Jump to content

The Great Psr Prophecy (With Graphs!)

Balance Gameplay

283 replies to this topic

#181 Nightbird

    Member

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

Posted 10 July 2020 - 05:57 AM

View PostAnomalocaris, on 10 July 2020 - 05:34 AM, said:

Hmm, the sweet zone doesn't seem to be expanding.....


It's gone up from 48.2 to 50.3 in the last two graphs. It's predicted to go up, just slowly compared to the WLR.

#182 MadcatX

    Member

  • PipPipPipPipPipPipPipPip
  • Big Brother
  • Big Brother
  • 1,026 posts

Posted 10 July 2020 - 06:35 AM

View PostNightbird, on 10 July 2020 - 05:57 AM, said:

It's gone up from 48.2 to 50.3 in the last two graphs. It's predicted to go up, just slowly compared to the WLR.


Faster would be better, I find things are a bit of a mess atm...

[EDIT]
Just was two matches (out of 4 today) with top 60 players from, a 4-man from 228th, JAY Z ironically, MW2 Annhiliator to name a few. This has happened since the PSR reset and it made sense at the time since we were all the same PSR, but now? I have absolutely no business playing matches with people of that caliber. I'm wondering, perhaps if I de-selected NA and European and exclusively played on oceanic only...

If team loses, my PSR goes down. If team wins, my PSR goes down because I'm matched against and with exceptional players. I am by no means bashing the top players, just the system that's matching me with them because the difference between my PSR and theirs must be big.

Gotta admit, the thought of running suicidal lights to force my PSR down to tier 4 just to avoid the top players is tempting, although I wouldn't... still tempting.

Edited by MadcatX, 10 July 2020 - 07:28 AM.


#183 Monkey Lover

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 7,918 posts
  • LocationWazan

Posted 10 July 2020 - 01:10 PM

View PostMadcatX, on 10 July 2020 - 06:35 AM, said:

This has happened since the PSR reset and it made sense at the time since we were all the same PSR, but now?

I hit tier2 in 4ish hours I can't see how these really good groups aren't teir1 right now. I think what we're really seeing is groups are coming back to the game . Now that the population is low, tier3 are playing with tier1 and as a lot of us said psr reset wasn't going to do much with a low population.

View PostMadcatX, on 10 July 2020 - 06:35 AM, said:

Gotta admit, the thought of running suicidal lights to force my PSR down to tier 4 just to avoid the top players is tempting, although I wouldn't... still tempting.

If you want to win more find a team to play with. Of course that's still not goign to stop you from getting smashed by top level teams or my win ratio would be lot higher lol

Edited by Monkey Lover, 10 July 2020 - 01:15 PM.


#184 Nightbird

    Member

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

Posted 10 July 2020 - 05:07 PM

I'm predicting that only 13% of the population will be outside of Tier 3 for 2 months so it's not gonna get better anytime soon.

Posted Image

#185 Monkey Lover

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 7,918 posts
  • LocationWazan

Posted 10 July 2020 - 05:17 PM

View PostNightbird, on 10 July 2020 - 05:07 PM, said:

I'm predicting that only 13% of the population will be outside of Tier 3 for 2 months so it's not gonna get better anytime soon.

Posted Image


I haven't really seen but based on the skill curve how many should be outside?

What I see in my matches it's really a handful of players that will tell me if this is going to be hard or next to impossible. Population so low easy to learn the names.

Edited by Monkey Lover, 10 July 2020 - 05:18 PM.


#186 Nightbird

    Member

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

Posted 10 July 2020 - 05:49 PM

Ideally 1/5 or 20% should be in each Tier when using a Tier system. The current PSR system doesn't put people in the right Tier anyways so it doesn't matter much.

Edited by Nightbird, 10 July 2020 - 05:52 PM.


#187 Monkey Lover

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 7,918 posts
  • LocationWazan

Posted 10 July 2020 - 09:06 PM

View PostNightbird, on 10 July 2020 - 05:49 PM, said:

Ideally 1/5 or 20% should be in each Tier when using a Tier system. The current PSR system doesn't put people in the right Tier anyways so it doesn't matter much.


Why an even number? Is skill in this game that even? I look at the leader board top few hundred players look to be the issue with the balance. These players can kill 1/4 or more of the team on thier own. You get two of these players on your team it's ggclose.

#188 warner2

    Member

  • PipPipPipPipPipPipPipPip
  • 1,101 posts

Posted 11 July 2020 - 03:49 AM

I am guessing many more people are average than great or terrible so the skill distribution isn't even. However, the PSR range for each tier could be adjusted to get a roughly similar number of people into each tier.

Edited by warner2, 11 July 2020 - 03:50 AM.


#189 Nightbird

    Member

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

Posted 11 July 2020 - 09:49 AM

View PostMonkey Lover, on 10 July 2020 - 09:06 PM, said:

Why an even number? Is skill in this game that even? I look at the leader board top few hundred players look to be the issue with the balance. These players can kill 1/4 or more of the team on thier own. You get two of these players on your team it's ggclose.


The PSR scale should be adjusted. If more people are in T3, then the range of T3 should be reduced to contain 1/5 the population. The whole point of tiers is to allow people to play inside their tier. If tiers haves so few people in them that you can't make matches inside them, then there is no point for tiers to exist.

My WLR suggestion effectively has no tiers.

#190 Nightbird

    Member

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

Posted 11 July 2020 - 07:59 PM

Posted Image

#191 OneTeamPlayer

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 399 posts

Posted 12 July 2020 - 08:41 AM

View PostMonkey Lover, on 10 July 2020 - 09:06 PM, said:

Why an even number? Is skill in this game that even? I look at the leader board top few hundred players look to be the issue with the balance. These players can kill 1/4 or more of the team on thier own. You get two of these players on your team it's ggclose.


