Forged Alliance Forever Forged Alliance Forever Forums 2020-03-22T21:48:36+02:00 /feed.php?f=2&t=17722 2020-03-22T21:48:36+02:00 2020-03-22T21:48:36+02:00 /viewtopic.php?t=17722&p=182751#p182751 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
keyser wrote:
TMM will most likely resolve this issue, but for this to happen we need most of the people to transition to TMM.

herein lies the problem, you need people to switch to tmm or it wont work, so you break their existing games so they have no choice but to play "new thing" that they didnt know existed until now, unable to play what "they were already playing", not a great way to introduce something new

the "random numbers" were custom lobbies matches vs not-custom lobbies matches (being ladder only atm), I think we all know that custom games dominate faf atm in % of players that play them, idk what amount of players are going to play tmm (preferably enough for it to work), but very likely most of people will play custom games anyway, you can replace 90% with "most of" if you dont like numbers
keyser wrote:
Also it should also be possible in the algorithm to implement a mirror function to put similarly rated players in front of each others. We have already a mirror option for launching the game, i'm not sure if it works well though.

mirror option doesnt always work, I already gave example on setons both navies getting stacked when using it, unless you want to manually write mirror code for every map so it balances it properly

another problem with auto balance is when there arent similar rated players on a map, and random balance puts unbalanced air slots vs each other, which means game is over as soon as better air gets t3 and goes strats while other air is building t1 bombers and land, hosts know air slot is important and can destroy the game if unbalanced and always put balanced airs vs each other so this doesnt happen, do you intend to put this in code as well (different for each map that has dedicated air slot), and this isnt only for air slots, many other slots have different importance on a map and depending on how good player that is on that slot is, it can ruin the match (like a pro in joe lobby on setons rock)

and all this because hosts couldnt balance on their own

rating being a reward for players is a problem in general (that they think it is a reward), and reason for mentioning it is that it shouldnt have anything to do with whether global rating exists, esp. not to point where you want to "punish them" for getting an integer to specific value by playing same map over and over, by removing it (idc about my rating inb4 insults and "if youre not at least 2000 you have no say in anything")

the reasons for whether global rating is removed or not should be only for balancing games and nothing more, including getting people to play tmm, which should be done different way than "you have to because you cant play custom anymore"

Statistics: Posted by Mach — 22 Mar 2020, 21:48


]]>
2020-03-22T20:41:40+02:00 2020-03-22T20:41:40+02:00 /viewtopic.php?t=17722&p=182750#p182750 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
keyser wrote:
Also most of the time people play gap & co., because there is nothing else in comparison to launch a game fast.

That is a good point actually that most people probably won't think of because it is not immediately obvious. Multiple people when asked why they play gap all the time have said it is because it is the only way to quickly start a game they are "allowed to play" and do not get kicked for rating/gamecount mis-match.

Statistics: Posted by UnorthodoxBox — 22 Mar 2020, 20:41


]]>
2020-03-20T18:31:06+02:00 2020-03-20T18:31:06+02:00 /viewtopic.php?t=17722&p=182702#p182702 <![CDATA[Re: Team Matchmaker Info&Discussion]]> Statistics: Posted by Azraeel — 20 Mar 2020, 18:31


]]>
2020-03-20T17:10:38+02:00 2020-03-20T17:10:38+02:00 /viewtopic.php?t=17722&p=182697#p182697 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
Mach wrote:
we arent lesser than you because we play them and so not worthy of a rating system.

you shouldn't do strawman, it makes you argumentation feels really bad afterward

Mach wrote:
rating isnt a "reward"

yes it shouldn't but it is. That's why we also try to change how the rating will be displayed for TMM/ladder but that's some other discussion.

Mach wrote:
90% of games that are being played being unbalancable

Not sure what's up with you putting up random numbers, anyway hopefully more than 10% of games will be in TMM. Also most of the time people play gap & co., because there is nothing else in comparison to launch a game fast. TMM will most likely resolve this issue, but for this to happen we need most of the people to transition to TMM.

