47 replies to this topic

[Aidan Crenshaw]

Anecdotes against statistics, really. Such games happened before the merge. Will a screenshot of a stomp from before the merge mean I'm right?

[Knight Captain Morgan]

Considering matchmaker puts T1 against T5 anyway, do whatever.

[Nearly Dead]

Statistics don't retain players.

[FRAGTAST1C]

It's not just about the resetting the tiers. There are still the following...

1. The point of assessing the threat value of a group. You're pitting weekend groups against comp groups

2. Fixing spawn locations where Assaults end up in Alpha lance even though they're not in a group

3. fixing the 3-man and 2-man groups ending up in the same team

4. Making sure BOTH teams have a group, let alone equally matched ones. Many times, one team will have a group and a full stack of 80-90%+ players while the other team has a cadet or two

[VonBruinwald]

Aidan Crenshaw, on 27 May 2020 - 03:10 AM, said:

Anecdotes against statistics, really. Such games happened before the merge. Will a screenshot of a stomp from before the merge mean I'm right?

It's a question of those statistics being valid. The source data isn't shared so nobody can confirm Paul's numbers. So instead we have to relay on the accumulation of anecdotes to see if the community can draw a comparable conclusion.

The alternative you're suggesting is to blindly trust Paul, which statistically, is not a sound option.

[Anomalocaris]

VonBruinwald, on 27 May 2020 - 09:26 AM, said:

The alternative you're suggesting is to blindly trust Paul, which statistically, is not a sound option.

You mind if I steal that for my forum signature? Real money quote there.

[Anomalocaris]

FRAGTAST1C, on 27 May 2020 - 06:25 AM, said:

It's not just about the resetting the tiers. There are still the following...

1. The point of assessing the threat value of a group. You're pitting weekend groups against comp groups

2. Fixing spawn locations where Assaults end up in Alpha lance even though they're not in a group

3. fixing the 3-man and 2-man groups ending up in the same team

4. Making sure BOTH teams have a group, let alone equally matched ones. Many times, one team will have a group and a full stack of 80-90%+ players while the other team has a cadet or two

I think you hit this right on the money. A PSR reset/rework has long been needed and I've bitched enough about PGI's intransigence on it that I'm just worn out. But it is still needed so I'm glad they've finally acknowledged the issue.

That said, the issues you highlight are some major pitfalls still in the way of making a PSR rework useful.

If they had done this before the merge queue, PGI could have avoided so many of the issues and really improved the health of the playerbase and match quality. But now, how will they balance groups? How will they separate player results in group drops vs. solo drops? I'm just not sure they can. And that's even assuming they get the PSR rework right in the first place (the issue is not gaining on a well played loss - it's gaining on a poorly played win, and Paul still doesn't seem to understand that).

[FRAGTAST1C]

Anomalocaris, on 27 May 2020 - 11:58 AM, said:

(the issue is not gaining on a well played loss - it's gaining on a poorly played win, and Paul still doesn't seem to understand that).

It's Paul, so, yeah...

Anyway, hope this is the first step on a long path to making MWO's Instant Action mode tolerable but I have my doubts.

[LordNothing]

i dont think it will matter much, any manipulation early on will just cause a bad time later as you have to wait longer for the system to readjust to your actual skill.

[VonBruinwald]

LordNothing, on 27 May 2020 - 07:09 PM, said:

i dont think it will matter much, any manipulation early on will just cause a bad time later as you have to wait longer for the system to readjust to your actual skill.

But if you tank it, you can farm pugs for longer, boosting your Jarl's....

(Of course, people can always look back at pre-reset for comparison)

[Anomalocaris]

VonBruinwald, on 28 May 2020 - 01:21 AM, said:

But if you tank it, you can farm pugs for longer, boosting your Jarl's....

(Of course, people can always look back at pre-reset for comparison)

Yep, if it works like the current Tier system does, tanking the first 20 matches might mean it takes you 100 matches to get out of Tier 5 vs. 10 to reach Tier 3.

While I think it would get rather boring for a high skill player to farm true Tier 5 players for hundreds of matches, every single MMO combat game on the market pretty much proves that there are always farmers out there. It definitely needs to be considered.

[O L L O]

Willard Phule, on 26 May 2020 - 11:14 AM, said:

I don't know about you, but I'm making Tier 5 a goal. 20 games of tongue dragging, LRM spewing complete incompetence.

Then we'll see how long it takes to get back to Tier 1. I'm going to guess less than a week.

I think you may be in for a surprise.

[LordNothing]

VonBruinwald, on 28 May 2020 - 01:21 AM, said:

But if you tank it, you can farm pugs for longer, boosting your Jarl's....

(Of course, people can always look back at pre-reset for comparison)

i have thousands of games on jarls, most of them bad. i dont think a few stomps against lower tiers is going to change that much.