The leaderboards aren't, put simply. They are some sort of scoreboard, but what they measure and how they do it is off.

I was #1 light for quite some time and i had quite a few kills, but kills aren't KMDD or Solo Kills and while i had quite a few of those as well, it was nowhere near the number of "kills" shown.

If a "kill" isn't a KMDD or solo it's not a player "killing" a mech, it's a player getting a shot off at the right time and confirming a kill (which while important isn't a solo metric).

As someone who has a lot of kills in my weight class i'm going to shout from the mountaintops that i consider any kill that I get that wasn't a solo kill a credit to the team, not me personally and solo kills were often made possible by something my team was doing (like distracting fire, a good push, or just a squirreling light).

It's telling that there's not any sort of "assists" column of stats in this supposedly "team" game.

Edited by OneTeamPlayer, 12 July 2020 - 08:42 AM.


#192 Tarl Cabot

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Tai-sho
  • Tai-sho
  • 7,771 posts
  • LocationImperial City, Luthien - Draconis Combine

Posted 12 July 2020 - 11:02 AM

The Leadership boards is able to sort by several different measures. The default selection is number of wins, but a user can select avg MS, Win/Loss or Kill/Death Ratio, number of loses, number of kills or deaths, and number of matches played.

PGI should add KMDD column, that would be interesting to see that number other than end of game results.

#193 Cluster Fox

    Member

  • PipPipPipPipPip
  • 104 posts
  • LocationStuck on a rock in Grim Plexus

Posted 13 July 2020 - 06:25 AM

View PostTarl Cabot, on 12 July 2020 - 11:02 AM, said:

PGI should add KMDD column, that would be interesting to see that number other than end of game results.


Along with that, AvgDMG per game or kill could be fun to have too. That could show kill efficiency.

#194 Nightbird

    Member

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

Posted 13 July 2020 - 08:39 PM

Posted Image

#195 Kroete

    Member

  • PipPipPipPipPipPipPip
  • 931 posts

Posted 14 July 2020 - 09:28 AM

View PostNightbird, on 10 July 2020 - 05:49 PM, said:

Ideally 1/5 or 20% should be in each Tier when using a Tier system. The current PSR system doesn't put people in the right Tier anyways so it doesn't matter much.

Juggling with numbers but missing the normal distribution?

The outliers (tier 1/5) are much smaller if you use 5 tiers, it would be more like 10%,20%,40%,20%,10% or you would missmatch a lot of players.

In the end it doesnt matter much,
because with that low playercount and 5 tiers all valves will be open most time.

Edited by Kroete, 14 July 2020 - 09:40 AM.


#196 Nightbird

    Member

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

Posted 14 July 2020 - 10:10 AM

View PostKroete, on 14 July 2020 - 09:28 AM, said:

Juggling with numbers but missing the normal distribution?


What normal distribution?

Posted Image

#197 Marquis De Lafayette

    Member

  • PipPipPipPipPipPipPipPip
  • CS 2023 Top 12 Qualifier
  • CS 2023 Top 12 Qualifier
  • 1,396 posts
  • LocationIn Valley Forge with General Washington

Posted 14 July 2020 - 10:51 AM

View PostNightbird, on 13 July 2020 - 08:39 PM, said:

Posted Image


I very much appreciate the work you are putting in here. I know it’s early on, but it’s a bit surprising to me that we have more of the playerbase with a .7 w/l (losing significantly more than winning) than we have with a 1.1 w/l (winning just a bit more than losing). That looks harsher than I would have thought (again it’s early though)...maybe it’s due to the dropping of everyone in tier 3 off the jump? Any thoughts as to why?

#198 Kroete

    Member

  • PipPipPipPipPipPipPip
  • 931 posts

Posted 14 July 2020 - 11:07 AM

View PostNightbird, on 14 July 2020 - 10:10 AM, said:

What normal distribution?

Posted Image

Looks a little like the blue line? Posted Image
Posted ImageBut red would be better with 5 tiers and matching the 10/20/40/20/10 i suggested.

Edited by Kroete, 14 July 2020 - 11:21 AM.


#199 Nightbird

    Member

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

Posted 14 July 2020 - 11:22 AM

View PostKroete, on 14 July 2020 - 11:07 AM, said:

Looks a little like the blue line ... Posted Image


You're not going to get a normal distribution from Jay Z's PSR system. It's not based on stats.

#200 Nightbird

    Member

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

Posted 14 July 2020 - 11:28 AM

View PostMarquis De Lafayette, on 14 July 2020 - 10:51 AM, said:

I very much appreciate the work you are putting in here. I know it’s early on, but it’s a bit surprising to me that we have more of the playerbase with a .7 w/l (losing significantly more than winning) than we have with a 1.1 w/l (winning just a bit more than losing). That looks harsher than I would have thought (again it’s early though)...maybe it’s due to the dropping of everyone in tier 3 off the jump? Any thoughts as to why?


The population <=0.7 is about the same as the ones >=1.3. That the peak is at 0.9 instead of 1.0 is disappointing though. It's because people who lose often quit the game. If everyone played and no one quit, then peak would be at 1. Get stomped 10 times in a row like many Tier 3 newbies are and they inflate the WLR of skilled players and after quitting early they never show in the chart for the low range.

Edited by Nightbird, 14 July 2020 - 11:29 AM.






10 user(s) are reading this topic

0 members, 10 guests, 0 anonymous users