27 replies to this topic

[Valore]

Helmer, on 24 February 2013 - 09:14 AM, said:

This post by Matthew Craig, Technical Director, explains why MM may have a few "wonky" matches stilll

http://mwomercs.com/...ost__p__1935843


"I wanted to just help clarify some points about the new match making system as there seems to be some confusion on these points:

Is the match maker taking weight class into account?

Yes the version currently live on production does take weight class into account. The way the match maker works it tries to group players within certain thresholds and grows those thresholds over time to ensure a match is made within currently a 2 minute time period.

These curves will likely see some tuning for both skill and weight as we monitor the data coming in and work to ensure more and more matches are created within good thresholds. This tuning will happen over the course of upcoming patches.

I'm still getting matches that seem unbalanced.

It will take time for the ratings to stabilize please bear in mind that this will likely happen over weeks not days. Also it's important to clarify that the system is only using your match history since we announced that we enabled it for pre-seeding. If you've been with us since closed beta your skill rating is essentially a few weeks old and only based on what you've done in that time. This is the only way to be fair to everyone.

Also currently after 2 minutes you may get an unbalanced match as the match maker increases the thresholds dramatically to ensure you get a match this is as designed to ensure that you don't wait longer than 2 minutes to find a match.

How does the match maker compose a teams Elo rating, is it average rating or closest to a target?

It's closest to a target value, so the match maker starts trying to make a match for an Elo of say 1300 and will pull in players to those teams closest to those values; however, as mentioned earlier within growing thresholds and those curves will be tuned. Currently it may be a bit 'sloppy' about how it's filling those buckets but over time it will be tuned to be much more precise.

We need to do this carefully over time as generally the cost of precision is time to find a match we want to slowly find a very nice balance between time to find a match and the number of matches that are correctly composed.

Can we expect match making to be tuned?

Yes as just stated we will monitor and tune over time to ensure it is working optimally and as everyone plays more games things will naturally settle out over time.

If your getting consistently hard or easy matches please bear with it while the system adjusts and know that we are monitoring and planning adjustments to ensure everyone gets fair matches."

Oh really? So there's an actual TECHNICAL problem with MM, and not me being punished by the mech gods because apparently according to people like SIXSIXSIX and Vassago we're a bunch of stat whoring, noob punching, sync dropping A-holes, and the problem is all our fault

Gee. I would never have known this. Thanks Helmer

Edited by Valore, 24 February 2013 - 09:18 AM.

[Glucose]

So the lighter tonnage side won the match, there were 4 deaths on both sides, and this is considered a bad matchup?

Who cares if the tons matched, it appears on paper to have been a decent match.

[Endbr1nger]

I look at this screenshot and see the opposite of what you do. This looks like a great match. A bunch of mediums and a smattering of heavies won a game against a bunch of assaults.

What are you complaining about exactly? That you actually had a close game? If you had been matched up with other assaults on your team with your high elo you would have won 8-0.

This is how MM is supposed to work. Mech tonnages don't matter, it's about close games, balanced between two good even teams or a good team against a heavier worse team.

Nice win.

[Helmer]

Valore, on 24 February 2013 - 09:17 AM, said:

Oh really? So there's an actual TECHNICAL problem with MM, and not me being punished by the mech gods because apparently according to people like SIXSIXSIX and Vassago we're a bunch of stat whoring, noob punching, sync dropping A-holes, and the problem is all our fault

Gee. I would never have known this. Thanks Helmer

I wouldn't say , and don't believe Mr Craig said, there is a "Technical" issue, more of a statistical one. With possible tweaks needed.

In my case, the grief mongers are correct in one aspect. I am an A-hole. Oh, and I do enjoy statistics , but I don't chase after them.




Cheers.

[SlXSlXSlX]

Helmer, on 24 February 2013 - 09:42 AM, said:

I wouldn't say , and don't believe Mr Craig said, there is a "Technical" issue, more of a statistical one. With possible tweaks needed.

In my case, the grief mongers are correct in one aspect. I am an A-hole. Oh, and I do enjoy statistics , but I don't chase after them.




