Mm Hates Me (Or Is There A Black List Of Players) ?
#101
Posted 13 August 2016 - 11:57 AM
Quite the transformation
#102
Posted 13 August 2016 - 08:01 PM
MrJeffers, on 13 August 2016 - 10:41 AM, said:
News flash - for every person who wins, one person will lose. So the global average must equal a 1:1 ratio.
The important part is the percentage of people that are at or near a 1:1 ratio. A perfect match maker would have 100% of the people at a 1:1 ratio, people are always playing their equals.
A good match maker would follow the 80/20 rule, where only the top 10 and bottom 10 percent is not close to that 1:1 ratio.
Putting it on a 5 point scale it would look like this:
5 - Perfect - 0% outside of the 1:1 range
4 - Good - <25% outside of 1:1 standard deviation range (top and bottom 12.5%)
3 - Average - <50% outside of 1:1 (top and bottom 25%)
2 - Poor - >75% outside of 1:1
1 - Awful - 100% outside
Using lost dragons numbers:
Bottom 22% are getting crap matches, likewise there would be >22% top scores that are getting matches below their skill level. The top percentage is larger because the average numbers are greater than 1:1 so there are more winners than losers, caused by eliminating people with less than ten matches, who must have more losses then wins.
So lets say that the top is 23%. That means for 45% of the players, the match maker is failing. That is on the poor side of average on the above scale. That's hardly something commendable, and likely could be done with a random match maker that didn't use any rating. Nearly half the players are getting crap matches.
I would bet real money that if we had the same set of numbers from the elo based match maker the results would be higher, nearer to good but defiantly on the good side of average and probably in the high 60% to 70% range.
Just because 22% of people are at .75 w/l or lower doesn't mean they are having bad matches. We all play with the same match maker, the people that have bad w/l have it because they are bad. The people who have good w/l have it because they are good.
As I explained several times previously you can't just have that lower 22% of the population play each other exclusively because of wait times. That is only a little over 7,000 people in time zones all over the world playing all different weight classes of mechs. If you only allow them to play people in that same group to try to push them closer to a 1:1 w/l they will wait several times longer to find a match than it takes to play a match most of the time. I would also argue that the people who are at a .5 or less w/l are the ones really hurting, but they make up less than 5% of the population, consisting of only about 1500 individuals, so trying to create a safe space for them to play only each other is just a pipe dream.
What you are asking for is absolutely not feasible. I doubt the results are going to be much different with ELO due to the nature of the game and the need to create matches in a reasonable amount of time.
#103
Posted 14 August 2016 - 08:20 AM
Lostdragon, on 13 August 2016 - 08:01 PM, said:
Just because 22% of people are at .75 w/l or lower doesn't mean they are having bad matches. We all play with the same match maker, the people that have bad w/l have it because they are bad. The people who have good w/l have it because they are good.
Actually it does mean just that - it means more often than not 43% the people are being matched against players that are better them them (losing record) or players that are worse then them (winning records). That does mean the match maker is failing at creating matches with opponents at your skill level. "Good" players should play "good" players, and "bad" players should play "bad" players. The problem is with PSR it cant tell. There are plenty of "bads" that are tier 2 or tier 1. MM considers them "good".
The people in the bottom shouldn't be playing against the people in the top, but because their PSR is close or the same they end up in the same matches.
Quote
As I explained several times previously you can't just have that lower 22% of the population play each other exclusively because of wait times. That is only a little over 7,000 people in time zones all over the world playing all different weight classes of mechs. If you only allow them to play people in that same group to try to push them closer to a 1:1 w/l they will wait several times longer to find a match than it takes to play a match most of the time. I would also argue that the people who are at a .5 or less w/l are the ones really hurting, but they make up less than 5% of the population, consisting of only about 1500 individuals, so trying to create a safe space for them to play only each other is just a pipe dream.
What you are asking for is absolutely not feasible. I doubt the results are going to be much different with ELO due to the nature of the game and the need to create matches in a reasonable amount of time.
I completely understand that there are population issues. And what I described previously is in no way infeasible because I didn't propose any changes to the match maker in that post. I just assigned a rating to it. What was infeasible about it?
Sure the lower 22% can't always play against themselves. But they shouldn't be playing against that top 23% if at all possible. The bottom 22% should be playing themselves and the other 55% to at least give a sporting chance. Likewise the upper 23% should be playing themselves and the 55%. The problem is PSR doesn't allow that because players with vastly different skill levels have PSR ratings that are same, something that doesn't happen an in elo system. Some, if not most, of those "bottom" 22% are likely Tier 3, tier 2, or tier 1 players who should not be. And if you go back and re-read my note, I only said an elo based system would be about 10-15% better, because I am fully aware of the population problems.
