129 replies to this topic

[Gerwig]

My experience has been great since the PSR implementation, tier 3, usually it seems I get placed with the 1s and 2s, overall a much better experience, communication and coordination between team more often. There are still stomps but not as much as before.

[JigglyMoobs]

Here are some of my observations playing over the last month:

MM seems to be still keeping track of some additional skill ratingi in addition to psr.

Win-loss can be streaky and dependent on the time you are playing at. During certain times of the day the player base on t2 is, shall we say, challenged and if MM expects you to carry, woe be onto you.

In carry mode the old carry mode tactics apply. Put yourself in the middle of the pack, study team positions more than enemy positions on the tac map, don't be the first one to peak and conserve health and armor for the later game. However if everyone is in carry mode team is ducked.

Number one game winner for me by far: voice comms. I play from my university office so sometimes I can talk and some times not. I win far more when I can talk on comms with the team. A "chatty" team usually stomps. The correlation between talking and winning has been pretty obvious.

[JigglyMoobs]

Gerwig, on 23 October 2015 - 11:56 PM, said:

My experience has been great since the PSR implementation, tier 3, usually it seems I get placed with the 1s and 2s, overall a much better experience, communication and coordination between team more often. There are still stomps but not as much as before.

Gerwig, it -is- great in t3, but then when you move into t2 it becomes a total crap shoot. Half the time teammates are much worse. I think the problem is that when you are in t2 MM expects you to carry more and more often gives you really bad teammates if you happen to be playing at a time with low population.

Edited by JigglyMoobs, 24 October 2015 - 02:32 AM.

[El Bandito]

JigglyMoobs, on 24 October 2015 - 02:32 AM, said:

Gerwig, it -is- great in t3, but then when you move into t2 it becomes a total crap shoot. Half the time teammates are much worse. I think the problem is that when you are in t2 MM expects you to carry more and more often gives you really bad teammates if you happen to be playing at a time with low population.

Then what is this T3 Elo hell people keep whining about, if T3 is great to be in? Do they just suck?

Edited by El Bandito, 24 October 2015 - 02:34 AM.

[JigglyMoobs]

El Bandito, on 24 October 2015 - 02:33 AM, said:

Then what is this T3 Elo hell people keep whining about, if T3 is great to be in? Do they just suck?

I don't know but t2 feels worse. T3 seems more even, better team work at least the first couple of days I was t2.

Then again I got in t2 around the same time the wolfhound dropped so that could have something to do with it. Everyone was trying to find miniature 1v1 matches all over the map, cheetahs were pursued with gusto and the herd of cats was impossible to keep together.

Edited by JigglyMoobs, 24 October 2015 - 02:41 AM.

[JigglyMoobs]

Just checked my stats, which perfectly reflect my matches under psr since I took a break for a year.

Surprisingly, my current and archived stats are almost exactly the same.

Kd: 2.14 v 2.11
W/L: 1.56 v 1.36
Cbills: 102463 v 90722
Exp: 900.1 v 718.3

So actually my stats improved a little despite games feeling more random.

[Water Bear]

Duke Nedo, on 22 October 2015 - 12:18 AM, said:

...and science and proper statistical treatment or huge data sets proved them wrong. I am just saying one should have an open mind and not jump to conclusions just because one think that there is a correlation among a shetload of other contributing factors...

Here are some nice examples:





My main point is that one player out of 12 players on the team slightly under-performing (I say slightly because if you got to Tier 2 there are limits how bad you can be and as Tier 2 you can't be expected to carry the whole team in most cases anyway) would not be enough to drag a team down so horrendously that you lose 83% of your matches. The way I see it, it's just a very unlikely explanation, there are many other more probable explanations because I think that 11 players + 1 AFK player will do better than 17% win rate.

One could argue that your success depends almost entirely on the 23 other people in the game and the map selection.

I think you need to consider the strength of evidence here. The assumption we are making that the appearance of the OP in a game causes failure is a much more justified one than the assumptions drawn from the image examples you posted.

From a scientific point of view, if the only variables you can change are the players in the game and the map before running your experiment (that is, observing the outcome of the match), and the outcome is over 80% failure, you have a very strong scientific reason for concluding that the one fixed constant really is causing failure.

I'm not making a funny argument here. This is basically how you set up a scientific experiment.

Edited by Water Bear, 24 October 2015 - 06:56 AM.

[Duke Nedo]

Water Bear, on 24 October 2015 - 06:52 AM, said:

From a scientific point of view, if the only variables you can change are the players in the game and the map before running your experiment (that is, observing the outcome of the match), and the outcome is over 80% failure, you have a very strong scientific reason for concluding that the one fixed constant really is causing failure.

