Jump to content

Impromptu Dev Update About Matchmaker And Faction Play With Ngngtv


55 replies to this topic

#21 KrocodockleTheBooBoxLoader-GetIn

    Member

  • PipPipPipPipPipPip
  • The Deadly
  • The Deadly
  • 308 posts

Posted 03 July 2022 - 06:21 AM

I actually really really like the 8v8 4 stock no groups set up in the current event. I think it's superior to both current QP and FW

#22 Thornhold

    Member

  • PipPip
  • Hero of Steiner
  • 27 posts

Posted 03 July 2022 - 06:29 PM

View PostVladokapuh, on 01 July 2022 - 10:31 AM, said:

just make quickplay solo 8v8, its better


totally agreed

#23 Beeper23

    Member

  • Pip
  • Little Helper
  • 12 posts
  • LocationUSA

Posted 03 July 2022 - 06:32 PM

Hmm an idea hit me. Some people are wanting faction play to be jammed into quickplay while others are wanting it to be its own thing. So why not do the first bit with event matches (even as a test)?
This is gonna show how little I bothered with it but if I recall in Solaris you could pick a sponsor right? So what about seeing if you can just yank that idea/code and slap it into event mode but the sponsors are just the factions? Can be the same points/system from faction, you already use drop decks, hell you can even say "its Solaris 2.0" if you really want and that'll allow IS with clan on same team if you give a **** about lore.
Best part is you can still try this no matter what else you do so if you want to keep faction play with its own maps its still on the table. Still on the table if you want to rework faction completely.
OK going back to drinking.

#24 King Samu

    Member

  • Pip
  • Ace Of Spades
  • Ace Of Spades
  • 19 posts

Posted 03 July 2022 - 07:15 PM

Simple solution:

1. Use a new metric similar to match score: simply the average damage (and maybe average damage + average kills, because that's pretty much what it all comes down to does it not?) for the weight class used by each pilot in a match, to determine the "tier" of a given pilot.

2. Increase the granularity of said "tier" in step 1 to account for the top 1% players (so a score from 0-100, where "tier" 100 is a top 1% player), but instead of separating tiers by matches, simply balance the teams in a match so each team has a similar total of said "tier" points.

In other words, if we balance the teams by a metric that matters most and is most impactful, we no longer need to separate matches by tiers, also tackling the population issue. This also gives an opportunity for current tier 5 players exposure to and to mingle with (and learn from) tier 1 players, which makes the community feel larger as well. Splitting matches by tiers is redundant and unnecessary if there's an effective way to balance teams.

This new score also takes into account the performance of a pilot in the weight class they're using, so we can have more flexibility in tonnage restrictions / attempts to balance by tonnage. Group limits also become more flexible, because as long as each side is balanced, the match is a go. We can effectively balance teams using one single metric. Simple solutions are the best solutions.

Edited by King Samu, 03 July 2022 - 07:25 PM.


#25 Voice of Kerensky

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 337 posts

Posted 03 July 2022 - 10:48 PM

Faster fix the matchmaker. Or refuse to combine solo and group queue in a quick play.
The quality of the matches caused a lot of complaints long before the combination of the solo and group queues. But instead of eliminating the shortcomings of the matchmaker that existed at that time, a (I think) completely rash and irresponsible decision was made to merge queues, which could only aggravate the situation with the quality of matches.

#26 TruMantas

    Member

  • Pip
  • Knight Errant
  • Knight Errant
  • 18 posts

Posted 04 July 2022 - 01:15 AM

This

View PostKing Samu, on 03 July 2022 - 07:15 PM, said:

2. Increase the granularity of said "tier" in step 1 to account for the top 1% players (so a score from 0-100, where "tier" 100 is a top 1% player), but instead of separating tiers by matches, simply balance the teams in a match so each team has a similar total of said "tier" points.


Isn't compatible with this

View PostKing Samu, on 03 July 2022 - 07:15 PM, said:

1. Use a new metric similar to match score: simply the average damage (and maybe average damage + average kills, because that's pretty much what it all comes down to does it not?) for the weight class used by each pilot in a match, to determine the "tier" of a given pilot.


Any metric that compares the performance of high tier vs low tier players will only widen the gap when in a system that doesn't separate players from different tiers. This shouldn't require saying. The changes you're proposing is like saying that we should have Premier League players and Juniors (U18) playing in the same team and keep track of who scores the most goals.

#27 Beeper23

    Member

  • Pip
  • Little Helper
  • 12 posts
  • LocationUSA

Posted 04 July 2022 - 01:29 AM

View PostKing Samu, on 03 July 2022 - 07:15 PM, said:

Simple solution:

1. Use a new metric similar to match score: simply the average damage (and maybe average damage + average kills, because that's pretty much what it all comes down to does it not?) for the weight class used by each pilot in a match, to determine the "tier" of a given pilot.



