Jump to content


Official Matchmaking Discussion Thread


  • Please log in to reply
2198 replies to this topic

j_rod #1901 Posted 04 November 2021 - 02:29 PM

    Forum Moral Police

  • Players
  • 36068 battles
  • 5,155
  • [III-C]
  • Member since:
    05-04-2011

First and foremost, I’m in complete agreement with VOP’s last post. The issue with skill based MM is that it punishes good play and incentivized bad play, driving everyone towards the mean. However, for the sake of discussion, I’ll play devil’s advocate and attempt to come up with a solution for better balance. 

 

Goal - to remove the most imbalanced teams (based on skill) from occurring and keep the skill levels between the two teams within a certain range

*It is not the goal of what I am proposing to create two evenly balanced teams, but rather to remove those instances where one team has a huge advantage over the other based on team skill*

 

Defining skill for the purpose of an algorithm - I propose defining as a combination of winrate and average damage (but I’m open to other suggestions) in order to come up with a numerical value that can be applied for a mathematical algorithm. For this Skill Value (SV), I would use the 90 day average for players with > 500 games in that timeframe  or the lifetime average for players with <500 games. To calculate the SV, I’d take ((Avg Damage + [5,000 * Winrate]) / 1000) with a max of 5 and a min of 2.5. So the lowest a player can score is 2.5 and the highest they can score is 5. The reason for this range is to avoid punishing the worst and best players disproportionately. 

 

Now, assuming the SV average to be approximately 3.5 (49% winrate with 1k avg damage), this means that the average SV total per team should be 24.5 and a range of 17.5 (2.5x7) to 35 (5x7). 

 

Determining an Acceptable Variation for Skill - Ok, so now that we have skill defined in a measurable way and know what the skill range would be, next we need to decide what’s “fair” from a skill discrepancy standpoint. If it’s too balanced, it disproportionally punishes skilled players and if it’s not balanced enough, it would be accomplishing it’s purported goal of minimizing skill disparity. While I could likely go find out what the Std Dev for the SV would be, I don’t have the time nor patience to do that right now, so it’s going to be my subjective opinion and my opinion is that if you were to take the difference between the top skill value of 5 and the low skill value of 2.5, the difference would be 2.5. So I’d take that difference, multiply it by 3 to come up with an acceptable skill disparity between teams of 7.5. 

 

Balancing the Teams - To start, MM would not change with how the teams are currently being balanced based on the class/tank type. So what this means is that if you took the total SV of two teams matched up using the current MM methodology and the difference between the two teams was 7.5 or less, then the teams would be within the acceptable range and the game would commence. Where MM would diverge from the current set up would be in the situations where the SV variation between the two teams was >7.5. In the event there was a disparity, the following would happen -

  • MM would pair two opposing players for the sake of analysis based on the following hierarchy -
    • High tier Heavy
    • High tier TD
    • High tier Med/Light
    • Low tier Heavy
    • Low tier TD
    • Low tier Med/Light
  • Then, MM would do a quick analysis based on the hierarchy to see if flipping two paired players brought the SV disparity back within an acceptable range. If it did, the flip would hold and the game would commence
  • If, after running this analysis/flip, the SV was still not within an acceptable range, MM would drop the low tier and refresh with a new high tier
    • The reason for this is that if the MM can’t make one change and balance the teams adequately, simply restarting the process or dropping top/low players on both teams is problematic 
    • This refresh would continue back and forth until two teams can be paired up adequately

 

Now, here are just a few of the challenges/concerns even with the suggestion above -

  1. How are platoons handled?
  2. What is a truly “fair” range of skill disparity between teams?
  3. Even if this does create more parity, is it actually addressing the frustration that many players have? I.e. - there are still going to be very bad players in the game and there are still going to be super-unicums
  4. What does this do to queue time? It’s designed to be relatively simple, but if there’s an impact and queue time is increased, it’s not like WG can just walk this back and remove it after designing it. 

 

Let me be perfectly clear that I am absolutely not advocating for this system. I’m in complete agreement with Cru and VOP and if anything, the issues that need to be addressed are trolls, AFKs and players that can’t get to a basic level (don’t crucify me on that one VOP). Trying to balance by skill is problematic and it’s like upping wages across the board - once you do it, you can’t really go back. 

 

That said, for those that continuously complain about MM, this is the type of suggestion/idea we’re looking for. If you want a reasonable debate/discussion, then put forward reasonable, thought out ideas. 


