51 replies to this topic

[SeventhSL]

I've always thought that the easiest way to solve PSR was to rank all players by match score, at the end of the match. The top 6 go up, the bottom 6 go down and the rest stay the same. Every game has exactly the same amount of people go up as down.

[Lily from animove]

SeventhSL, on 15 January 2018 - 05:45 AM, said:

I've always thought that the easiest way to solve PSR was to rank all players by match score, at the end of the match. The top 6 go up, the bottom 6 go down and the rest stay the same. Every game has exactly the same amount of people go up as down.

would work if we had a better matchscore calculation, but as long as damage si the major factor in it and those lrm kiddos do 400damage per mech they meet not killign them, we would get a lot of inefficient people with high matchscores in the wrong ranks.

[Jonathan8883]

I'm just glad we don't have "promotion rounds" like ranked play in League of Legends. Talk about toxic attitudes!

[Fleeb the Mad]

I think a much simpler solution would simply be if the no change score for PSR had a bigger range, so that mediocre scores on wins didn't promote upwards. If you're scoring average in your games it tends to imply you're more or less in the right place.

It's honestly pretty difficult to use a matchmaker based on one statistic because they can all be manipulated and sometimes cause a feedback loop. For example, win/loss rate would initially be a decent metric for gauging players if you could filter out group queue stats. Initially. A matchmaking system by definition must attempt to balance teams, so one that used win/loss exclusively would force rates for players of all skill ranges towards 1.0 over time regardless of their individual performance. That's just because high W/L outliers would have to be balanced by either the same on the enemy team, or lower ones on their team. After that point you need a new metric.

[Brain Cancer]

MrMadguy, on 15 January 2018 - 04:02 AM, said:

That happens only when Tier 1 guy plays on Tier 5 alt account.

You'd think so, but no.




I end up killing a quarter of the enemy team singlehandedly, someone secures #5 on my fifth KMDD, and I do the same for kill #6 (the last enemy standing). If I'd performed like the rest of my team, we'd have lost- not horribly, but we'd have lost. Carrying is very, very real and it happens at T1 like it happens everywhere else. And I'm NOT a super-killer legendary T1 player.

Edited by Brain Cancer, 15 January 2018 - 01:38 PM.

[Stinger554]

Brain Cancer, on 15 January 2018 - 01:37 PM, said:

You'd think so, but no.




I end up killing a quarter of the enemy team singlehandedly, someone secures #5 on my fifth KMDD, and I do the same for kill #6 (the last enemy standing). If I'd performed like the rest of my team, we'd have lost- not horribly, but we'd have lost. Carrying is very, very real and it happens at T1 like it happens everywhere else. And I'm NOT a super-killer legendary T1 player.

Not trying to say you didn't make a difference in that match but the enemy team only had two players with 300+ damage whereas your's had 8 not including yourself(9 if you do include yourself) I'm pretty sure this is just a mismatch of skill.

Which PSR doesn't accurately represent.

[Brain Cancer]

Stinger554, on 15 January 2018 - 01:42 PM, said:

Not trying to say you didn't make a difference in that match but the enemy team only had two players with 300+ damage whereas your's had 8 not including yourself(9 if you do include yourself) I'm pretty sure this is just a mismatch of skill.

Which PSR doesn't accurately represent.

Not really. Checking the S18 stats, they're actually all around my level of stats heck, at least one of them is notably better than me. What happened is I pretty much cleansed the entire alpha lance off the field by myself plus the Blackjack in very short order (plus the Jagermech killshot at the end, but that was a total gang-beating), thus the low damage scores. At that point, it was essentially 8v12 two minutes in, and we STILL ended up with 8 guys down because those big damage numbers came from a bunch of team-mates that basically couldn't hit the same spot twice.

Again, I'm not a great player. But it's depressing to see that many games, you're solo-killing 3-4+ targets when there's a win to get that win, or someone else is because compensating for the lack of others is generally a fact of life in MWO.

Matches where nobody ends up with more than 2 kills and damage tends to be pretty much evenly spread out across the team in terms of who dealt what are rare birds indeed. But they tend to be the stompiest of stomps to me, as that's where a clear mismatch of skill has happened.