I'm not making a funny argument here. This is basically how you set up a scientific experiment.

Yes, you can set up a scientific experiment like that, but from that experiment alone you cannot correctly conclude that the OP is a bad player. Scientific method doesn't work like that. For what it's worth, I'm a scientist.

[Ted Wayz]

El Bandito, on 24 October 2015 - 02:33 AM, said:

Then what is this T3 Elo hell people keep whining about, if T3 is great to be in? Do they just suck?

Matchmaker didn't change so why would you think the quality of matches would change overall? Seems like the status quo has remained the same but people are attributing any change to PSR. But there is no evidence PSR has affected matchmaking. So yes, people are essentially saying matchmaker is giving good and bad matches like it always has.

What we do know about PSR is that new players get seeded into high T4. So if PSR does have an effect it would be more evident here.

[Oncoshi]

Kristian Radoulov, on 20 October 2015 - 06:40 PM, said:

They need to tighten up the tier restrictions:

5 with 4 only
4 with 5 and 3
3 with 4 and 2
2 with 3 and 1
1 with 2 only

20 min searching guaranteed.

[JigglyMoobs]

IMHO maybe ther solution is opening elo up for voting by other players at the end of a match. this could be used as an additional input into match making system.

since mm will be determined by elo, there is extra incentive to correctly vote up or down people who are good or bad. people also tend to be better judges of skill than computers so mm will tend to get better data. there might be some vote-griefing but that will just be in the noise once there are enough votes.

the other thing that mm could do is show you the distribution of skills on the queue at a particular time, and also show the median and average skills on both teams once you get in gane. so you'll see how much you have to carry and could plan accordingly.

finally, if there is large skill disparities in the team, i think that people should be able to see who the "elders" are on the team so the less skilled will know who to follow.

Edited by JigglyMoobs, 24 October 2015 - 11:15 AM.

[Simbacca]

JaxRiot, on 23 October 2015 - 07:43 PM, said:

MM gave me a stunning beating today.

Ive been doing well all week in Tier 4. Getting good scores and performing well in all of my mechs.

But today I hit Tier 3 and the very first match I got devastated. It was as if I had no armor at all and was shooting rainbows. These guys were hitting me like freight trains with impeccable aims. Get shot two times and half my HBR would be gone. It was amazing the difference it was from Tier 4 to Tier 3.

And I was being cautious. I knew Tier 3 was going to be tough and Ive never been the Rambo rush in style player so I was trying to be as tactical as I could. Not exposing myself and picking my shots and all that. It didnt help at all.

Of course me and my PSR get get pummeled back down to Tier 4 (thankfully) where I thought things would get back to normal for me. But nope. MM wasnt done with me and Ive been getting woefully and impressively stomped all day.

I seemed to have angered the MatchMaking Gods

Just tonight whilst playing the Cicada-3F(L) had 7 losses in a row.
3 losses = balanced teams
1 loss = Tourmaline - none of our teams actually engaged. Our base was capped. Total match damage was less than 600 for everyone combined. -- A unique way of losing, but no complaints.
3 losses = My team was murdered.

Though constant losses and their nature seems unfortunately a usual occurrence for me.

[Thorqemada]

Rule of Thumb (from my experience in Tier 3):

When is use my Jenners F/S, the Locust P or my Mediums (Cicada 3M, HBK F, SHD P) i lose and i lose often and i lose big regardless how good or bad i play - PSR shrinks slightly.

When i use my Heavy (Thunderbolt P) or Assault (Battlemaster P) i win slightly more than i lose pretty regardless how good or bad i play and my PSR rises up slowly (sometimes rises up big time).

Edited by Thorqemada, 24 October 2015 - 08:52 PM.

[Karamarka]

JigglyMoobs, on 24 October 2015 - 02:32 AM, said:

Gerwig, it -is- great in t3, but then when you move into t2 it becomes a total crap shoot. Half the time teammates are much worse. I think the problem is that when you are in t2 MM expects you to carry more and more often gives you really bad teammates if you happen to be playing at a time with low population.

I'd agree with this.

I have a screenshot of someone using a trial mech because they are new in my game.

Also, so many people doing less than 100 damage. Insane. How can this be ?

[Ex Atlas Overlord]

Water Bear, on 24 October 2015 - 06:52 AM, said:

From a scientific point of view, if the only variables you can change are the players in the game and the map before running your experiment (that is, observing the outcome of the match), and the outcome is over 80% failure, you have a very strong scientific reason for concluding that the one fixed constant really is causing failure.

