268 replies to this topic

[Villz]

Kaijin, on 28 September 2013 - 04:17 AM, said:

Didn't read the entire thread so don't know if this has been suggested or not.

Roughly, 3 queues. 4v4, 8v8, and 12v12. Players can choose to queue up for 1, 2 or all 3. In the event a player queues up for more than one size drop, the matchmaker would first try to find matches for the smallest match size. High and low Elo payers would probably quickly get a 4v4 match. The 4v4 should probably be limited to the small maps.

A u didnt read the thread
B this is the patchfeedback forum not the advice suggestion forum wtf r u even posting that here for?

[DragonsFire]

Villz, on 28 September 2013 - 02:55 AM, said:

I would have helped but i am banned off comastar NS by ghostbear one of the bAdmins. Ps my group broke up for that so fu

ps check my account history we had alotttt of failed to find matches about 2 hrs ago

Chavette, End Crescendo, Delta Spectra and me

TY Mathew <3

Hey Villz,
You might have been filled in by Ryan on this, but when we did the tuning it was strictly solo drops. I would imagine group dropping would potentially entail other difficulties and perhaps further tuning would need to be done to help account for that, but that's just my best guess. Perhaps Matt can give us some feedback on that front?

[Chavette]

Yea, the tuning didn't cut it for us. I need to search for 20 minutes to get in a match... and when I'm in premade It takes 10 minutes.

[Corvus Antaka]

Chavette, on 29 September 2013 - 12:59 AM, said:

Yea, the tuning didn't cut it for us. I need to search for 20 minutes to get in a match... and when I'm in premade It takes 10 minutes.

I thought the timer only ran for 3 minutes max?

Big concern I'm seeing is that playing with guys new to the game and trying to help them is best done by not grouping with them.

Edited by Colonel Pada Vinson, 29 September 2013 - 01:53 AM.

[Chavette]

Colonel Pada Vinson, on 29 September 2013 - 01:50 AM, said:

I thought the timer only ran for 3 minutes max?

I counted in the times I have to re-que, Its about 3-4 times on average, plus the 1-2 minutes till I notice it stopped searching.

[retakrew7]

matchmaking is working for us now, definitely a good step in the right direction.

[Too Much Love]

Matthew, thanks for direct participating in the discussion. It's nice to see developer on this forum.

First of all I would like to say that, IMHO, after the last patch matchmaking quality in PUGs dramatically increased. Since then I hadn't long losing streaks (like 20 matches in a raw), when I was placed with the team of 11 not-so-good-players. I don't say that MM is perfect, but it became better.

Don't know if you tweaked MM or not, but I noticed difference in matchmaking this saturday/sunday. Saturday it was much more balanced, sunday it was very striped: win-lose-win-lose etc.

I want to tell in my opinion very important thing about matchmaking: it should give player a feeling that his/her actions matter. That he is not a toy in the hands of Doom (matchmaker), but a player, who affects the outcome of the game. Unfortunately, most of the time I don't have such a feeling. Just the opposite, every match I see that I'm placed either in the Good Team, or in the Bad Team.

Time after time today when I pugged I witnessed how some 4 men team won and in the next match they were placed with the weak teammates and lost. Matchmaker just switched the sides with the weak gamers and the ending was quite predictable. After every victory matchmaker gave me 90% losing variation of teammates.

I mean matches should be competitive, not "equal" (in terms of even number of victories and defeats). Hope you will find this useful in your work.

[Corvus Antaka]

Chavette, on 29 September 2013 - 02:04 AM, said:

I counted in the times I have to re-que, Its about 3-4 times on average, plus the 1-2 minutes till I notice it stopped searching.

yeah ok I figured that might be it.

I've noticed it seems more extreme with high KDR lights or assaults than the other wieght classes, with my assaults and my high KDR lights I can sometimes still fail 2-3 in a row though it seems rarer the last few days.

I've noticed a bigger issue grouping with new players, they seem to get an elo pairing as a lance that is not conducive to the new player experience.

[warner2]

Colonel Pada Vinson, on 29 September 2013 - 09:10 AM, said:

yeah ok I figured that might be it.

I've noticed it seems more extreme with high KDR lights or assaults than the other wieght classes, with my assaults and my high KDR lights I can sometimes still fail 2-3 in a row though it seems rarer the last few days.

I've noticed a bigger issue grouping with new players, they seem to get an elo pairing as a lance that is not conducive to the new player experience.

The ELO of a group is average yeah.

What else is the match-maker going to do? Tough on the new player I guess. On the bright side he's got three high ELO mates backing him up right?

[Chronojam]

Colonel Pada Vinson, on 29 September 2013 - 09:10 AM, said:

yeah ok I figured that might be it. I've noticed it seems more extreme with high KDR lights or assaults than the other wieght classes, with my assaults and my high KDR lights I can sometimes still fail 2-3 in a row though it seems rarer the last few days. I've noticed a bigger issue grouping with new players, they seem to get an elo pairing as a lance that is not conducive to the new player experience.