[justcallme A S H]

MrMadguy, on 15 January 2018 - 04:16 AM, said:

Again, this happens just because you're being matched with/against lower Tier players, so you really can win matches solely it this case. But you should know, that it happens just because current MM is broken and you should also know, that for lower Tier players situation is completely opposite. For example I can go AFK and it won't change result of match, cuz my <100 dmg and 0 kills aren't such big deal.

There isn't the population for that to work though.

Only having the carry players in T1, they'd never find a match. You inevitably will end up with T2-T3 in your matches.

Edited by justcallme A S H, 15 January 2018 - 05:28 PM.

[arcana75]

Brain Cancer, on 15 January 2018 - 01:37 PM, said:

You'd think so, but no.




I end up killing a quarter of the enemy team singlehandedly, someone secures #5 on my fifth KMDD, and I do the same for kill #6 (the last enemy standing). If I'd performed like the rest of my team, we'd have lost- not horribly, but we'd have lost. Carrying is very, very real and it happens at T1 like it happens everywhere else. And I'm NOT a super-killer legendary T1 player.

It's not very useful to use an anecdotal screenshot to prove a general point. An end game summary is a snapshot of a match, and fails to account for the entire match's dynamics. Imagine trying to get your kid to take a profile photo. The photo might eventually come out great, but it didn't account for the earlier 10 minutes to get him to sit still and smile properly at the camera.

Same with end-game screenshots. There could be any number of reasons to explain why a pilot's damage/score is low, and none of them would indicate a player's poor overall skill.

Cases in point, see these 2 screenshots.




In either images, one could draw any number of conclusions to support whatever BS you wanted to put forth, for example, that TTB was a terrible player, that there were many bad players in either match, etc etc. But either failed to capture the battle's flowing dynamics, and none of it is related to the matchmaker or a player's lack of skill.

Note tho, they do demonstrate individual players' skills over other players, but showing that A has more skill than B does not also mean that B does not have skill. In that match with me and the stormcrow dominating the game, it wasn't so much skill than having the courage to push, and even then how would you know it to be the case without a video to watch?

[Xavori]

Just some tips if you want to lower your PSR without being a total loss to your team:

- Use TAG and Narcs to mark targets for your teammates. You'll get some match score out of this, but nowhere near as much if you were doing damage.

- Share all your armor all the time. If you stand next to a teammate and are focused on torso twisting, you can go a long ways towards saving them and letting them get more damage/kills.

- AMS. This one also works pretty well for c-bill farming, but you need to be careful as it also generates decent match score if you do any fighting along with it.

- Just using the R key to lock targets while sitting back from a fight. Your teammates can use your lock to maneuver towards or missile fire at a target. And since you're not actively shooting, you're a lot less likely to be noticed and can therefore focus more effort towards keeping your target locked.

- Be the squirrel. Firing a laser near an enemy while running a fast mech will usually get them to chase you. If you can tie up 2 or 3 mechs chasing you that means your team is only down 1 mech while the enemy is down several. That's quite a bit of tonnage you've taken out of the main fight giving your team a pretty significant advantage.

[C E Dwyer]

adamts01, on 15 January 2018 - 01:40 AM, said:

Yes, PSR and MM are broken. I also think it's important to make clear that PSR and MM aren't at all related. How players are ranked and how players are grouped are two separate issues, both in a bad situation. "How PSR should work" is broken as well. At a very quick glance you can see that PGI's Moderate Drop combined with a Small Drop averages a more sever small drop. But on your chart, a small drop and a small drop average a moderate drop?.... That's not how averages work. But let's forget all that.

Any system heavily based on Match Score is doomed to fail, as two teams each putting up a good fight, which should be the end goal, will see every player progress. We could still see a situation where everyone in the match gets a PSR increase, which is part of our current problem.

So here's the simple solution, which accounts for support builds doing a good job, team leading, awareness, calling good targets...

• PSR is purely based on W/L alone
  
• We have 3 tiers, and someone has to bump someone down from tier 1 to move up from tier 2
  
• You only fight your own tier
  
• Each tier composes 1/3 of the active playerbase
  
• You lose 1 tier per month of inactivity
  
• Your first 50 matches back after a month gone yield double points

this system also has several problems attached to it.

The system won't protect complete newbies from being farmed.

A fair amount of the population doesn't play more than 50 games a month.

