35 replies to this topic

[Luminis]

... and meanwhile, we're still playing with a "match maker" that lumps apprx. 80% of the population (such that it is) together and randomly throws them at each other. 180 average matchscore, 500 average match score - meh, close enough. Severly lopsided teams? Also meh, close enough.

I mean, by all means, keep fiddling with faction balance. Who cares if only one in a dozen games feels worth playing. Priorities.

[eminus]

this is how PGI sees the MM and weapon Balance right now

[The Lighthouse]

eminus, on 11 June 2018 - 09:10 PM, said:

this is how PGI sees the MM and weapon Balance right now

I really, really love this picture IMHO.

[El Bandito]

Luminis, on 11 June 2018 - 08:51 PM, said:

... and meanwhile, we're still playing with a "match maker" that lumps apprx. 80% of the population (such that it is) together and randomly throws them at each other. 180 average matchscore, 500 average match score - meh, close enough. Severly lopsided teams? Also meh, close enough.

I mean, by all means, keep fiddling with faction balance. Who cares if only one in a dozen games feels worth playing. Priorities.

And, have you done anything about it? I personally sent multiple suggestions to Russ for improving MM, via Twitter, but it would be more effective if there are tons of other players who openly voice their concerns to him through Twitter, as well. Forums are much less effective.

Edited by El Bandito, 11 June 2018 - 09:36 PM.

[Kaeb Odellas]

Luminis, on 11 June 2018 - 08:51 PM, said:

... and meanwhile, we're still playing with a "match maker" that lumps apprx. 80% of the population (such that it is) together and randomly throws them at each other. 180 average matchscore, 500 average match score - meh, close enough. Severly lopsided teams? Also meh, close enough.

I mean, by all means, keep fiddling with faction balance. Who cares if only one in a dozen games feels worth playing. Priorities.

I think all this can be pretty easily fixed. The only question is how long are you willing to wait for matches?

[Luminis]

El Bandito, on 11 June 2018 - 09:36 PM, said:

And, have you done anything about it? I personally sent multiple suggestions to Russ for improving MM, via Twitter, but it would be more effective if there are tons of other players who openly voice their concerns to him through Twitter, as well. Forums are much less effective.

Made suggestions around here, as has been done by many others.

PGI not bothering with their own forums and communicating strictly via Twitter isn't going to make me create a Twitter account, I'm afraid. Besides, I'm a little baffled that we've come to accept them being that oblivious.

Kaeb Odellas, on 11 June 2018 - 09:41 PM, said:

I think all this can be pretty easily fixed. The only question is how long are you willing to wait for matches?

Evenly spreading the players among the two teams ought to take zero additional time. It's an equation the servers ought to handle in miliseconds. Quick and easy bandaid is anyone gave a damn.

Edited by Luminis, 11 June 2018 - 09:59 PM.

[Kaeb Odellas]

Luminis, on 11 June 2018 - 09:52 PM, said:

Made suggestions around here, as has been done by many others.

PGI not bothering with their own forums and communicating strictly via Twitter isn't going to make me create a Twitter account, I'm afraid. Besides, I'm a little baffled that we've come to accept them being that oblivious.


Evenly spreading the players among the two teams ought to take zero additional time. It's an equation the servers ought to handle in miliseconds. Quick and easy bandaid is anyone gave a damn.

AFAIK, the only thing the matchmaker takes into account are the player's weight class and PSR tier. We both know there are wild variances in mech performance within a weight class and player skill within a tier. The fact that one T1 player can be far superior to another T1 player, and one assault mech can be far better than another complicates things.

And suppose they start arranging teams by player stats. How do they go about determining a player's effectiveness? K:D? Win:Loss? Average match score? What if a player has a high win rate but mediocre k:d or average match score? He could either be a brilliant shot-caller or he could be repeatedly carried in group matches. What about the opposite, where a player has high k:d and match score but low wins? Is he strong and unlucky, or a coward who hides behind his teammates? How does the system account for that?

And what about mech effectiveness? They could try to rate a mech's effectiveness by calculating its average win rate or kill rate, but that doesn't take into account the variety of mech builds. Omnimechs throw the distinctions between variants right out of the window.

I don't think it's as simple as just "evenly spreading the players"

[Elizander]

It can't ever take into account player behavior either. When I want to have fun with a dumb mech I will no matter what tier I reach

I still don't know why PGI will not crank up the PSR penalties for Tier 2 and Tier 1 players. If I'm not serious 100% of the time I shouldn't be Tier 1. A 1:3 W/L ratio is fine to work until T3 but penalties should scale up more in T2 and even higher in T1 so that any "non-serious 24/7" player will never reach T1.

Edited by Elizander, 12 June 2018 - 01:53 AM.

[Luminis]

Kaeb Odellas, on 11 June 2018 - 10:17 PM, said:

I don't think it's as simple as just "evenly spreading the players"

Sorting by average match score isn't the perfect solution, but the observation is simple.

In the vast majority of cases I've checked, lopsided games correlate to a lopsided distribution of average match score. In a 12-0,12-1,12-2 blow out and so on, the winning team had a significantly higher avg. MS total - I bet results are similar if one was to compare KDR, WLR, player percentile or whatever other metric. Avg. MS is, however, arguably easier to scrape than player percentile and not accurate than, say, KDR.

I mean, we can choose to not do anything until someone offers a perfect solution or we could try to improve the situation as best as possible.

[Sjorpha]

It's a false dichotomy though, both things are important to work on.

[Johnathan Tanner]

Without a zero sum match maker any discussion is moot. As pgi has said zero sum will NEVER happen any discussion of how match maker sucks is automatically moot. /thread

[Sjorpha]

Johnathan Tanner, on 12 June 2018 - 02:33 AM, said:

Without a zero sum match maker any discussion is moot. As pgi has said zero sum will NEVER happen any discussion of how match maker sucks is automatically moot. /thread

That's way to categorical. Certainly a zero sum rating system might be better but it's not the only alternative.