Mach wrote:
everyone knows game quality % means nothing if one of slots is completely outmatches

Also it should also be possible in the algorithm to implement a mirror function to put similarly rated players in front of each others. We have already a mirror option for launching the game, i'm not sure if it works well though.

Statistics: Posted by keyser — 20 Mar 2020, 17:10


]]>
2020-03-20T16:09:06+02:00 2020-03-20T16:09:06+02:00 /viewtopic.php?t=17722&p=182695#p182695 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
Mach wrote:
wrong reasons again. rating isnt a "reward"

Yes it is? In both faf and other games there are tons of people that want to get as high of a ranking as possible.

Statistics: Posted by Farmsletje — 20 Mar 2020, 16:09


]]>
2020-03-20T14:58:49+02:00 2020-03-20T14:58:49+02:00 /viewtopic.php?t=17722&p=182690#p182690 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
keyser wrote:
The goal of hiding global rating is to give an incentive to people to play TMM. Rating is important for people, so this will turn custom games into games that are more casual.

wrong reasons again. rating isnt a "reward", we need it so we can balance custom games. everyone knows game quality % means nothing if one of slots is completely outmatches other, I dont want a 1000 (team 1) to play vs 2000 (team 2) on front and 1800 (team 2) vs 800 (team 1) on navy, just because I cant see which is which, yet it shows equally balanced as 2000 v 1800 and 800 v 1000 according to game quality.

you have to find another way to "incentivize" people to play tmm other than by wrecking custom games into casual roflstomps every match just because you consider them such. we dont. when we play custom games we want some degree of balance in them. we arent lesser than you because we play them and so not worthy of a rating system.

instead of forcing people to play what you consider a test of skill, you have to play around maps they want to play, and effect their rating accordingly, yes its more difficult than simply having everyone play what you need them to so rating system can work properly without changes, but at a cost of 90% of games that are being played being unbalancable

It would be fine if you hid global rating outside of custom lobbies so "people cant brag about it" or whatever your problem with it is, but we need it visible in custom lobbies to be able to balance, game quality % is meaningless for it and any auto balance system is always going to have problems (ex. optimal balance always stacks both setons navies, and sometimes other slots as well) tmm rating is meaningless for it again for multiple reasons (if people are way better on specific map and for people that dont play tmm)

Statistics: Posted by Mach — 20 Mar 2020, 14:58


]]>
2020-03-20T13:21:29+02:00 2020-03-20T13:21:29+02:00 /viewtopic.php?t=17722&p=182689#p182689 <![CDATA[Re: Team Matchmaker Info&Discussion]]> we introduce some algorithm that i linked before, to balance the game in the lobby. (as i said this algorithm need improvement, to take into account which group want to play together)
People can still choose their spot (in opposition to the TMM) and choose the map they are going to play.
The balance quality (%) is still displayed, so host can do manual adjustement and mesure the impact on the balance quality. (but it isn't require as the algorithm should sort out the team well enough).
The algorithm use the global rating to balance the team. So the global rating exist, and is being used, but isn't displayed.

The goal of hiding global rating is to give an incentive to people to play TMM. Rating is important for people, so this will turn custom games into games that are more casual. But we still have a global rating hided, so people playing only 1 map will be able to play in balanced games.

This is a middle ground, between removing the global rating (big incentive for TMM, but hardly balanced custom games) and keeping a global rating and displaying it ("balanced" custom games, but competition between the global rating and the TMM rating).