Whether P.G.I do a complete overhaul and produce the best MM system, it's still going to be crippled by low population .

I suspect this last one is why P.G.I haven't bothered, and are very resistant to even minor changes.

I also and I could be very wrong on this, but I have a slight recollection of P.G.I wanting to make it hard to drop, so that long standing people couldn't game their rating to farm newbies.

This doesn't mean it shouldn't be harder to go up, but if I am correct, it makes sense to try and improve the NPE

Edited by Cathy, 15 January 2018 - 03:08 PM.

[C E Dwyer]

Here's something else to throw at you.

Let's suppose, P.G.I find a way of realistically getting all the really skilled players in a place away from the potatoes.

what will it change..?

For a start there will be no 1k damage carries, because everyone is skilled right ?

It is likely to stop such excessive nascar, or is it ? Will someone be more likely to listen to another's calls ? Or will there be many cases of I'm not listening to that A-hole from unit X ?

Will it stop stomps ?

I have my doubts, and I think it's likely that stomps will become more common, because losing one good skilled player first is far more likely to cause a domino effect, than losing a potato or two.

I'm just throwing this out there for people to talk around, I don't really have an opinion on this, but I don't think it's all so cut and dried that removing the potatoes will stop stomps

[arcana75]

Cathy, on 15 January 2018 - 03:22 PM, said:

Here's something else to throw at you.

Let's suppose, P.G.I find a way of realistically getting all the really skilled players in a place away from the potatoes.

what will it change..?

For a start there will be no 1k damage carries, because everyone is skilled right ?

It is likely to stop such excessive nascar, or is it ? Will someone be more likely to listen to another's calls ? Or will there be many cases of I'm not listening to that A-hole from unit X ?

Will it stop stomps ?

I have my doubts, and I think it's likely that stomps will become more common, because losing one good skilled player first is far more likely to cause a domino effect, than losing a potato or two.

I'm just throwing this out there for people to talk around, I don't really have an opinion on this, but I don't think it's all so cut and dried that removing the potatoes will stop stomps

We only need to look at the MWO championship match summary scores to find out.

[MrMadguy]

Brain Cancer, on 15 January 2018 - 01:37 PM, said:

You'd think so, but no.




I end up killing a quarter of the enemy team singlehandedly, someone secures #5 on my fifth KMDD, and I do the same for kill #6 (the last enemy standing). If I'd performed like the rest of my team, we'd have lost- not horribly, but we'd have lost. Carrying is very, very real and it happens at T1 like it happens everywhere else. And I'm NOT a super-killer legendary T1 player.

First thing you need to do - to prove, that all this players are "real" Tier 1 players. As I remember, I see you in my matches quite often. And I recently was in Tier 3, while my real skill - is Tier 4 level.

Second thing - I also have bunch of cherry picked screenshots in my MWO folder back from that days, when I was playing at potato ratings.

Again, great matches happen, when both teams are below your skill level. When enemy team is above your team - you get stomped. If your team is above enemy team - they stomp it too fast, so you can't show great performance. Great matches happen, when enemies are potatoes, but your team is full of potatoes too, so all glory goes to you.

Something like this:

[Khobai]

Even if PSR worked games still wouldnt be balanced

because the game still wouldnt balance weight classes or quality of mechs.


matchmaker needs to balance three things not just one thing.

player skill, tonnage/weight class distribution, and mech quality

Edited by Khobai, 15 January 2018 - 08:39 PM.

[Brain Cancer]

MrMadguy, on 15 January 2018 - 08:33 PM, said:

First thing you need to do - to prove, that all this players are "real" Tier 1 players. As I remember, I see you in my matches quite often. And I recently was in Tier 3, while my real skill - is Tier 4 level.

There are no such things as all-tier 1 games. There aren't enough of them at any given moment to get 24, meaning all games are T1-T3. The potatoes are the same, the only difference is the T1 ones tend to potato longer. Once you hit T3, welcome to the Old Spuds Retirement Home.

Quote

Again, great matches happen, when both teams are below your skill level. When enemy team is above your team - you get stomped. If your team is above enemy team - they stomp it too fast, so you can't show great performance. Great matches happen, when enemies are potatoes, but your team is full of potatoes too, so all glory goes to you.

