16 replies to this topic

[JC Daxion]

I read the FW patch notes and he made a hint about a better PSR system.


I get that it is working atm.. vets, verse vets.. But the problem is great vets stacked on one side. I dunno if this is a simple fix or not. but how about...


tier, step one

something similar to jarl's as step 2.. If you are always up at the top for months on end, you get a rating, and a good one..

BV, the mech you are in.. If you are in a total crap mech, maybe that drops you a little.



the match maker takes 24 people, then starts layering them. 2 top 10 players, different teams.. 2 top 50.. and one top 10... maybe its one top 10 verse 2 top 50... The idea being, a bit more to the MM, and maybe stack matches every few matches.. aka instead of letting a match go every 2 mins.. maybe its 5, sorta like solaris. Is this a match every 2 mins the bigger issue, or is it better that we wait a little for something a bit more even?


what do you gues think?

[K O Z A K]

this would only work with an solos only que, FP is a group que with solos filling the empty spots

[Monkey Lover]

Its not vets vr vets. Its groups ver groups and vets fill in the slots. Solo v solo isnt really cared about in a group based mode.

[justcallme A S H]

Yep. There was a video



It says it will be based on ELO. We've now found out that it's based on PSR. PSR is broken, therefore I cannot see how FP MM is ever going to work if it's based / using - totally useless input in terms of player skill.

We all know PSR is not representative of player skill. Just how many games you play, which means you can be low skill and make it into Tier 1.

Edited by justcallme A S H, 11 April 2019 - 07:26 PM.

[FRAGTAST1C]

Is the PSR separate from QP? I have played only 1 FW match before the PTS (where I played 4) and I'm Tier 3. There is no way I would be a Tier 3 FW player.

[K O Z A K]

justcallme A S H, on 11 April 2019 - 07:26 PM, said:

Yep. There was a video



It says it will be based on ELO. We've now found out that it's based on PSR. PSR is broken, therefore I cannot see how FP MM is ever going to work if it's based / using - totally useless input in terms of player skill.

We all know PSR is not representative of player skill. Just how many games you play, which means you can be low skill and make it into Tier 1.

PSR is garbage for skill rating, I really don't understand the effort that was put towards creating a FP matchmaker, while using a totally useless input for the matchmaking process

[justcallme A S H]

FRAGTAST1C, on 11 April 2019 - 07:41 PM, said:

Is the PSR separate from QP? I have played only 1 FW match before the PTS (where I played 4) and I'm Tier 3. There is no way I would be a Tier 3 FW player.

No.

It is using QP data... That is the entire issue. There are people that play FP that are high skill, but never play QP and thus are Tier 3-4... And you have 1,000s of low skill users in Tier 1. It's non-workable as a basis for a match maker (the PSR data)

[Prototelis]

Don't worry OP.

I too was stoked about the possibility of a new matchmaker that could eventually be adapted to QP.

Apparently not going to happen.

[PhoenixFire55]

JC Daxion, on 11 April 2019 - 05:10 PM, said:

I read the FW patch notes and he made a hint about a better PSR system.

Thats not even in Soon^TM category yet. Judging by previous experience ... in 3-4 years maybe. So effectively never.

[Grus]

justcallme A S H, on 11 April 2019 - 10:44 PM, said:

No.

It is using QP data... That is the entire issue. There are people that play FP that are high skill, but never play QP and thus are Tier 3-4... And you have 1,000s of low skill users in Tier 1. It's non-workable as a basis for a match maker (the PSR data)

Never understood why they even went with that's as a way to determine player skill... experience? Sure. Not skill.

[yrrot]

Using PSR at least means the FP matches shouldn't be drastically worse than normal QP matches we have now. Of course, it does have a potential exploit for people that only care about winning FP matches to try to keep their PSR low to get easier FP matches.

The nice changes here are that smaller groups and solos can get into FP as fill and be less likely to end up going against 12 mans. If Paul ends up using something more Elo-like based on FP performance instead of PSR, great. Perhaps bringing in some of their previous Elo implementation into the improved matchmaker?

Either way, we at least get faster matchmaking (potentially) for FP and perhaps more consistent matches than we might have otherwise. Net gain is still better than the status quo.

Edited by yrrot, 12 April 2019 - 07:27 AM.

[Knight Captain Morgan]

Hazeclaw, on 11 April 2019 - 05:31 PM, said:

this would only work with an solos only que, FP is a group que with solos filling the empty spots

That’s what the proposed changes will be. Currently FW is groups slaughtering solos.

[Grus]

Just had a thought. If say a a group of 12 people on one side is comprised of the top 10% in the que.. and another group of 12 on the other side is comprised of the bottom 10%.. are both going to wait for ever to get a match? Or are they going to be dropping against each other? If the latter, then what's the point of a MM?

[Racerxintegra2k]

Anyone else play Counter Strike way back ? They had a matchmaker that if your team crushed the other team, it would activity switch players to the side that was wrecked so it wouldn't happen again. Would anyone mind playing against a team mate or two in the name of a more balanced fun match ?

[Dee Eight]

justcallme A S H, on 11 April 2019 - 10:44 PM, said:

No.

It is using QP data... That is the entire issue. There are people that play FP that are high skill, but never play QP and thus are Tier 3-4... And you have 1,000s of low skill users in Tier 1. It's non-workable as a basis for a match maker (the PSR data)

Not to mention all the players who either because they're infamous streamers or repeatedly banned, keep creating new accounts... the former are usually skilled enough they SHOULD NOT BE mixing with tier 4 & 5 players and the later shouldn't be allowed to be playing at all, especially not the ones who continue to do the things that got their previous accounts banned.

Edited by Dee Eight, 12 April 2019 - 06:06 PM.

[Feral Clown]

Dee Eight, on 12 April 2019 - 06:01 PM, said:

Not to mention all the players who either because they're infamous streamers or repeatedly banned, keep creating new accounts... the former are usually skilled enough they SHOULD NOT BE mixing with tier 4 & 5 players and the later shouldn't be allowed to be playing at all, especially not the ones who continue to do the things that got their previous accounts banned.

That's a great point. I have many accounts and none of them should be mixing with lower tiers. I flipped between gimping myself to now just getting out of lower tiers as fast as possible. Be great if PGI let me link all of my accounts and automatically put them all in tier one. I have seen first hand alt accounts that are either predatory or outright trolls.

[Asym]

Grus, on 12 April 2019 - 05:34 AM, said:

Never understood why they even went with that's as a way to determine player skill... experience? Sure. Not skill.

Because skill and experience are "relative Grus".... If you play with the top 10% of any sport, your skill level will reflect that level of play... Changes in modes of play, mechs and weapons (nerfs and buffs) affect skill rating a lot !

There is an answer, but in implementing a Combat Effectiveness (CE) metric into an ELO platform will NOT give you what you want "from old data...." Like I have said before, CE has been programmed and implemented in several prototype games I've had the opportunity to demo and "it's (CE) not for the easily affected or thinned skinned types of players..." When you actually measure "individual effectiveness", the results can be pretty harsh.... PSR is trash. A great many elite players are elite because of the teams they built their skills lists from......and, as individuals are average players.... we see this every event where a "big name" wanders in to play a match or two and us potatoes wonder what the hype is about as they get massacred early doing noob things....... And, just the opposite: a noob on comms is an outright death sentence at tier 4.....???

CE actually measure the effectiveness of "what you do with weapons".... Not just aim. Not just damage. Not just winning or losing and the results, in my experience are a bit humbling......

This game couldn't use CE. The population wouldn't support it and the players would have to start all over and build their resumes from scratch..... It'd take too long and with so few players, simply would be value added to do..........