Cheers.

So for clarification, MM, ie ELO, ie Elaine, does not guarantee you a perfect 50/50 rate, nor a perfectly balanced match every time, and will be tweaked.

Valore, on 24 February 2013 - 09:17 AM, said:

we're a bunch of stat whoring, noob punching, sync dropping A-holes, and the problem is all our fault

Glad were finally seeing eye to eye Valore.

[p00k]

Stoicblitzer, on 24 February 2013 - 08:10 AM, said:

lmao. i've seen that a lot. Need BV.

i always laugh when someone pokes in and says BV will make things better. between OP, and other examples
http://mwomercs.com/...03#entry1947503
http://mwomercs.com/...15#entry1949715
teams that are outtonned, and out-BV'd, come out victorious, sometimes dramatically so. BV didn't work well in tabletop, and would be worthless in mwo. we don't roll dice to aim, and mechs are not a sum of their parts, and more tonnage and more parts are not necessarily better.

Glucose, on 24 February 2013 - 09:21 AM, said:

So the lighter tonnage side won the match, there were 4 deaths on both sides, and this is considered a bad matchup?

Who cares if the tons matched, it appears on paper to have been a decent match.

this. 4 kills on each side, and while the winners won by base cap, and damage is pretty even when you consider there's less damage to soak up on the lighter team (and thus less damage needed to kill those same 4 ppl). not sure why OP is complaining

[Valore]

SlXSlXSlX, on 24 February 2013 - 09:51 AM, said:

So for clarification, MM, ie ELO, ie Elaine, does not guarantee you a perfect 50/50 rate, nor a perfectly balanced match every time, and will be tweaked.



Glad were finally seeing eye to eye Valore.

Back to working the system then. Won't be the end of the world running 2 X DDC 2 X StaLRMker for a few days until Elaine gives in or the entire system craps itself. Might be a relaxing time seeing how many 1337 pros in mediums the MM can throw at us, since apparently that's how its supposed to be working *rolls eyes*

p00k, on 24 February 2013 - 10:04 AM, said:

this. 4 kills on each side, and while the winners won by base cap, and damage is pretty even when you consider there's less damage to soak up on the lighter team (and thus less damage needed to kill those same 4 ppl). not sure why OP is complaining

Because balancing with a flawed ELO alone, and throwing most weight consideration out of the window is a terrible idea.

There's a reason why weight considerations were put in place in the first place. Now there isn't any weight checks, we'll just have to grind the system down until we produce enough bitter tears on the forums that PGI rectifies the issue.

IMHO, the ideal matching would be ELO * MWO customised BV to give a third value, and this third value is used to balance teams. That would probably be the closest we can get to a good system.

Edited by Valore, 24 February 2013 - 10:44 AM.

[Stoicblitzer]

p00k, on 24 February 2013 - 10:04 AM, said:

i always laugh when someone pokes in and says BV will make things better. between OP, and other examples
http://mwomercs.com/...03#entry1947503
http://mwomercs.com/...15#entry1949715
teams that are outtonned, and out-BV'd, come out victorious, sometimes dramatically so. BV didn't work well in tabletop, and would be worthless in mwo. we don't roll dice to aim, and mechs are not a sum of their parts, and more tonnage and more parts are not necessarily better.

Are you saying then that nothing should be done? You seem quick to laugh and criticize but offer no ideas. I think a partial solution that would go a long way towards preventing roflstomps and increasing coordination for pug/premade teams with the current MM system is integrated VOIP. This should be #1 priority imo. I know; different teams work on different things. Well they need to add more people to the "VOIP" team. Premades currently have a huge advantage over pugs that goes beyond the current MM system. 4 people using COM-2Ds on TS can focus fire a D-DC to death. The solution isn't to force pugs to join premades. I guess the current system is supposed to adjust for that as people in premades will typically have a much higher W/L ratio and likely a much higher Elo score. Eventually they will be put up against another premade with a similar score. I guess the wait time should be increased so that the MM system doesn't just say "**** it, throw some noobies in there."