It was never fair... but it was fun! - Krietenstein34 
 

“There are no stupid questions, only stupid people.” - Mr Garrison (South Park)

 


Shaymin_Loves_You #1902 Posted 04 November 2021 - 02:33 PM

    First Sergeant

  • Players
  • 7213 battles
  • 1,582
  • Member since:
    02-07-2021

View Postj_rod, on 04 November 2021 - 07:29 AM, said:

First and foremost, I’m in complete agreement with VOP’s last post. The issue with skill based MM is that it punishes good play and incentivized bad play, driving everyone towards the mean. However, for the sake of discussion, I’ll play devil’s advocate and attempt to come up with a solution for better balance. 

 

Goal - to remove the most imbalanced teams (based on skill) from occurring and keep the skill levels between the two teams within a certain range

*It is not the goal of what I am proposing to create two evenly balanced teams, but rather to remove those instances where one team has a huge advantage over the other based on team skill*

 

Defining skill for the purpose of an algorithm - I propose defining as a combination of winrate and average damage (but I’m open to other suggestions) in order to come up with a numerical value that can be applied for a mathematical algorithm. For this Skill Value (SV), I would use the 90 day average for players with > 500 games in that timeframe  or the lifetime average for players with <500 games. To calculate the SV, I’d take ((Avg Damage + [5,000 * Winrate]) / 1000) with a max of 5 and a min of 2.5. So the lowest a player can score is 2.5 and the highest they can score is 5. The reason for this range is to avoid punishing the worst and best players disproportionately. 

 

Now, assuming the SV average to be approximately 3.5 (49% winrate with 1k avg damage), this means that the average SV total per team should be 24.5 and a range of 17.5 (2.5x7) to 35 (5x7). 

 

Determining an Acceptable Variation for Skill - Ok, so now that we have skill defined in a measurable way and know what the skill range would be, next we need to decide what’s “fair” from a skill discrepancy standpoint. If it’s too balanced, it disproportionally punishes skilled players and if it’s not balanced enough, it would be accomplishing it’s purported goal of minimizing skill disparity. While I could likely go find out what the Std Dev for the SV would be, I don’t have the time nor patience to do that right now, so it’s going to be my subjective opinion and my opinion is that if you were to take the difference between the top skill value of 5 and the low skill value of 2.5, the difference would be 2.5. So I’d take that difference, multiply it by 3 to come up with an acceptable skill disparity between teams of 7.5. 

 

Balancing the Teams - To start, MM would not change with how the teams are currently being balanced based on the class/tank type. So what this means is that if you took the total SV of two teams matched up using the current MM methodology and the difference between the two teams was 7.5 or less, then the teams would be within the acceptable range and the game would commence. Where MM would diverge from the current set up would be in the situations where the SV variation between the two teams was >7.5. In the event there was a disparity, the following would happen -

  • MM would pair two opposing players for the sake of analysis based on the following hierarchy -
    • High tier Heavy
    • High tier TD
    • High tier Med/Light
    • Low tier Heavy
    • Low tier TD
    • Low tier Med/Light
  • Then, MM would do a quick analysis based on the hierarchy to see if flipping two paired players brought the SV disparity back within an acceptable range. If it did, the flip would hold and the game would commence
  • If, after running this analysis/flip, the SV was still not within an acceptable range, MM would drop the low tier and refresh with a new high tier
    • The reason for this is that if the MM can’t make one change and balance the teams adequately, simply restarting the process or dropping top/low players on both teams is problematic 
    • This refresh would continue back and forth until two teams can be paired up adequately

 

Now, here are just a few of the challenges/concerns even with the suggestion above -

  1. How are platoons handled?
  2. What is a truly “fair” range of skill disparity between teams?
  3. Even if this does create more parity, is it actually addressing the frustration that many players have? I.e. - there are still going to be very bad players in the game and there are still going to be super-unicums
  4. What does this do to queue time? It’s designed to be relatively simple, but if there’s an impact and queue time is increased, it’s not like WG can just walk this back and remove it after designing it. 

 

Let me be perfectly clear that I am absolutely not advocating for this system. I’m in complete agreement with Cru and VOP and if anything, the issues that need to be addressed are trolls, AFKs and players that can’t get to a basic level (don’t crucify me on that one VOP). Trying to balance by skill is problematic and it’s like upping wages across the board - once you do it, you can’t really go back. 

 

That said, for those that continuously complain about MM, this is the type of suggestion/idea we’re looking for. If you want a reasonable debate/discussion, then put forward reasonable, thought out ideas. 


 

 

 

 getting rid of trolls and afks from the regular mm is something I completely agree with


Edited by Droodles_Little_Noodle, 04 November 2021 - 02:36 PM.

Unicum, Don't Unicum

Yesterday is history, tomorrow is a mystery, but today is a gift. That is why it is called the present.

