75 replies to this topic

[Davegt27]

I was looking at Jarl's list

https://leaderboard.isengrim.org/

and my overall rank is 26259
edit: 27212 on 15 Jan

26259! I don't think you can get much lower then that
it also means if you see me on your team your chances of winning just
went down I mean come on why am I even in tier 1 it makes no sense

its obvious the current PSR system is not working
also PGI will never listen to me no matter how much money I spend
since my low scores mean in there eyes it qualifies me as low like scum

can any of you try hards talk some sense into PGI
the PSR is not working

GLHF

Edited by Davegt27, 14 January 2018 - 11:46 PM.

[Helsbane]

Yeah, the wonderful XP bar we got for tier level sorting is pretty much a joke. I'm low tier two material on a good day, but I've played enough to fill up my XP-o-meter, so I must be great, right? But, coding is hard, and sorting even harder, so here we all stay with pleas falling on the deaf ears of the devs for pretty much anything really...

'VISION'

[Kamikaze Viking]

PSR has needed a reset and an adjustment for a long time. The XP bar has failed to do what it was intended to do as people of all skill levels such as the OP eventually reach Tier 1. Hence the skill variance at tier 1 is the greatest, and the match quality is the worst. Tier 5 games on my alt at least most of the team is on the same page even if they haven't the skills, it seems they all get along better down there.

I cant tell you the number of clearly brand new players I've been up against that shouldn't have to deal with me and other comp players while they are still learning.

And remember folks the Match Maker Works (long live Karl Berg) but the PSR system is feeding it junk data. Fixing PSR will also fix the match maker.

[Nema Nabojiv]

As I agree the xp bar is not indicative of anything, with the population that low, it doesn't really affect our MM quality either.

We basically have two buckets now - tier5+tier4s, and tier2+tier1 with tier3 being in limbo and having a chance to be mathced with each group depending on current number of players.

[Kamikaze Viking]

Nema Nabojiv, on 10 January 2018 - 04:23 PM, said:

with the population that low, it doesn't really affect our MM quality either.

I think this is a false assumption. I trying to explain the fix to a friend last night and he assumed the same.

It is possible to fix it, yes the population is 'lower' than it was, but its not really all that much different to when the MM + PSR was new and worked (before PSR went bad as predicted).

The other factor is 'some' players will cry if they get reset to lower tiers. Ignore them for 2 weeks and ask again and those same players will thank you for the improved match quality.

[Xiphias]

You typify the type of player that shouldn't be in T1 and show what is wrong with the system. I think most of the "true" T1 pilots agree that they system should be fixed instead of having an upward bias towards T1, that unfortunately pushes players like you who don't want to be in and really shouldn't be in T1.

[Nema Nabojiv]

Kamikaze Viking, on 10 January 2018 - 04:31 PM, said:

I think this is a false assumption. I trying to explain the fix to a friend last night and he assumed the same.

It is possible to fix it, yes the population is 'lower' than it was, but its not really all that much different to when the MM + PSR was new and worked (before PSR went bad as predicted).

The other factor is 'some' players will cry if they get reset to lower tiers. Ignore them for 2 weeks and ask again and those same players will thank you for the improved match quality.

I have no other explanation why it takes so long to find a match then. Like three to five minutes long, outside of combined EU+US primetime,. It may be my subjective impression though as I didnt sat with timer, but it feels uncomfortably long.

[FireStoat]

I'd be completely happy with all tiers being stripped from players with a hard reset, and a new system in place to toss out the experience bar mechanic. I'm a good tier 3 player / mediocre tier 2 player. I'm not a one and there are tons like me. Tier 1 players in my mind are the tryhard guys that play a rotation of 4 total different mechs and like, that's it. They focus on their specialized gameplay for countless hours on top of having superior map awareness and situational judgement.

Maybe PGI could give it a whirl when Solaris hits as a draw to bring older players back as well as divide attention from traditional quick drops / group drops wedged against Solaris as the new shiny toy? Because both rankings would be different and separate?

[fat4eyes]

Don't get too hung up on leaderboard stats, your numbers will go down as you get to fight players closer to your skill level. As an example, here's me

https://leaderboard....arch?u=fat4eyes

While here is my tier 2 alt:

https://leaderboard....ch?u=TattleTale

My alt is ranked higher than me even though I use less meta mechs on that one (I use it mainly to experiment on mechs that I don't have on my main account, and for faster matchmaking during low pop hours). I'm guessing it's because I'm more often matched with Tier3/4 on my alt than on my main. As long as you're around 1.0 WL you're not dragging your team down.

[justcallme A S H]

Davegt27, on 10 January 2018 - 03:37 PM, said:

can any of you try hards talk some sense into PGI
the PSR is not working

A lot of us have been trying

Long as your games above 200m/s are better than the games below 200m/s, you'll make T1. Without knowing the exact current values (Be nice if PGI could chime in and provide). I would say it needs to work in the ballpark of the below:

Below 120 m/s - large drop
Between 121-200 m/s - small drop
Between 201-250 m/s - even - no change.
Between 251-300 m/s - small increase
Between 301-380 m/s - moderate increase
Above 381 m/s - large increase

There are far too many people with ~250 m/s or below averages in T1. That should be a T3 level, maybe T2.