[VonBruinwald]

LordNothing, on 28 May 2020 - 04:15 PM, said:

i have thousands of games on jarls, most of them bad. i dont think a few stomps against lower tiers is going to change that much.

You've got to play the meta on this one:

Currently you're a T3 in T1 getting clubbed. Your current stats are what you managed to chip out between clubbings.

Post-PSR reset you will no longer be over-tiered. You'll be a T3 fighting T3's, putting you are equal footing. As a result your stats will rise because you're no longer the perpetual-underdog.


This is how you play it across... If you happen to tank your first 20 games, well, you can always blame it on your team/matchmaker and claim it wasn't your fault. We've all had winning and losing streaks.

[LordNothing]

VonBruinwald, on 29 May 2020 - 04:43 AM, said:

You've got to play the meta on this one:

Currently you're a T3 in T1 getting clubbed. Your current stats are what you managed to chip out between clubbings.

Post-PSR reset you will no longer be over-tiered. You'll be a T3 fighting T3's, putting you are equal footing. As a result your stats will rise because you're no longer the perpetual-underdog.


This is how you play it across... If you happen to tank your first 20 games, well, you can always blame it on your team/matchmaker and claim it wasn't your fault. We've all had winning and losing streaks.

id like to think im at least t2 material. but t3 wouldn't surprise me.

Edited by LordNothing, 29 May 2020 - 03:20 PM.

[Nesutizale]

I wonder where I will end up when the change goes life because I took a break in Nov'19. When I started again in Mar'20 my stats where worse then when I stopped but that is to be aspected but then I jump from beeing around 50% to 80% ?

I doubt that I got that much better within the few games I played but I also doubt that everyone else got so much worse. Also, except for a handfull of games I played mostly solo so haveing a group that helped me can't be the reason.

Did the playerbase change that drastical in one month ? Something seams off here or is it so strange because of the few numbers of games I played so the statistic dosn't realy say something.
How many games are neccessary to have a relevant basis for placeing a player in a certain Tier?

Here are the last 3 Jarls stats. Personaly I think T3 should be good for me.
Season - Percentile - Games Played - W/L Ratio - K/D Ratio - Average Score
45 (Apr '20) - 80% - 15 - 0.88 - 1.33 - 262
44 (Mar '20) - 46% - 25 - 0.56 - 0.61 - 222
40 (Nov '19) - 52% - 119 - 0.95 - 0.66 - 222

[Kubernetes]

Wait, I've been away too long. They're resetting PSR? When? The body's been dead and decomposing for years and PGI just now is getting around to clearing the airway.

[FRAGTAST1C]

Kubernetes, on 30 May 2020 - 08:38 PM, said:

Wait, I've been away too long. They're resetting PSR? When? The body's been dead and decomposing for years and PGI just now is getting around to clearing the airway.

Don't worry. It's not the PSR reset that but other things like lopsided tonnage balance, group threat value assessment, spawn fix, etc., aren't being addressed. But... baby steps, right? I mean, the guy is basically old enough to drive but is only now getting potty trained

[BARRY SHlTPEAS]

Aidan Crenshaw, on 27 May 2020 - 01:57 AM, said:

I don't think so. Following Paul's recent posts, PSR deviation has been quite tight the past days, without a significant rise of wait-time.

Yo dude, do you understand PSR deviation? Could you explain it to me please?

[VonBruinwald]

-Verti-, on 31 May 2020 - 05:11 AM, said:

Yo dude, do you understand PSR deviation? Could you explain it to me please?

Nobody understands PSR deviation. Paul goes silent when we quiz the underlying mechanics. What we have learned is PSR is more granulation than 1-5 as the tiers present them.

Now supposedly (from Paul's discoveries):

Based on 8 matches:
When the matchmaker isn't trying to balance anything we get a Average PSR difference of 1077 with an SD of 406.
When the matchmaker is tuned to balance PSR we get a Average PSR difference of 149 with an SD of 45.
When the matchmaker is tuned to tonnage we get a Average PSR difference of 1076 with an SD of 407.

Seems consistent, you get the same PSR deviations when you throw PSR balance out the window. What we currently have is a combination of PSR and tonnage balance with an Average PSR difference of 111 with an SD of 107. This result was based on 14 matches.

The problems with this are the abysmal sample sizes, I would like to assume Paul used more and these were just illustrative; the complete lack of preceding date for what we had prior to the merge; and no sense of scale for what the true PSR system is based on (what are the PSR thresholds for tiers and maximum PSR value?).

It's also worth noting that the MM doesn't appear to balance the tonnage of mechs but rather the classes as gleemed from this tidbit.

Average wait time in QP Queue: ~1 min 30 seconds (up from ~48 seconds)
Average Tier separation between teams: ~300 (tightened from ~1800)
Average Weight Class difference: +/- 1 or 2 in a given weight class (5 heavies, 3 assaults for example) (tightened from 12)

Edited by VonBruinwald, 31 May 2020 - 09:55 AM.