Edited by MrJeffers, 14 August 2016 - 09:10 AM.
#104
Posted 14 August 2016 - 09:49 AM
MrJeffers, on 14 August 2016 - 08:20 AM, said:
The people in the bottom shouldn't be playing against the people in the top, but because their PSR is close or the same they end up in the same matches.
I completely understand that there are population issues. And what I described previously is in no way infeasible because I didn't propose any changes to the match maker in that post. I just assigned a rating to it. What was infeasible about it?
Sure the lower 22% can't always play against themselves. But they shouldn't be playing against that top 23% if at all possible. The bottom 22% should be playing themselves and the other 55% to at least give a sporting chance. Likewise the upper 23% should be playing themselves and the 55%. The problem is PSR doesn't allow that because players with vastly different skill levels have PSR ratings that are same, something that doesn't happen an in elo system. Some, if not most, of those "bottom" 22% are likely Tier 3, tier 2, or tier 1 players who should not be. And if you go back and re-read my note, I only said an elo based system would be about 10-15% better, because I am fully aware of the population problems.
T1 and T5 are supposed to be excluded from one another so that is already built in to exclude the top and bottom from playing each other. I see anecdotal reports on the forum all the time that they are not but I've yet to see any hard proof. If T1 and T5 are in fact playing against each other I imagine it is after many release valves have been triggered to try to make a match, which again goes back to the fact that if you want timely matches you have to play with a wide skill range, there is just no way around this.
You also fail to take into account many other factors. Maybe the people that are losing a lot are running mechs like Commandos and Mist Lynx a lot. Maybe they are running mixed builds. Maybe they don't care that they lose a lot and would rather play how they want than run a meta mech that would help them win more often.
I am sure there are improvements that can be made to PSR, but I don't think there is anything you can do that will move the bar in the way you want to get everyone closer to 1:1 w/l when the vast majority of the population is already there. Better skill level classification only goes so far and will still run into the population problem.
#105
Posted 11 September 2016 - 05:22 AM
Have fun, I say and the rest will fall into place nicely and eventually.
#106
Posted 11 September 2016 - 06:17 AM
#107
Posted 11 September 2016 - 12:19 PM
Mister Blastman, on 12 August 2016 - 09:00 AM, said:
Ah-hah! Thank you for that! Interesting that the average w/l is 1.09 but I believe this figure is the result of Group Queue being combined with PUG Queue stats.
Group Queue naturally skews the win/loss ratio above 1:1 (provided you have decent teammates).
This is a big reason why Solo and Group stats should be separated for leaderboard purposes.
You guys are ducking idiots! 12 win and 12 loose every match, of course the average will be 1:1! !!!!
How stupid can anyone be.
EDIT: Not who i quoted, quoted for his logic in above 1
Edited by Bobzilla, 11 September 2016 - 12:28 PM.
#108
Posted 11 September 2016 - 12:39 PM
#109
Posted 11 September 2016 - 12:40 PM
Bobzilla, on 11 September 2016 - 12:19 PM, said:
How stupid can anyone be.
EDIT: Not who i quoted, quoted for his logic in above 1
But it isn't, genius. It's 1.09:1. Something's wrong.
#111
Posted 11 September 2016 - 01:01 PM
Edited by Pilotasso, 11 September 2016 - 01:02 PM.
#112
Posted 13 September 2016 - 10:11 AM
EDIT: I did just read a bit further into the thread and noticed the posts about basically giving up and losing the will to fight on because it feels pointless. Here's how I cope with it: If we lose, as long as I did personally well I am happy. We lost but I got 900+ damage and 4 solo kills? Well, we lost, but damn I did good. So I'm happy about it. Perhaps if you realize you are losing look to personal gain as an incentive to still continue to do the best you can?
Edited by Mole, 13 September 2016 - 10:14 AM.
#113
Posted 13 September 2016 - 10:13 AM
PORONOPAPOS, on 13 September 2016 - 09:35 AM, said:
My main account is T1 from last December (when Tier system was introduced I was placed at start of T2). Many say that tiers are only XP bar and eventually all will reach T1 by playing for long time. OK maybe but I went from T2 to full T1 having all this time W/L ratio less than 1, there is only one way to do this while losing.
My alt account is currently half T2. I created last June just to see if my bad luck continues, matches there are quit different although I meet mostly the same known T1 players as in my main account. I can't say much about T4 matches It took my 5 or 6 matches to get to T3 (playing only trial mechs).
So you have different PSR ratings on this two accounts? This explains it all.
