11 replies to this topic

[Remains Intact]

Thread topic: Improvement for current PSR system.
Not thread topic: Whining about PSR system, or saying it "needs to be removed QQQQ".

All statistics are as a solo pug player, I haven't grouped for more than 20 matches before the start of this december.

So as a tier 2 player slowly creeping towards tier 1 the PSR system seems to be doing it's job under one specific circumstance:

It matches me as a "carry the team, this person is strong" player in most of the matches. And this is the role I tend to play in the matches in which I play very specific mechs that I am very experienced with. In these chassis I pull a 60+% win rate. Which translates as an individual player adding a 10%+ chance for my team to win (which is pretty suprising).

This system doesn't compensate for "the fun factor" of piloting chassis that I am absolutely terrible at, or that just don't perform in general due to build limitations. This causes the matchmaker to think i'm a "solid player, that is trying very hard", and results in a significant detriment to my team. The extreme example of this is a Chassis that I contribute a detriment of -10% chance to win to my team. Thats a swing of expected value to the team of around 22% win chance. And even on one anomalous chassis I seem to contribute a 30% increase in loss rate.

So, long story short: If PSR system is sticking around, please refine it to consider per mech skill modifiers when placing people on teams. I and many players I know would love to run goofy fun builds without being expected to carry hard.

Edited by Remains Intact, 30 December 2015 - 01:25 PM.

[Deathlike]

I simply don't see it happening.

When we had Elo, it was massively requested that it would be per chassis... ideally per variant. For Elo's existence, it was always per WEIGHT CLASS... even though a Dire Wolf is not on the same category as an Awesome.

Considering that PGI dropped that altogether, they simply don't plan on doing that for whatever inane, illogical reason (outside of "not enough samples"... in other words, if you don't drop in the mech enough, the values don't mean anything).

[Dread Render]

If they want us to keep purchasing new Mechs they need to do this.
Otherwise, if you hate the Meta builds... soon that is all there will be.
You will only see two or three Mechs per weight class.
.... no, the OP is correct.
Also these calculations can be done faster than you can blink.
this is totally doable.

[C E Dwyer]

weight class, like it used to be, would make more sense than the nonsense we have now

[EgoSlayer]

Remains Intact, on 30 December 2015 - 12:57 PM, said:

Thread topic: Improvement for current PSR system.
Not thread topic: Whining about PSR system, or saying it "needs to be removed QQQQ".

All statistics are as a solo pug player, I haven't grouped for more than 20 matches before the start of this december.

So as a tier 2 player slowly creeping towards tier 1 the PSR system seems to be doing it's job under one specific circumstance:

It matches me as a "carry the team, this person is strong" player in most of the matches. And this is the role I tend to play in the matches in which I play very specific mechs that I am very experienced with. In these chassis I pull a 60+% win rate. Which translates as an individual player adding a 10%+ chance for my team to win (which is pretty suprising).

This system doesn't compensate for "the fun factor" of piloting chassis that I am absolutely terrible at, or that just don't perform in general due to build limitations. This causes the matchmaker to think i'm a "solid player, that is trying very hard", and results in a significant detriment to my team. The extreme example of this is a Chassis that I contribute a detriment of -10% chance to win to my team. Thats a swing of expected value to the team of around 22% win chance. And even on one anomalous chassis I seem to contribute a 30% increase in loss rate.

So, long story short: If PSR system is sticking around, please refine it to consider per mech skill modifiers when placing people on teams. I and many players I know would love to run goofy fun builds without being expected to carry hard.

PSR takes hundreds or thousands of matches to get to your "true" PSR rating. There is no way it would work at a chassis level in its current incarnation. Elo by comparison took about 40-60 matches to get your target elo, which is why it worked at a chassis level and not at a variant level.

Edited by EgoSlayer, 30 December 2015 - 02:59 PM.

[Mystere]

Just scrap the public queues already and divide it between the Training Academy and Solaris:

• New players can learn in a much safer environment via the Training Academy.
  