Elo is not based on KDR as far as my understanding of it goes.

[Deathlike]

Chronojam, on 29 September 2013 - 11:00 AM, said:

Elo is not based on KDR as far as my understanding of it goes.

Tends to lean towards Wins, but I think also how the win is achieved (it's hard to figure out how useful match score is actually).

Edited by Deathlike, 29 September 2013 - 11:04 AM.

[DirePhoenix]

New players start off in the middle of the Elo range, not at the bottom. Their performance over time will move them above or below that middle rating. New players with a mid-range Elo ratings don't have issues getting matches as there are significantly more mid-range Elo players than extreme high- and low-end players.

[Almeras]

I found the ELO system much worse. Tonight was by far the worst I have seen it. My 4man were repeatedly dropped against one or the other of the same two 4mans for the whole of our session.

One we could just about beat 1/3 times depending on the rest of our team, while the other we never stood a chance against and at one point got them 3 times in a row.

None of the games in the whole 2hrs would be considered fun. Or even fair it just felt like the ELO decide to serve us up for a beating which left a very sour taste.

It feels like the match maker wants to grind people down vs having fun.

[Literally]

Matthew Craig, on 26 September 2013 - 12:39 PM, said:

I'd be happy to but not for a while, as you can imagine there are changes coming to the match maker to support community warfare so this is probably best discussed after that work is in place.

Since Community Warfare has been pushed back at least another 6 months again (announced at launch event), wouldn't it be OK to tell us how it works for now?

[Xandre Blackheart]

Still roflstomps.

The Roflstomp train will not stop.

ALL ABOARD!

Edited by Xandre Blackheart, 29 September 2013 - 10:47 PM.

[FREDtheDEAD]

Putting new players at medium ELO is like categorising L plate drivers on their first lesson as average drivers. No one is going to get any worse than they are when they start a game as tactically complex and reliant on game-specific hand-eye skills as MWO.

What the hell do you have to do to get a LOW ELO? Mount a 140 standard engine on an assault mech, strip off its armour and weapons, then run into the middle of the enemy line with a fruit knife???

Proposed new weapon: Fruit Knife. Allows you to damage the camo on an enemy mech and just scratch it up, halving resale value.

Edited by Xajorkith, 30 September 2013 - 12:34 AM.

[Corvus Antaka]

Chronojam, on 29 September 2013 - 11:00 AM, said:

Elo is not based on KDR as far as my understanding of it goes.

It has something to do with if X kills Y X goes up Y goes down add infinitum, thus KDR does have a direct relation to ELO rankings iirc.

Edited by Colonel Pada Vinson, 30 September 2013 - 12:38 AM.

[Wintersdark]

Colonel Pada Vinson, on 30 September 2013 - 12:38 AM, said:

It has something to do with if X kills Y X goes up Y goes down add infinitum, thus KDR does have a direct relation to ELO rankings iirc.

No, Elo rankings are based solely on winning and losing. How much it changes depends on the relative team Elo scores (and thus how likely the matchmaker feels your victory or defeat is) but kills has no bearing at all on it. Other than, obviously, being a contribution to winning of course.

Thus, if you play on a team that wins most of its matches, your personal Elo score will be quite high - and equally so - even if you never kill anyone.

The math behind it is shown in the Elo command chair post.

In short, its purely a matter of winning and losing, though.

[Arete]

What wintersdark said^^

If two groups of 1500 rated player meet about, the winning team gets 20+ rating, losing team gets -20.

If a 1000 group meets a 2000 group and the 2000 group wins (which is expected), the 2000 group probably won't get any + at all, and the 1000 group won't lose any rating. And vice versa, in the unlikely event that the 1000 rated group wins, they can get +200 and the 2000 group -200.

Then they have some nifty modifiers for different ratings within a group aswell, so you gain/lose different amounts depending on the internal rating of the group.

Kills mean nothing for your rating.

(All numbers are pure speculation, we don't know much about the exact math)

[Arete]

Almeras, on 29 September 2013 - 03:53 PM, said:

I found the ELO system much worse. Tonight was by far the worst I have seen it. My 4man were repeatedly dropped against one or the other of the same two 4mans for the whole of our session.

One we could just about beat 1/3 times depending on the rest of our team, while the other we never stood a chance against and at one point got them 3 times in a row.

None of the games in the whole 2hrs would be considered fun. Or even fair it just felt like the ELO decide to serve us up for a beating which left a very sour taste.

It feels like the match maker wants to grind people down vs having fun.

I've found that group matches seem to be wonkier than solo drops when it comes to matching. It's kind of hard to know why though, since we have no idea what the actual rating is for different players in a group. But judging from other games I've played with Elo matchmaking where the player rating is visible the match enjoyment usually goes down (or at least varies) the bigger the rating difference is in the group. It's just theorycrafting but unless I'm reasonably sure that all members of a 4-man drop is at roughly the same rating I kind of expect the match to be unbalanced.