I don't know, I think, that PSR being just XP bar - is just a myth, created by PGI in order to avoid upsetting players, who were thinking, that their rating was much higher, than it really was. Also may be another faulty assumption was made by PGI - they assumed, that absolutely all players can play at exactly the same level of skill - they just need to get more experienced. I.e. they assumed, that there is no physical skill cap, player just can't overcome. I.e. PGI's statement meant "Play more - get more experienced - and your PSR will rise".
At the end current MM is pretty simple - it tries to keep your W/L = 1 and average match score around 200. Ladderboard analysis confirms it. Despite of what PGI say, W/L is out of your control - it's MM, that decides, which of two teams should be stomped. So, only average MS - is what depends on you. Yeah, you may assume, that it's very easy to earn 200MS. But what all players seem to forget: is that the higher your rating is - the higher skill your enemies have - the harder to obtain the same 200MS. So player is assumed to stabilize at some point. For example my rating fluctuates between 75-95% Tier 3.
Most terrible thing about current PSR - bias toward increasing. It can take 100500 loses in a row to drop your rating just a little bit. But then all of a sudden after 100500 loses in a row your start to have decent matches. But not for so long... As your rating increases much faster, than drops - 3-5 decent matches and you gain 30% PSR again. And say hello to loses again. In order to see, how terrible current MM is - see my ladderboard position. I'm on my losing phase again.
And despite of stats, being much lower, than needed W/L = 1 and MS = 200 - my PSR just refuses to drop. I'm still at exactly the same 90% Tier 3. As you know, if I would deal 1K dmg - my rating would rocket jump immediately. 3-5 matches - and I would be in Tier 2. This happens every time, when I use my LRM60 boat to finish even as fast, as possible. But dropping back isn't so easy. Yeah, that's how fair PSR system is. Once gained in good 'Mech - rating just can't be dropped.
There is no "Very large drop" here. Also gaining 100MS - is pretty easy. PGI give you 20 for free just for victory. Yeah, it's double W/L dependence.
So. MM is terribly broken. And PGI simply ignore this fact and refuse to fix it. The only option for you to deal with this problem - just to quit this game and find some other game, developers care about. Overwatch may be?
Edited by MrMadguy, 13 September 2016 - 10:24 AM.
#114
Posted 13 September 2016 - 12:18 PM
PORONOPAPOS, on 13 September 2016 - 12:08 PM, said:
Since you mention the same known Tier 1 players, are the results related to which side they are on or whether one side just has more of them? The goal of MM is to use PSR as a skill metric, when it is not. Could it be that you are now seeing the results of newer players climbing up into tiers they should not be in, and then having them often end up on your team?
#115
Posted 13 September 2016 - 01:52 PM
PORONOPAPOS, on 13 September 2016 - 01:39 PM, said:
I can't be sure but the feeling is that some times top players (what may call them "tier 0") fall in the same team while the other team may gets "generic" T1/T2 players who as you describe are "newer players climbing up into tiers they should not be in (at least yet I would add)"
I'll give you an example from a match I had last month in tourmaline desert map : I was playing with a high alpha Mad Dog and moved alone in a flanking behind enemy line. So I went unoticed behind a rifleman, a Ebon and another IS heavy or assault (don't remember). They were shooting my teammates while standing still close to each other. I shot one alpha in Ebon's back then some chain fire and killed him without any reaction from him. I did the same with the rifleman, still no reaction, then I shot the third one's back and he turned only to face a good alpha. I got 3 easy kills, but I'm sure no experienced player no matter his Tier would stay and die with no reaction. So the question is how these gus were in a supposed high Tier match ?
You very likely have a good point with this topic. Use the complaint button if players on your team are doing less than 100 damage and things like that and also try to swing the matches your way if possible.
If the trolls smell blood that just sends them into a frenzy and there most definitely is trolling going on in game especially at times when not a lot of players are on and the queues can be full of guys on team speak on both teams.
Either way consider it hard mode when in a tough match and an opportunity to up your game. This is not an easy game to master though, and a good score is good, win or lose.
Also, epic scores are not possible on a good team.
Edited by Johnny Z, 13 September 2016 - 02:11 PM.
#116
Posted 13 September 2016 - 01:55 PM
PORONOPAPOS, on 12 August 2016 - 06:03 AM, said:
Not trying to be rude, I've seen you in QP matches the last few days. Most of the time you are not with your team, even in an assault. The bulk of your team is grouped up 2-3 map squares from where you are hiding/sniping/whatever. Either way, totally outta position, swarmed, die.