• Via Solaris, grouped players can compete with other groups for placement in leaderboards. Solo players tired of carrying others can compete in N+ free-for-alls for individualized recognition.

Community Warfare, on the other hand, should be more lore-based and stripped of any eSports trimmings.

Finally, casuals have the freedom to chose where they want to go.

This is not rocket science.

Edited by Mystere, 30 December 2015 - 02:52 PM.

[Remains Intact]

EgoSlayer, on 30 December 2015 - 02:38 PM, said:

PSR takes hundreds or thousands of matches to get to your "true" PSR rating. There is no way it would work at a chassis level in its current incarnation. Elo by comparison took about 40-60 matches to get your target elo, which is why it worked at a chassis level and not at a variant level.

I would argue an intelligent person could come up with a system that can weight both your general psr, and a per chasis psr via some sort of weighted average until a large enough sample size can be obtained to only use the chassis psr.

Edited by Remains Intact, 30 December 2015 - 03:04 PM.

[Alistair Winter]

If I ever get to Tier 1 (the struggle grind is real), I would very much like to stay there. Even when piloting my Gauss Cicada.

[spartync]

EgoSlayer, on 30 December 2015 - 02:38 PM, said:

PSR takes hundreds or thousands of matches to get to your "true" PSR rating. There is no way it would work at a chassis level in its current incarnation. Elo by comparison took about 40-60 matches to get your target elo, which is why it worked at a chassis level and not at a variant level.

One way to handle this would be to look at your relative success per chassis and apply a multiplier (positive and negative) to your overall PSR.

The thing that would frustrate people is that we don't see the "math" behind the PSR. That would ultimately mean an unsatisfied player base complaining about whether they were being weighted appropriately in each chassis. You'd have folks who run into issues where they get JUST over the line to where a chassis they didn't think would qualify for a tier increase ... suddenly does and there would be chaos trying to calm people down.

The "XP bar" style that's in place works because of the overall lack of specificity and transparency in the process. Players might be upset, but the information provided is sufficiently vague to produce a mythology around things like the number of games you need to be successful in to move up in PSR...

[Love in an Annihilator]

Alistair Winter, on 30 December 2015 - 03:06 PM, said:

If I ever get to Tier 1 (the struggle grind is real), I would very much like to stay there. Even when piloting my Gauss Cicada.

I didn't notice a big difference. If you can reach it with the cicada, then you'll stay up.

OT: I think per chassis would be too complicated (for the MM), but we really should have different PSR per class.

[Remains Intact]

If it were designed well, your psr rating would have nothing to do with matchmaker load. Your psr would a pre determined stat in a spreadsheet that's just sitting there waiting to be used (just like it is now).

EDIT: It would just have many numbers to choose from
ex: Laser vomit baddass meta mech psr: 2
ex: All machine gun direwolf: 4

Edited by Remains Intact, 30 December 2015 - 03:22 PM.

[EgoSlayer]

spartync, on 30 December 2015 - 03:06 PM, said:

One way to handle this would be to look at your relative success per chassis and apply a multiplier (positive and negative) to your overall PSR.

The thing that would frustrate people is that we don't see the "math" behind the PSR. That would ultimately mean an unsatisfied player base complaining about whether they were being weighted appropriately in each chassis. You'd have folks who run into issues where they get JUST over the line to where a chassis they didn't think would qualify for a tier increase ... suddenly does and there would be chaos trying to calm people down.

The "XP bar" style that's in place works because of the overall lack of specificity and transparency in the process. Players might be upset, but the information provided is sufficiently vague to produce a mythology around things like the number of games you need to be successful in to move up in PSR...

The "mythology" about how long it takes comes straight from the source - Paul Inouye the guy who created it.

https://www.reddit.c..._tiers_and_psr/

With some more information here:
http://mwomercs.com/...-tiers-and-psr/