I kinda got a bit of a nickpick/gripe/whatever on the first part mostly two parter
A: I dont think damage would be the best metric or at least still allow objectives to give score since already people will throw games if it meant they can get one more shot in for said damage as it is now. unless skirmish was the only quickplay mode lol.

B: "weight class used by each pilot in a match" in a match is the biggest negative I have to the idea mostly because its one of the main issues with the system now. It wont matter what you play if one assault decides because they go 70+ they can go scout only to end up getting shredded thus giving the enemy team easy damage (more or less making you lose rating for somebody else's actions). Sadly this one suggestion I got to fix it shouldn't be that hard but probably the least likely to have done and that is maybe once a day/week/whatever just get the averages of every player for each weight class and check that to end game score then go from there.

Honestly I think the only real issue to the idea is what I said in B because of it being one of those devil in the detail things that tends to bite people in the ***. Outside of that I like the idea and down for giving it a go so if you ever want to try and push it I'm down to support it.

Edited by Beeper23, 04 July 2022 - 01:36 AM.


#28 Roodkapje

    Member

  • PipPipPipPipPipPip
  • The 1 Percent
  • 344 posts

Posted 04 July 2022 - 05:19 AM

View PostThe Chancelor, on 03 July 2022 - 05:26 AM, said:

Well I guess that is why most of us are the same age like moadepe or blue^^

I have no idea who those guys are ?! Some YouTube/Twitch dudes I am guessing ?? Posted Image

#29 Alreech

    Member

  • PipPipPipPipPipPipPipPip
  • Little Helper
  • Little Helper
  • 1,632 posts

Posted 04 July 2022 - 12:45 PM

View PostNomad Tech, on 02 July 2022 - 07:12 AM, said:

Adding bigger groups will make it even more lopsided.. Groups are the problem not the size of them.. Yes i know its an online game but groups are killing this online game..

if you want large groups knock the teams up to 24x24..

Groups are the main reason i don't play often, getting slaughtered every match by coordinated high tiers isn't fun after a while.

At the moment the matchmaker needs 24 players to start one match (12 vs 12).
That 24 players are the absolut minimum, without the matchmaker taking tiers and mech class into account.
So for fair matchmaking the matchmaker needs more players than 24.
Knocking up the teams to 24 vs 24 would make that worse, the matchmaker would need 48 players if it ignores tiers and mech class...
Reducing the match size to 8 vs 8 would need only 16 players.
IMHO the best solution for all players:
  • Make quickplay solo only and 8 vs 8 to speed up matchmaking.
  • Add group quickplay with small groups use 4 vs 4, 8 vs 8, 12 vs 12 depending on the number of groups aviable. Use drop decks in group quickplay to remove mech class as factor from matchmaking.
  • Keep the old faction play for bigger groups


#30 Avalon91211

    Member

  • PipPipPip
  • The Star
  • The Star
  • 96 posts
  • LocationAn experimental SLDF McKenna class Battleship, waiting for Clan Nova Cat to return to the Inner Sphere in 3160

Posted 04 July 2022 - 02:00 PM

I'm sorry but everyone calling for 8v8 is in the wrong. Thats my opinion and if you don't like it, oh well. 8v8 will not solve anything with the matchmaker being broken as it is, it'll just make things worse. On top of which everyone here keeps complaining it takes them forever to find matches. Ok so is 5 minutes or less forever for you? Because that's my average wait time, 5 minutes or less. I haven't had to wait more than that for awhile now. I also don't play strictly on the NA server, I play on every available server. Perhaps try that if you can't get a match super fast? Or better still, slow down, chill out, and have fun.

#31 Sawk

    Member

  • PipPipPipPipPipPip
  • Star Captain
  • Star Captain
  • 273 posts

Posted 04 July 2022 - 03:23 PM

HMM all kinds of folks commenting, so i will put my 2 CENTS in.
FIRST OFF - find a mode -- i'll use SCOUTING because i had blast playing it for years, when i could and when others did.
SECOND OFF if you pick my SCOUTING, you advertise it, so folks can build mechs, maybe 2 man 4 man teams, ect ect.
THIRD OFF do it on a weekend start, with an EVENT top teams--- top pilot-- TOPGUN LOL -- was thinking 5 days, from say Thursday thru monday,
FOURTH OFF i know you looking for 1 mode to use but it is not that easy, but i think if you ADVERTISE, the event, SPOCK IS IN TROUBLE and need help on vulcan, HIGH HEAT, HEAVY GRAVITY, ECT ECT, you will find your a group of folks BUILDING MECHS and making freinds with HOW TO WIN ECT ECT