Statistics: Posted by keyser — 20 Mar 2020, 13:21


]]>
2020-03-20T01:19:54+02:00 2020-03-20T01:19:54+02:00 /viewtopic.php?t=17722&p=182683#p182683 <![CDATA[Re: Team Matchmaker Info&Discussion]]> Statistics: Posted by FtXCommando — 20 Mar 2020, 01:19


]]>
2020-03-20T00:54:43+02:00 2020-03-20T00:54:43+02:00 /viewtopic.php?t=17722&p=182681#p182681 <![CDATA[Re: Team Matchmaker Info&Discussion]]> Statistics: Posted by Mach — 20 Mar 2020, 00:54


]]>
2020-03-19T22:29:26+02:00 2020-03-19T22:29:26+02:00 /viewtopic.php?t=17722&p=182677#p182677 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
Blodir wrote:
Btw I feel like everyone in this discussion is expecting TMM system to be perfect from the get go. More than likely there will be at least some technical issues, shitty map pools, not being able to play with friends etc. etc. If you remove global there is no coming back


Anyone expecting that didn't read my first post.
1) The discussion is built on the assumption of minimal problems with matchmaker implementation
2) It is built on the assumption that 2v2 and 4v4 both exist at a minimum.

The first public matchmaker will be getting released as a 2v2 matchmaker, meaning that there would be a transition period between removing global and having matchmaker.

Keyser's solution (talked more about it on zulip so I had a better understanding) is something I'm willing to accept as another solution to global here. Thanks for presenting it.

In regards to Box's post, that is basically my current mindset explained by another person.

Statistics: Posted by FtXCommando — 19 Mar 2020, 22:29


]]>
2020-03-19T22:26:49+02:00 2020-03-19T22:26:49+02:00 /viewtopic.php?t=17722&p=182676#p182676 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
Next step for us is to move onto making sure the server is going to be capable of making proper balanced games based on the rating of players.

Statistics: Posted by FtXCommando — 19 Mar 2020, 22:26


]]>
2020-03-19T20:20:18+02:00 2020-03-19T20:20:18+02:00 /viewtopic.php?t=17722&p=182673#p182673 <![CDATA[Re: Team Matchmaker Info&Discussion]]> Statistics: Posted by Blodir — 19 Mar 2020, 20:20


]]>
2020-03-19T18:12:50+02:00 2020-03-19T18:12:50+02:00 /viewtopic.php?t=17722&p=182672#p182672 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
Mach wrote:
last I checked all of you are ~2000 rated, yet you say you know what low level lobbies are like


He had to get to 2k somehow... Smart guy

Statistics: Posted by Azraeel — 19 Mar 2020, 18:12


]]>
2020-03-19T17:46:11+02:00 2020-03-19T17:46:11+02:00 /viewtopic.php?t=17722&p=182671#p182671 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
global rating removal will like ftx said definitely cause a problem after some time, as custom only players that dont play tmm (something you cannot know whether they will or not) will keep their rating after by playing customs only they get way better than their rating suggests, and it is absolutely unnecessary to do as tmm doesnt depend on global rating removal

Statistics: Posted by Mach — 19 Mar 2020, 17:46


]]>
2020-03-19T16:16:31+02:00 2020-03-19T16:16:31+02:00 /viewtopic.php?t=17722&p=182670#p182670 <![CDATA[Re: Team Matchmaker Info&Discussion]]>
Now of course it could also be argued that team game players are also farming their rating on 3 maps only, but the point is not who is farming rating or not. It is that in normal games like counter strike that have ranked and custom modes, the custom modes provide no rating and there is no balancing, yet plenty of people still play them and have fun. One of the biggest hurdles in faf is that different maps completely change the meta play style. even different 10x10 team game maps have vastly different playstyles; for example, canis river 5v5 is focused mainly on only 2 players fighting with units, while the other 3 focus mostly on eco. Then you take wonder 5v5 open, and have 4 players focus on fighting, with 1 player each ecoing/sniping other players. This isn't rigid, and you can play differently or do well, but it is just an example of metas formed.

Global rating right now is mostly ignored by smart hosts, and only paid attention to in order to make people who complain about game quality percentage happy. Hosts familiar with the players in their lobby usually can guess who will beat who, regardless of what their rating is, and even more so in the dual gap and setons community. Balancing already can be done with no rating at all, but optimal balance would no longer work, which probably should be looked at.

Really the point I am trying to make is that team matchmaker should be looked at similarly to ranked modes in other games, and global games should be seen simlar to custom servers.

Statistics: Posted by UnorthodoxBox — 19 Mar 2020, 16:16


]]>