing with parties in Social Matchmaking and Warzone as well?Yes.

Could we have Opt-In seasons?This isn’t something we can do for Halo 5, but I thought it was a cool idea. I could see having an ongoing CSR that never resets (though we’d probably need some decay), and then one for the season if you opt-in. If we did this, I would only give out rewards to players who decided to opt-in to the season. There are other complications, like, which to show on the PGCR, but could probably be worked through.
So If I understand this right ......at reset last season i ranked gold 6 and play against platinum 6 and higher this whole season and end up going down in rank all the way to silver 3 at one time... and as of last night i still played against platinum mid and high rank players......at a high silver rank.... so i played this whole season of Slayer where it was 4 low to mid gold rank against 4 platinum that was more mid and high rank because they play like gold..........with respect i didnt feel like that they was playing like gold players..... really felt like all my games was out match and boosting VS them.. even tho 70% of them i was in the positive other players was gettting req i mean as i was my self too ..... but to still play against the some level of rank even tho i drop from g6 to sliver 3 at one time i cant even get back to gold because it is platinum im match against still.......
yes ...I play with randoms 90% of the time.......all other play list i play i am platinum ....... slayer is only one i have this situation .... but all and all you guys and ladys rock...... I know you all do and give all u can to this ............. my hat is off to you all mad respect...............
  1. 1
  2. ...
  3. 2
  4. 3
  5. 4
  6. ...
  7. 6