It's a different story if you're in a light/med that can relocate quickly easily or relying information to the team/harrassing etc. In the mechs you were in, you can't afford to be alone.
In an alt-account T3-5, that style of play might work (I've never run an alt, so I don't recall how different the play is), you'll be able to deal a lot of damage etc. But throw a few T1/knowledgeable players into the enemy team that work as a team and you're going to have a bad day not sharing assault armour and attempting to be the last mech standing etc.
#117
Posted 13 September 2016 - 01:59 PM
justcallme A S H, on 13 September 2016 - 01:55 PM, said:
Not trying to be rude, I've seen you in QP matches the last few days. Most of the time you are not with your team, even in an assault. The bulk of your team is grouped up 2-3 map squares from where you are hiding/sniping/whatever. Either way, totally outta position, swarmed, die.
It's a different story if you're in a light/med that can relocate quickly easily or relying information to the team/harrassing etc. In the mechs you were in, you can't afford to be alone.
In an alt-account T3-5, that style of play might work (I've never run an alt, so I don't recall how different the play is), you'll be able to deal a lot of damage etc. But throw a few T1/knowledgeable players into the enemy team that work as a team and you're going to have a bad day not sharing assault armour and attempting to be the last mech standing etc.
I played against his team the other day and it wasn't pretty, no one player could have done very well in that match. He is most likely doing fine. Takes quite a while to do really well in this game.
I use a light mech most of the time lately anyway, but a light mech is the easiest mech to swing a match in also. Because I can be on the other team before they have a chance to do much of anything.
Stick with the same mech until mastered and fully modified(moduled), perfect the load out and practice until its second nature for good scores.
I had my fresh Locust taken out at the start of 2 matches yesterday. Once from an arty strike, totally fresh and another from a LRM volley. Had a couple excellent matches to.
Edited by Johnny Z, 13 September 2016 - 02:05 PM.
#118
Posted 13 September 2016 - 03:07 PM
Your data confirms this. You play T1, your entire team dies horribly. You play T2, where the enemy is less skilled, your mistakes dont have as big an impact.
The snowball effect is strong. 12 vs 12 turns into 12 v 5 very quickly if even one person does something dumb, and waste's his tonnage.
As an example from my epic T5 greatness (LoLz). My 2 favorite mechs are both huge alpha monsters. The Vulture (Maddog) with SRM spam, and a Masakari (Warhawk) built with a huge alpha.
My vulture, I try to flank, It's fast and 6 srm sixes will one shot most mechs if I can get behind them while they are tunnel visioned on the mass of my team. If my flank is succesful, I get a good one shot on a marauder sniper, and take it out, then get some damage out on whoever sees me, the rest of the match is up to my team, I did a fair trade or better. If I can get multiple kills through this flank attack, it becomes a roll. If I'm spotted and focus fired without even getting a shot off, it's pretty much a guaranteed loss, just like when you start a match with a player discoed.
Same deal in my alpha Masakari. If the first guy I charge goes down quick and I can get more damage on even more mechs before they finally kill me, My team can usually win it. If I mange to take out 3 or 4 people, obviously the rest of the team is going to have an easy go of it, and we have a roll. If I fail to line up my shots, get killed without taking anyone else with me, i just denied my team 85 tons, and they are now likely to lose.
Unless of course, those one or two guys who can carry pick up my slack. What it boils down to is attrition, that can be offset with skill. If you suck, you need someone else with better than average skill to offset your weakness. And that's gonna be rare in T1, where everyone skill level is well, tier one.
I spectate every fight I die in, I never leave a match (i'm still trying to learn), and I can tell you. The bad in T5 is everywhere. So many people dont even use a mouse. So me having a good match all by myself can sway it. Or visa versa.
Edited by HauptmanT, 13 September 2016 - 03:28 PM.
#119
Posted 13 September 2016 - 03:19 PM
Your average pug will consist of three shooters, two average players, one totally new guy and six each total gits.
Now don't get me wrong we all go git at times.
You, me everyone of us.
After playing since closed beta I willing to speculate that of 75% of us are total gits.
Maybe better at COD or BF4, but in this game Total Git-Town USA.
So how does PGI balance this out?
Matchmaker unfortunately for PGI the 75% rules the roost and it has to balance that out with average to pro players.
And 25% does not go a long way.
Occasionally MM seem to burp and throws the whole 25% on the opposite team which in turns into a rollstomp.
Both side have the 25% average so you do the math.
Sometimes you can just look at your team and tell if it's gonna be one.
Sigh.
#120
Posted 13 September 2016 - 03:20 PM
3 user(s) are reading this topic
0 members, 3 guests, 0 anonymous users