SAWK CLANNER

#32 Guardian Soul

    Member

  • Pip
  • 15 posts

Posted 04 July 2022 - 08:36 PM

View PostKrocodockleTheBooBoxLoader-GetIn, on 03 July 2022 - 06:21 AM, said:

I actually really really like the 8v8 4 stock no groups set up in the current event. I think it's superior to both current QP and FW


TL;DR
I'm not hard set on 8v8 4 'mech matches or anything, but I think bringing some form of backup 'mech(s) into standard quick play could be quite nice as it gives a bit of a buffer against bad experiences from getting blown up very early in the match and it gives the matches a bit more texture and draws out the fights in a nice way.

Long(ish) version
Want to give this a big +1. Quick play 12v12 can be fine but there's downtime in the beginning of the match as the teams find each other and it can really hurt to come right from that into getting caught out and burned down in moments (yes yes I know, position better and play with the team but sometimes mistakes happen). Being able to hop into another 'mech instead of having to go back to queue is really nice, and it's fun to have scrappy moments where more players are at low health consistently through the match instead of just towards the end.

It doesn't feel like it makes the matches take that much longer, and I really can't stress enough how much better it feels to have four 'mechs to go through in a match instead of just the one try to have a good match, especially considering the time spent in queue, loading screen, countdown, walking to the fight, etc. I know that makes quick play a lot more like faction play-lite but I think having the standard quick play objectives and maps and keeping it framed as the 'casual' game mode could still give it a different texture to faction play.

The 8v8 all heavies 4 drops each event matches were honestly the best time I've had with MWO and I like the dang game a lot, so that's saying something.

#33 Alreech

    Member

  • PipPipPipPipPipPipPipPip
  • Little Helper
  • Little Helper
  • 1,632 posts

Posted 05 July 2022 - 10:05 AM

View PostLockheed_, on 04 July 2022 - 06:16 PM, said:

On top of that PGI's goal is to retain and get more players. For years the game has been 12v12 if suddenly the game is changed to 8v8 this might cause people to leave. It's very unlikely thought that keeping the 12v12 format is going to impact player numbers.
Also some of major points of criticism that caused people to leave were the skill tree, matchmaking and faction play, so moving to 8v8 is also not even going to bring back droves of old players, like addressing these other issues are more likely to do.
It's a high risk low reward move.

The key problem for matchmaking are the low player numbers.
With 12 vs 12 the matchmaker needs 24 players, and can't even balance player tier and mech class.
Moving to 8 vs 8 the matchmaker only needs 16 players (1/3 less than before), allowing the matchmaker to set up fair matches by player tier & mech class.

View PostAvalon91211, on 04 July 2022 - 02:00 PM, said:

I'm sorry but everyone calling for 8v8 is in the wrong. Thats my opinion and if you don't like it, oh well. 8v8 will not solve anything with the matchmaker being broken as it is, it'll just make things worse.

Having a opinion is fine, backing it up with facts why you have this opinion is even better.

Quote

On top of which everyone here keeps complaining it takes them forever to find matches. Ok so is 5 minutes or less forever for you? Because that's my average wait time, 5 minutes or less. I haven't had to wait more than that for awhile now. I also don't play strictly on the NA server, I play on every available server. Perhaps try that if you can't get a match super fast? Or better still, slow down, chill out, and have fun.

Players don't only complain about waiting time to get a match.
They also complain that the 2 tier distance between players OR the mech class balance is ignored by the matchmaker.
They also complain about player groups spoiling the fun of solo players by using unfair mechanics like teamwork with friends.

So the problems are:
  • long waiting time in quickplay
  • bad player balance in quickplay
  • bad mech balance in quickplay
  • groups in quickplay
  • waiting time in faction play
Any idea how to fix that without reducing the player numbers per match and/or removing groups from quickplay?

#34 Void Angel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 5,750 posts
  • LocationParanoiaville

Posted 05 July 2022 - 01:12 PM

This is your periodic friendly reminder that "ELO" is the Electric Light Orchestra, and that the name of the player rating system is "Elo," a word derived from its creator's name.

#35 Tarteso

    Member

  • PipPipPip
  • The People's Hero
  • 73 posts
  • LocationSpain

Posted 05 July 2022 - 01:19 PM

View PostAlreech, on 05 July 2022 - 10:05 AM, said:

The key problem for matchmaking are the low player numbers.
With 12 vs 12 the matchmaker needs 24 players, and can't even balance player tier and mech class.
Moving to 8 vs 8 the matchmaker only needs 16 players (1/3 less than before), allowing the matchmaker to set up fair matches by player tier & mech class.