I'm not making a funny argument here. This is basically how you set up a scientific experiment.

Except no "scientific experiment" would ever work that way.

What you'd ACTUALLY do is set up 23 constants.... THEN if you added the 24th as the OP and the failure rate changed.... THEN you'd have a result worth mentioning.

Anyone that submitted results that said "Hey we decided to see if this thing affected this other thing, but we didn't control for 23 of the 24 variables" would be laughed out of the room.

Edited by The Atlas Overlord, 24 October 2015 - 09:59 PM.

[JigglyMoobs]

Actually they do that all the time. You simply can't control that many parameters in the real world.

Instead, you control -for- them.

So for the op, we'd have to examine his games to see if the other players on his team performed statistically as teammates normally do for t2 players. If there are no deviations from normal t2 expectations, then we might say that the op is underperforming in t2.

If his teammates consistently underperforming, then we might say that op is unfairly burdened by MM.

If his teammates outperform, then we might say that op is possibly disastrous for his teammates.

(I have a PhD in science)

[JigglyMoobs]

PS - my advice to op:

STOP WORRYING ABOUT MM AND START TALKING WITH YOUR TEAMMATES ON VOIP.

Talking is the single biggest game winner in solo queue right now. It is totally OP if several of you start communicating.

Don't worry about saying the wrong thing or giving the wrong info. Just do your best and don't be an a-hole. The important thing is that once you start talking more teammates will feel comfortable talking too, and a minute into the match you'll have good communication established.

I played about 8 games this afternoon, only lost one or two.

Closest match was on Viridian. I got caught out early and died. Both teams were neck and neck. At the end the few survivors on our side used comms to coordinate perfectly and eek out a win.

Best game was on Tourmaline. We started talking early. I saw an isolated king crab and two friends on tac map and told team to pounce. Team responded! King crab was mince meat. Then our guys in the rear were in trouble, called out for help and team responded! We ate up two of their lights in a cross fire.

Finally remaining enemies nascared around and started pushing up the hill. I took a peak to see if I could get favorable trades and that was a big NOPE! Radioed team to push them from the flank and team did! As soon as they turned those of us at their front advanced and caught them totally confused. Executed a perfect pincer.

It was of course a stomp. At the end guy on other team complained about NASCAR and bad teammates. As solo games go his team was actually pretty good and coordinated. He had no inkling how much our coordination over matched theirs.

Start talking people. Talking doesn't guarantee you a win but it will, with high probability, give you a less frustrating, more enjoyable match.

Edited by JigglyMoobs, 25 October 2015 - 12:19 AM.

[Ex Atlas Overlord]

The Atlas Overlord, on 24 October 2015 - 09:49 PM, said:

Anyone that submitted results that said "Hey we decided to see if this thing affected this other thing, but we didn't control for 23 of the 24 variables" would be laughed out of the room.

The Atlas Overlord, on 24 October 2015 - 09:49 PM, said:

but we didn't control for 23 of the 24 variables" would be laughed out of the room.

The Atlas Overlord, on 24 October 2015 - 09:49 PM, said:

didn't control for 23 of the 24 variables

The Atlas Overlord, on 24 October 2015 - 09:49 PM, said:

didn't control for

JigglyMoobs, on 24 October 2015 - 11:37 PM, said:

Actually they do that all the time. You simply can't control that many parameters in the real world.

Instead, you control -for- them.

Really?

JigglyMoobs, on 24 October 2015 - 11:37 PM, said:

(I have a PhD in science)

Maybe you should re-take reading comprehension 101

[JigglyMoobs]

The Atlas Overlord, on 25 October 2015 - 12:17 AM, said:

Really?

Ha-ha did not see that. The first part of your post lost me.

You didn't describe control for in the right way. Setting up a constant in the way you described is to -control- a variable, not to control -for- it. One of the ways of controlling -for- a -set- of variables is to calculate a statistical score that you can use for comparison, but, its also not the only way.

You can also control -for- a variable by -controlling- the variable, but again, not the only way or the preferred way as the rest of your post -implied-.

Prior post you were replying to had some problems, but so does your preferred experimental plan. Most importantly, its not really practical.

Edited by JigglyMoobs, 25 October 2015 - 01:08 AM.

[Duke Nedo]

Methods vary between fields, in some fields you can isolate variables, in others you have no choice but to work with complex processes. In this case, the sample of 100 games is just to small to say anything.

Matches go up and down, see for example: http://mwomercs.com/...omm-is-working/

By selecting the worst stretch of 100 games from this graph you could end up with a pretty bad result.