ELO systems aren't zero sum for example, but it's open ended so there is no limit to the granularity meaning that you can always sort players correctly regardless of how high the ratings get.

You could have the current system without the positive bias for example, that would be better without being zero sum. There is basically an infinite amout of possible rating systems both better and worse than the current one.

It's always a bad idea to say discussion is meaningless, trying to declare ultimate breakdowns of rationality in a given subject etc. Not only is does it not acheive anything and makes it harder to talk about things, it's also melodramatic and a bit childish.

Edited by Sjorpha, 12 June 2018 - 03:57 AM.

[Luminis]

Johnathan Tanner, on 12 June 2018 - 02:33 AM, said:

Without a zero sum match maker any discussion is moot. As pgi has said zero sum will NEVER happen any discussion of how match maker sucks is automatically moot. /thread

We absolutely need a zero sum matchmaker - and since that, as you pointed out, isn't happening, I'd at least like to make the best out of what we have.

If PGI straight up refuses to implement anything that reduces the skill gap among players within a given match, I'd at least like to try and balance the two teams in the match as best I can. Y'know, try and get something to improve the current situation by implementing something that has not been ruled out.

Otherwise, what are we supposed to do? Sit around with our thumbs up our asses and just accept that matchmaking is gonna stay FUBAR for as long as MWO stays online?

[MechaBattler]

I think a big part of the problem is the variable you can't control, teamwork. You can't make people work together by sheer dint of the matchmaker. Even in group queue, usually the coordinated larger group has the edge. Can't make up for Joe deciding to go left, while the team went right. Even though he's actually good at the game otherwise.

[LordNothing]

the matchmaker is broke because psr is dumb. it would work better if it was more than an exp bar, a noob shield and an epeen fluffer. you know, an actuall measure of skill.

Edited by LordNothing, 12 June 2018 - 09:58 AM.

[Flyby215]

Last night I found the majority of matches quite enjoyable. Most matches were close, with players I recognize and chat with, a variety of strategies, load-outs, and mechs. Not even a single Deathstrike to be seen, and no dual-HGR builds all night. The only dual guass build I saw came from a warhammer that became an unfortunate victim to RNG when my RAC5s miraculously didn't jam when I bull-charged.

Being a Monday night, I wonder if the release valves had been opened and it was just random mash-ups? The distinct lack of flavour-of-the-month mechs/builds really made me wonder what was going on.

[process]

Maybe a dumb question, but how are we presupposing that a lopsided match result is proof of a poorly made match? It just takes one or two guys getting picked off to create a snowball effect that results in those 12-0, 12-1, 12-2 matches. I would even say a 12-6 is more common than a 12-10 or 12-11. What's the metric for determining an even match? I'm highly skeptical even a perfect matchmaking system would produce the results some of you are looking for.

[FupDup]

MechaBattler, on 12 June 2018 - 09:45 AM, said:

I think a big part of the problem is the variable you can't control, teamwork. You can't make people work together by sheer dint of the matchmaker. Even in group queue, usually the coordinated larger group has the edge. Can't make up for Joe deciding to go left, while the team went right. Even though he's actually good at the game otherwise.

In theory somebody's capacity for teamwork would show up in their stats over time in some way like more wins or whatever.

[MechaBattler]

You know I never considered it as my capacity for teamwork. So much as my intuition on how the situation was going. When I was still playing in earnest. I found that I developed a sixth sense for moving into the right position relative to the team and the enemy. Now that I play more infrequent it doesn't come as naturally as it once did. It's not simply enough to work with the team, often enough the team will unwittingly work against you.

I joke you have to play like you're Liao, your team are Davrats, and the enemy are Clanners.

[Lightfoot]

I just don't like the current weapon and armor balancing. I should not take two SRM hits and lose all my weapons and then die in 2 seconds of small laser and MG firing. That is not MechWarrior and it is ridiculous balancing.

The worst of all is the difference between the Invulnerability=ON Light mechs and the insta-dead Medium mechs. Some Mediums are good, but most are made of tissue paper. I can't understand how both these mech Classes exist in the same game.

Above all MWO mechs have either not enough armor or hit boxes that are too big. What gives the Light mechs survivability is their small hit boxes and speed. Anyway the tier 2 tech made MWO's flawed Mech weaknesses much much worse. Most of my Clan mechs are pure garbage now. To have any fun and feel like I am driving a mech I have to go Inner Sphere. I don't really care which Faction does what, I just want the mechs to feel like actual Battlemechs.

As for the MM, I don't see a solution for that. QP will likely always have to draw on the full player base to move quickly. Best to rely on yourselves to organize the teams you end up with better. What you really need for consistent quality matches is to be on a Merc unit in Faction Warfare. Playing with a team you know is the most rewarding, but of course teammates are not always available and not all Merc units have the right combination of qualities. Then there is FW which pits me into Siege too often for my liking and lacks the notoriety and rewards you would expect. As a Lone Wolf, you get nothing really and I guess Clans can hire Merc Units in MWO which is forbidden in BattleTech. Oh well.

Fix my Clan mechs please... and my Thanatos and Shadow Cat.

Example: Arctic Cheetah vs. Black Lanner.

Which is faster? Arctic Cheetah.
Toughest? Arctic Cheetah
Most Damage? Arctic Cheetah
Most Agility? Arctic Cheetah
Jumping? Arctic Cheetah
Tonnage? Black Lanner

What's wrong with MWO's Mediums?

Edited by Lightfoot, 12 June 2018 - 02:00 PM.