Having a opinion is fine, backing it up with facts why you have this opinion is even better.


Players don't only complain about waiting time to get a match.
They also complain that the 2 tier distance between players OR the mech class balance is ignored by the matchmaker.
They also complain about player groups spoiling the fun of solo players by using unfair mechanics like teamwork with friends.

So the problems are:
  • long waiting time in quickplay
  • bad player balance in quickplay
  • bad mech balance in quickplay
  • groups in quickplay
  • waiting time in faction play
Any idea how to fix that without reducing the player numbers per match and/or removing groups from quickplay?



If PGI/cauldron really want to fix the balance, they should begin experimenting with some bold changes in this game:
  • long waiting time in quickplay > low population, less games played per season: unfixable unless you could attract more players or keep people playing for longer. Live with it. Or go again for the 8 vs 8.
  • bad player balance in quickplay > prioritize people's skill over waiting times (tier, average match score, whatever)
  • bad mech balance in quickplay > limit total tons (min-max, average tons per player something like 60) or go for the 3/3/3/3 formula
  • groups in quickplay > see what happens after changing 2 and 3, no less than 1-2 months of testing, and remove groups. Better? Same? Worse? Don't figure it, TRY IT
  • waiting time in faction play > attract more people to play this mess, e.g., groups playing QP


#36 Void Angel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 5,750 posts
  • LocationParanoiaville

Posted 05 July 2022 - 01:26 PM

On-topic, 8v8 may solve some balance issues, or at least amelorate them. Having four more people on the enemy team to shoot at you makes a lot of things more punishing or difficult - newbie positioning errors and brawling, for example. I'm not sure if I would go 8v8, but this is still a fact that should be acknowledged when discussing it.

As for faction play, one of the big reasons my guildmates, Back in the Day, told me they didn't like Faction Warfare was that the rewards simply didn't reflect the amount of time and organization that went into it. They also felt like the strategic level of FW was simplistic and flat - and I agreed. These are probably the major factors that led to the formation of giant, multi-unit "alliances" to dominate FP with numbers (yes, some of MercStar's players were very good; but most of them were average, and it was their massive numbers that made them hard to deal with.)

Overhauling the faction play mechanics is a large undertaking - honestly, I think you should hire a board game designer for that, but I digress. The best short-term fix for Faction Warfare would likely be to find some meaningful rewards for individualunit performance, modified by how well their faction did, and implementing a seasonal reset system to both keep things moving and to allow iteration and testing of FW rewards and mechanics changes.

#37 Nomad Tech

    Member

  • PipPipPipPipPipPipPipPipPip
  • Shredder
  • Shredder
  • 3,392 posts
  • LocationAustralia

Posted 05 July 2022 - 03:21 PM

If they kept event queue i'm all for groups in quickplay.. Event queue is so much more fun.

#38 Sawk

    Member

  • PipPipPipPipPipPip
  • Star Captain
  • Star Captain
  • 273 posts

Posted 05 July 2022 - 04:05 PM

WOOT GUYS you just explained a 3 tier way to fix things ---- QUICK PLAY 12 man 5 man groups to test the new, ok 4 man, and event to test then weekly, they really need BUY or BUILD mechs to play here - and then a FACTION PLAY that reflects the TRUE PILOT with a 4 man 4 mech DECK, that can also work in the event.
I WAS HOPING for some scouting missions, LIKE take over COMMS tower, in 4 man groups, its what i trained for in the ARMY, maybe you need real live army guys to HELP defend and ATTACK LOL

SAWK CLANNER

#39 Voice of Kerensky

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 337 posts

Posted 05 July 2022 - 06:17 PM

View PostAlreech, on 05 July 2022 - 10:05 AM, said:

So the problems are:
  • long waiting time in quickplay
  • bad player balance in quickplay
  • bad mech balance in quickplay
  • groups in quickplay
  • waiting time in faction play
Any idea how to fix that without reducing the player numbers per match and/or removing groups from quickplay?




Let me comment on your post.
Of the five issues you listed, four were either exacerbated by the merging of the group and solo queues, or directly brought by the putting of groups into the single player quick play.