I regularly see players with ~170 m/s score in my games. Consistently doing no more than 150dmg as they "will play the game their way"... Which I guess would be fine if it wasn't a 12 v 12 team oriented FPS.

The big issue overall is though... Population & match maker.

Over the last 12 months the population has dropped by a whopping 25%... That is a big drop. This means less players, more wait time and thus the MM is constantly expanding to basically include anyone in any QP drop so it will never really work even if there is a PSR reset or calculation changes simply because there isn't the players to make it work effectively. So putting a bunch of people back to T3 or even T4/T5 won't stop them from ending up in matches regularly with a bunch of T1s. All it'll mean is the true T1s - long wait times for the same end result.

Edited by justcallme A S H, 10 January 2018 - 06:53 PM.

[justcallme A S H]

fat4eyes, on 10 January 2018 - 05:41 PM, said:

Don't get too hung up on leaderboard stats, your numbers will go down as you get to fight players closer to your skill level. As an example, here's me

https://leaderboard....arch?u=fat4eyes

While here is my tier 2 alt:

https://leaderboard....ch?u=TattleTale

My alt is ranked higher than me even though I use less meta mechs on that one (I use it mainly to experiment on mechs that I don't have on my main account, and for faster matchmaking during low pop hours). I'm guessing it's because I'm more often matched with Tier3/4 on my alt than on my main. As long as you're around 1.0 WL you're not dragging your team down.

You're missing a key point though. You didnt start the Alt till S7. By this stage your skill floor is much higher from the outset. This is why it is ranked higher simply because of historical numbers more than anything else, even if it's weighted more for newer seasons.

[Revis Volek]

Davegt27, on 10 January 2018 - 03:37 PM, said:

I was looking at Jarl's list

https://leaderboard.isengrim.org/

and my overall rank is 26259

26259! I don't think you can get much lower then that
it also means if you see me on your team your chances of winning just
went down I mean come on why am I even in tier 1 it makes no sense

its obvious the current PSR system is not working
also PGI will never listen to me no matter how much money I spend
since my low scores mean in there eyes it qualifies me as low like scum

can any of you try hards talk some sense into PGI
the PSR is not working

GLHF

If you had better MM i bet you would not be stuck so low in the sea of puglandia. Based on your numbers id say you currently belong in low t2 to mid t3. Your numbers are not abysmal and correct MM with like skilled players would do wonders for those players who are not in the top % IMO.


IF you are gonna reset tiers it should be from when this system started IMO. Lets get rid of the old data from ELO or whatever Frankenstein we had before and just use the data from this system to reset everyone. Then in another PSR SEASON (6 months or whatever) reset everyone and base it off the past data and keep it current and fresh based on how you have played RECENTLY not HISTORICALLY from day one.

Edited by Revis Volek, 10 January 2018 - 07:04 PM.

[slide]

@ OP,

I am not even sure you can compare Jarls list to PSR in any realistic way. Take my stats.

Jarl's list puts me at 8212 yet on this account I am not even at T1 (about 90% T2)

My alt account is a at 3311 (only one season old though) which is only about 50-60% through T3.

Both sets of stats suggest that I am better than the OP and yet PSR would say otherwise. MM can only work with the data it has, however that data is patently flawed.

I pay no attention to a persons tier on these forums as it is completely meaningless. For the same reason I don't display mine.

(note: if you are wondering why this account is only T2 it is because I played nothing but CW for 2 years, PSR can't move if you don't play)

[justcallme A S H]

Revis Volek, on 10 January 2018 - 06:57 PM, said:

Your numbers are not abysmal and correct MM with like skilled players would do wonders for those players who are not in the top % IMO.

The other thing to think about is, what is the top %? It is about 2% of the population, literally. You look at the avg match scores of players in the 99.XX percentile vs the 95.XX and it's a big difference. Huge in-fact.

I'm - 99.86% / rank 63 - 379 avg m/s on the Jarls list. Lets go to rank 1200?

At rank 1200 you have players at... 97% / 280 avg m/s... Just making T1 over a long period (if using my goalposts as above).

Now there are some 28,000 active across a month at the moment. So just think about that, only 1200 odd players around the T1 mark. The majority of the MWO population is at a T3 level and should be ranked as such. That is where the "meat" in the bellcurve is and where the majority should be. If Dave is in T1 then PSR is absolutely broken as a system (no offence Dave, i know you ain't taking offence even so.).

The other issue is that the PSR bar counts GroupQ and SoloQ. GroupQ is generally MUCH harder to get higher match scores in as if you are playing even in a 4man with competent pilots. You can shred the entire 12 on the other side pretty much on your own where SoloQ you are doing the bulk of the heavy lifting as a true T1 pilot, on your own, most of the time.

I know my rank and in-game scores are far higher than many pilots that are much better than I am but they GroupQ more and I SoloQ more.