I just looked at matchscore, W/L, and K/D. There is no way to confirm tier for an entire match's worth of players, as there is no record of actual Tier rank outside the forum posters. But my W/L and K/D tend to be pretty darn stable on the march up the ranks. Season 1 when I was T3? Just above 1.0 W/L and K/D. It's remained about the same the whole way up through to T1.

Mixed T1-T3 matches are the "standard" once you get up there- they almost are one player pool at this point due to the contortions the MM makes.

Khobai, on 15 January 2018 - 08:37 PM, said:

matchmaker needs to balance three things not just one thing.

player skill, tonnage/weight class distribution, and mech quality

I fully expect they will do something like take Solaris tiering and tie it into the matchmaker for chassis ratings (not that I agree, but....). As for tonnage? It'd probably work better split five ways rather than 4: 20-30,35-50,55-65,70-85, 90-100. But that's just me.

[MrMadguy]

Brain Cancer, on 15 January 2018 - 10:03 PM, said:

There are no such things as all-tier 1 games. There aren't enough of them at any given moment to get 24, meaning all games are T1-T3. The potatoes are the same, the only difference is the T1 ones tend to potato longer. Once you hit T3, welcome to the Old Spuds Retirement Home.

You've just confirmed, what I said in my initial post. You can affect result of match, only when you represent 30-50% of skill value of your team. When everybody is "fair" 1/12 of team and 1/24 of match - it's not possible to distinguish one's effort towards victory.

And you can't use "We have too few players" as argument, cuz it's not reason to have bad MM - it's actually consequence of having it. For example I play much less this days exactly cuz MM started being terrible around implementation of 3/3/3/3 and release of Clans. And I had been warning PGI about it and possible loss of playerbase due to it for around 2 years before I stopped caring and almost quit this game. They've ignored this warnings and now we have, what we have.

And if PGI don't want to ditch this game and want it to start recovering - they should fix problems, that have caused decline of this game in a first place. You can't blame players, while game itself is still broken.

Edited by MrMadguy, 15 January 2018 - 11:16 PM.

[sceii]

PGI said they do PSR thing to separate novice players from players who understand how to control mech.
If you're just bad but you know a bit about this game you go up.
AND PSR DOES IT.
Meanwhile we have threads every day about psr not working, when it completely does thing it was made for.

[adamts01]

MrMadguy, on 15 January 2018 - 11:12 PM, said:

You've just confirmed, what I said in my initial post. You can affect result of match, only when you represent 30-50% of skill value of your team. When everybody is "fair" 1/12 of team and 1/24 of match - it's not possible to distinguish one's effort towards victory.

W/L is the best way to do things. Look at it this way, if you're such a potato that you literally have no impact on the game, then your W/L will represent that and you'll move down until you start having an impact, then you'll reach a balancing point where your W/L is closer to 1:1. The worst player absolutely has an affect on the outcome, he helped his team lose.

sceii, on 15 January 2018 - 11:15 PM, said:

PGI said they do PSR thing to separate novice players from players who understand how to control mech.
If you're just bad but you know a bit about this game you go up.
AND PSR DOES IT.
Meanwhile we have threads every day about psr not working, when it completely does thing it was made for.

PSR exists for MM, and yes, that apparently lives up to PGI standards, but it's still broken because it doesn't live up to industry or player standards, it doesn't even come close to being satisfactory. I hate that MWO has failed to become anything more than 12v12 death match, but if that's all it's going to be then a decent MM really is critical, otherwise you'll get the pissed off and dwindling community you see today.

[MrMadguy]

adamts01, on 15 January 2018 - 11:41 PM, said:

W/L is the best way to do things. Look at it this way, if you're such a potato that you literally have no impact on the game, then your W/L will represent that and you'll move down until you start having an impact, then you'll reach a balancing point where your W/L is closer to 1:1. The worst player absolutely has an affect on the outcome, he helped his team lose.

You know, best proof, that you're wrong - is leaderboard itself. All, W/L based MM cares about - to make W/L as close to 1, as possible, cuz W/L - is only variable in equation. And the best player with W/L = 1 in our leaderboard has AvgMS = 451, while the worst one has 37... Yeah. And according to W/L based MM they all would have equal skills, lol.

Edited by MrMadguy, 16 January 2018 - 12:56 AM.