1) The waiting time for a match with the introduction of groups could only increase so that the matchmaker could select teams comparable in terms of the level of players (groups) and (ideally) equalize the tonnage of the teams. In order to avoid a long wait time for a match, PGI had to sacrifice two things that you said, and I have already mentioned them. About them below.
2) The balance of the players, which before the putting of groups in a solo quick play was already limping on both legs, with the putting of groups could only receive a crushing blow. After all, now the matchmaker (ideally) needs to take into account the level of players in groups in order to select teams of comparable strength. But this directly led to an increase in the waiting time for the match. Therefore, PGI was forced to force the matchmaker to ignore this variable.
3) The balance of tonnage in teams also had problems before the putting of groups in a solo quick play. And the matchmaker's balance for this variable PGI was also sacrificed in order not to increase the match waiting time to unacceptable values.
4) And the problem of groups in a solo quick play was simply struck by the very fact of the merging of the group and solo queue in a quick play.
As you can see from the first three points, the PGI introduced too many new, hard-to-balance variables into the matchmaker in one action. And this could not but lead to fatal consequences for some aspects of the game. The PGI decided to sacrifice the level of players (groups) balance and the tonnage balance in order to maintain an acceptable match waiting time.
I'm coming to the conclusion that out of the five problems you've mentioned, four (80%) were brought (or greatly exacerbated) by the merging of solo and group queues in quick play.

But I would like to mention a couple of non-obvious problems that arose as a result of this merger:
1) I myself have repeatedly witnessed how people simply left the battle at the very beginning, seeing that their team gathered players at the best of an average level, and all the MWO stars were collected in the enemy team. The first thing that directly follows from this is an even greater deterioration in the quality of the match. The second is a possible outflow of players from the game in principle. I am almost sure that there are people with a fine mental organization who come to the MWO to relax and have fun, but getting completely opposite emotions, they may well not return to the game or take a break for an indefinite time.
2) The second non-obvious problem for PGI. Many people are used to playing in a band and enjoying it. Can they refuse to play if the PGI decides to remove groups from quick play? I think some of these people might actually stop playing MWO. And I see that the PGI may face a difficult choice: if it is impossible to please everyone, then whom to sacrifice? Who will be forced to leave the MWO: people who do not want to see groups in a quick game; or vice versa, people who only want to play as part of a group?

I continue to believe that (given an extremely emaciated player base) PGI made an amazingly reckless, verging on criminally negligent decision to merge the group and solo queues in quick play. This action could not but worsen existing problems and not give rise to new ones.

I apologize for the long and not very optimistic text.

Edited by Voice of Kerensky, 05 July 2022 - 06:20 PM.


#40 SimpleUkyo

    Member

  • Pip
  • CS 2021 Bronze Champ
  • CS 2021 Bronze Champ
  • 17 posts

Posted 05 July 2022 - 11:25 PM

for QP match making:

here are some 3 notable issues:
1. there's no balance between the 2 sides, with or without groups. Sometimes you get to see a stack of 2 high tier groups, and their opponents are appears to be solo that can't even be match against their opponent
2. there's no balance with weight classes. you can have a very synchronized team with huge wolfpack that can dominate a full group of sitting duck assaults
3. there are people who likes to group up casually that are not a problem in the match making because they themselves can't really make a difference in a match unlike a group of highly competitive pilots that stomp the queue

here are some recommendations to address the 3 points:
1. Reduce the matches to 8v8 to lower the wait times
2. Include weight class and balanced ELO rating distribution across the two teams.
3. Include groups that can run 4 mans but it has to maintain the weight class and ELO rating distribution across both teams. Meaning, if you have 4 99% rating group running 2 Assault, 2 Lights, then another 4 solo dropping any combination of weight classes, they should have a match on the other side that has also equal or close to equal in rating with similar weight classes that are either group up like them OR 8 99% dropping solo that matches each side's weight classes.

Basically each side needs to have the same or close to the same number of weight class categories (like world of tanks), as well as equal competitive or casual players on either side.


Additional note: Tier system is not accurate as you have a bunch of T1 are not reflective of skill gaps on T1. There should be higher T1 or there should be no cap at all using ELO maybe. Balance should strike on both teams to have either side a balance ELO. The current system is faulty because there are 99% T1 with 90% T1 and that gap is huge and is not representative of actual balance. Being accurate on player rankings in QP by not placing a cap on the pilots and use that actual ELO system to balance the match making will make the MM better in terms of balancing.


For Faction Play:
This should be treated as a group match making where units should thrive. Faction play died when the incentive to play as a unit was scrapped.

Bring back the planets being controlled by units where their actions matter when conquering planets and it should be rewarded as such.
Just bring back the relevance of having a unit and being part of a faction where playing as a unit in Faction Play as a group relevant. Right now, groups are thriving mainly in QP where it should have been in Faction Play.

Edited by Yamete Kudasai Onee-sama, 05 July 2022 - 11:39 PM.






1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users