Jump to content

What Is The Point Of Cooperative In New Ranking System?


89 replies to this topic

#81 General Solo

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • 3,625 posts

Posted 22 December 2020 - 08:16 AM

plus its not the player bases job to fix these things
...just saying

In the mean time when playing in a pit of vipers
Be the mongoose
nom nom nom

#82 Capt Deadpool

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 305 posts

Posted 22 December 2020 - 12:39 PM

View PostGARION26, on 22 December 2020 - 07:58 AM, said:

Snip


I was trying to keep my post brief, but there are solutions to address your points using W/L. There are hundreds of pages of posts about the topic of you are interested, so no need to relitigate it, and you are of course entitled to your opinion in any case. The finer point, in this instance, was that there are ways to incentivize ONLY in-game behaviors that are conducive to winning (including many more subtle/nuanced ones than raw inaccurate damage output), and do not increase PSR for behaviors that are not conducive to winning, whether flat out exploits akin to Nightbird's example, or other ineffective tactics and criteria currently incorporated into match score.

#83 GARION26

    Member

  • PipPipPipPipPipPip
  • Giant Helper
  • 301 posts

Posted 22 December 2020 - 12:53 PM

Agree with much of what you have said.

I don't think the current matchscore is perfect and you can if you want to engage in 'perverse' behaviors that work against winning the game while trying to maximize your match score. You could also argue for tweaks on match score that better represent the contributions of mechs to the match.

Optimally behaviors that max your win rate would max your match score and would cause PSR rises- or allow whatever matchmaker system used to identify people who are really good or really bad at winning matches.

I think despite some very loud voices on this forum - it's pretty rare in the overall player base that someone is so focused on PSR rather then winning that they go down that "I don't care if I lose more, I'll sacrificing winning to max my PSR rises" route.

#84 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 22 December 2020 - 12:53 PM

View PostGARION26, on 22 December 2020 - 07:58 AM, said:

This system isn't perfect (match score or PSR) but I think it does a better job of sequentially winnowing the better players away from the worse players then WLD would.


There's enough evidence posted already to show you are 100% wrong on this.

#85 Gagis

    Member

  • PipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 1,731 posts

Posted 22 December 2020 - 01:04 PM

Note that Match Score formula has not changed. The score and the c-bill earnings have always rewarded the same kind of play as they do now.

On that note tho, the MS could probably benefit from both damage dealt and AMS missiles destroyed kickers being taken down a notch. They seem excessive compared to other stuff such as kills, kmdds and solo kills, which seem to be well in line with each other.

Edited by Gagis, 22 December 2020 - 01:05 PM.


#86 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 22 December 2020 - 01:08 PM

View PostGagis, on 22 December 2020 - 01:04 PM, said:

other stuff such as kills, kmdds and solo kills, which seem to be well in line with each other.


bold claim, got any proof?

#87 Gagis

    Member

  • PipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 1,731 posts

Posted 22 December 2020 - 01:37 PM

View PostNightbird, on 22 December 2020 - 01:08 PM, said:

bold claim, got any proof?

Only that the players don't seem to experience a strong incentive to focus on killing blows only thanks to kmdd's being relevant enough to balance that nor incentive to focus on kmdd's only thanks to solo kills being relevant enough. Only damage and AMS seem to stand out. It's actually quite amazing how seldom you hear anyone speak in terms of kill stealing in MWO, compared to other games. That's a very good sign.

Its MS, so ultimately fairly arbitrary and handwavy no matter what.

Edited by Gagis, 22 December 2020 - 01:37 PM.


#88 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 22 December 2020 - 02:21 PM

View PostGagis, on 22 December 2020 - 01:37 PM, said:

Only that the players don't seem to experience a strong incentive to focus on killing blows only thanks to kmdd's being relevant enough to balance that nor incentive to focus on kmdd's only thanks to solo kills being relevant enough. Only damage and AMS seem to stand out. It's actually quite amazing how seldom you hear anyone speak in terms of kill stealing in MWO, compared to other games. That's a very good sign.

Its MS, so ultimately fairly arbitrary and handwavy no matter what.


I see, I also agree that the c-bill and xp reward system is fairly good at making kill stealing not prevalent, but I don't think that's related to Match Score and Matchmaking weighting. There is no need for MS to be arbitrary, it's fairly simple to compute exactly what the size of each kicker should be relative to each other. We just have no data, so we don't know.

If we have n+1 data points available (n being the number of kickers we know exists) it's simple to compute the current kicker coefficients (linear algebra). Main problem is not feasible to collect enough data about what the kickers should be.

#89 Zaver11

    Member

  • PipPip
  • Little Helper
  • Little Helper
  • 25 posts

Posted 23 December 2020 - 07:49 AM

View PostNightbird, on 22 December 2020 - 02:21 PM, said:

I see, I also agree that the c-bill and xp reward system is fairly good at making kill stealing not prevalent, but I don't think that's related to Match Score and Matchmaking weighting. There is no need for MS to be arbitrary, it's fairly simple to compute exactly what the size of each kicker should be relative to each other. We just have no data, so we don't know.

If we have n+1 data points available (n being the number of kickers we know exists) it's simple to compute the current kicker coefficients (linear algebra). Main problem is not feasible to collect enough data about what the kickers should be.


What are the odds PGI is willing to release that information?

#90 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 23 December 2020 - 08:35 AM

View PostZaver11, on 23 December 2020 - 07:49 AM, said:

What are the odds PGI is willing to release that information?


0

Not necessarily because they don't want to, but they don't collect the info either and it would require work on their part to collect it. They're not willing to do any work for this.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users