To truly fix it GroupQ and SoloQ need to be split out and GroupQ avg's actually lowered compared to SoloQ.

So I agree it would improve the matches for some players but at the same time it would also increase the wait times for others.

A delicate situation it is all in right now simply because of how badly the system has been for so long. Hard to fix the situation after so long.

Edited by justcallme A S H, 10 January 2018 - 07:18 PM.

[Revis Volek]

justcallme A S H, on 10 January 2018 - 07:15 PM, said:

The other thing to think about is, what is the top %? It is about 2% of the population, literally. You look at the avg match scores of players in the 99.XX percentile vs the 95.XX and it's a big difference. Huge in-fact.

I'm - 99.86% / rank 63 - 379 avg m/s on the Jarls list. Lets go to rank 1200?

At rank 1200 you have players at... 97% / 280 avg m/s... Just making T1 over a long period (if using my goalposts as above).

Now there are some 28,000 active across a month at the moment. So just think about that, only 1200 odd players around the T1 mark. The majority of the MWO population is at a T3 level and should be ranked as such. That is where the "meat" in the bellcurve is and where the majority should be. If Dave is in T1 then PSR is absolutely broken as a system (no offence Dave, i know you ain't taking offence even so.).

The other issue is that the PSR bar counts GroupQ and SoloQ. GroupQ is generally MUCH harder to get higher match scores in as if you are playing even in a 4man with competent pilots. You can shred the entire 12 on the other side pretty much on your own where SoloQ you are doing the bulk of the heavy lifting as a true T1 pilot, on your own, most of the time.

I know my rank and in-game scores are far higher than many pilots that are much better than I am but they GroupQ more and I SoloQ more.

To truly fix it GroupQ and SoloQ need to be split out and GroupQ avg's actually lowered compared to SoloQ.

So I agree it would improve the matches for some players but at the same time it would also increase the wait times for others.

A delicate situation it is all in right now simply because of how badly the system has been for so long. Hard to fix the situation after so long.

Well there in lie the big questions, where do we actually break the tiers?


What is T1? T2? etc.. You hit the nail on the head because a 550 ms vs a 250 is a HUGE gap that not even groups can cover often. But i also agree that group q should be weighted more then solo because of exactly what you said. IT much easier to get a high MS by yourself carrying every match then with 5 or 6 guys foaming at the mouth to get theirs before all 12 are dead.

But where yo break it up is IMO what they are missing, we have a system that needs fixing so it works better and then we go from there. Even if it means shrinking some Tier over time by raising the MS requirement to get into it as players get better sorted.

[justcallme A S H]

Indeed the big question.

PGI have the data. They just need to interpret/use it properly and ensure the 'bulk' are in T3, just need to work out what that bulk/median average point actually is. Right now it's too low, way too low.

Edited by justcallme A S H, 10 January 2018 - 09:05 PM.

[TheArisen]

I honestly don't experience these problems that people talk about but I do agree the system has a few flaws.

[PocketYoda]

Weird i'm tier 4 and it says my rank is 24941 wtf am i higher than the op?

Edited by Samial, 10 January 2018 - 09:52 PM.

[UnofficialOperator]

Samial, on 10 January 2018 - 09:52 PM, said:

Weird i'm tier 4 and it says my rank is 24941 wtf am i higher than the op?

Sort of yes. But not really. Its meaningless or meaningful depending on who you ask. His season 18 matchscore at 228 places him at 19323. But then supposedly in the optimal scenario that MM is working, you guys are playing different tiers of opponents anyway...

[MischiefSC]

Davegt27, on 10 January 2018 - 03:37 PM, said:

I was looking at Jarl's list

https://leaderboard.isengrim.org/

and my overall rank is 26259

26259! I don't think you can get much lower then that
it also means if you see me on your team your chances of winning just
went down I mean come on why am I even in tier 1 it makes no sense

its obvious the current PSR system is not working
also PGI will never listen to me no matter how much money I spend
since my low scores mean in there eyes it qualifies me as low like scum

can any of you try hards talk some sense into PGI
the PSR is not working

GLHF

Sorta. The Jarls List is a great resource for stats -

However all it cares about is damage farming and playing smaller mechs. It even adjusts your stats to try and resolve for your weight class and the relative imbalance between weight classes.

In reality for PSR what matters is your win/loss. Most people are closer to 0.80 w/l. So because you win 0.94 you're winning more than the average player... which means slowly, inevitably, you'll climb the PSR tier.

The #2 ranked players on the Jarls List right behind kaffeangst and 12 ranks better than Proton has a 1.52 w/l. However he got great scores in lights for a few hundred matches over 18 months which makes him the 2nd best player in the game.

Again, the Jarls List gets a lot of great data together and in a very useful format. However the rankings are about as arbitrary as PSR is - a few pages into either shows the issues inherent in the ranking system.

However PSR is still a bad system and no, you shouldn't be in T1 with a 0.97 w/l. However from a matchmaking perspective and the small population we have, does it make things better overall for you to be filler for matches with high skilled players (given your experience with the game) or you romping through the matches with nubs.

No real good answer to that.