The more you take, the less you have

 


j_rod #1903 Posted 04 November 2021 - 02:39 PM

    Forum Moral Police

  • Players
  • 36068 battles
  • 5,155
  • [III-C]
  • Member since:
    05-04-2011

View PostDroodles_Little_Noodle, on 04 November 2021 - 08:33 AM, said:

 getting rid of trolls and afks from the regular mm is something I completely agree with


Just some friendly advice - when replying back to a post, if you are only replying back to a small portion of the post, only quote that. Otherwise it  takes up half the screen for no reason. 

 

Or just reply back in a new post with something like - @j_rod - I totally agree that getting rid of trolls and AFKs from the regular MM is needed


It was never fair... but it was fun! - Krietenstein34 
 

“There are no stupid questions, only stupid people.” - Mr Garrison (South Park)

 


Shaymin_Loves_You #1904 Posted 04 November 2021 - 02:47 PM

    First Sergeant

  • Players
  • 7213 battles
  • 1,582
  • Member since:
    02-07-2021

View Postj_rod, on 04 November 2021 - 07:39 AM, said:


Just some friendly advice - when replying back to a post, if you are only replying back to a small portion of the post, only quote that. Otherwise it  takes up half the screen for no reason. 

 

 

 

got it


Unicum, Don't Unicum

Yesterday is history, tomorrow is a mystery, but today is a gift. That is why it is called the present.

The more you take, the less you have

 


JefeMac #1905 Posted 04 November 2021 - 09:08 PM

    Foro Fuego Jefe (Jubilado)

  • Players
  • 11655 battles
  • 2,470
  • [ROTI]
  • Member since:
    11-22-2014
Anyone know the criteria for mm during special mode events like Uprising that’s running right now? I’m waiting 5-13 mins on mm, esp in a toon, but Uprising gets me into a battle ASAP. Wondering what the difference is? 


 

“And yet when you hear people say ‘conspiracy theory’ that’s designed to pluck this little part of your brain that says, well, that thing’s not true,” said Attkisson. “And I always keep an open mind and say, that crazy thing that they say is a conspiracy theory may well have some truth in it.”
 


Shaymin_Loves_You #1906 Posted 04 November 2021 - 09:09 PM

    First Sergeant

  • Players
  • 7213 battles
  • 1,582
  • Member since:
    02-07-2021

View PostJefeMac, on 04 November 2021 - 02:08 PM, said:

Anyone know the criteria for mm during special mode events like Uprising that’s running right now? I’m waiting 5-13 mins on mm, esp in a toon, but Uprising gets me into a battle ASAP. Wondering what the difference is? 

Because of uprising, the player base of regular mm drops by a lot. while uprising has a majority of the playerbase playing it


Unicum, Don't Unicum

Yesterday is history, tomorrow is a mystery, but today is a gift. That is why it is called the present.

The more you take, the less you have

 


ruffseas #1907 Posted 04 November 2021 - 10:32 PM

    Sergeant

  • Players
  • 270 battles
  • 370
  • [ROH]
  • Member since:
    01-16-2015

View PostJefeMac, on 04 November 2021 - 01:08 PM, said:

Anyone know the criteria for mm during special mode events like Uprising that’s running right now? I’m waiting 5-13 mins on mm, esp in a toon, but Uprising gets me into a battle ASAP. Wondering what the difference is? 

Pretty sure fun modes use the similar MM to ratings. 


Bring back the ingame forum link!

 

"We've all heard that a million monkeys banging on a million typewriters will eventually reproduce the entire works of Shakespeare. Now, thanks to the Internet, we know this is not true.”Robert Wilensky


whatzup22 #1908 Posted 05 November 2021 - 12:25 AM

    First Sergeant

  • Players
  • 4043 battles
  • 1,555
  • Member since:
    03-25-2017

View PostDroodles_Little_Noodle, on 04 November 2021 - 09:40 PM, said:


I mean bad as in, panther 8,8 stg, t-34-3. those tanks are basically normal tanks because they're balanced

yea if they’re balanced tanks why not just play the tech tree counterpart (panther 2, t-34-2) . i dont know anyone who play ratings to grind credits or show off a shiny new tank. thats for normal mode 



Shaymin_Loves_You #1909 Posted 05 November 2021 - 02:10 PM

    First Sergeant

  • Players
  • 7213 battles
  • 1,582
  • Member since:
    02-07-2021

View Postwhatzup22, on 04 November 2021 - 05:25 PM, said:

yea if they’re balanced tanks why not just play the tech tree counterpart (panther 2, t-34-2) . i dont know anyone who play ratings to grind credits or show off a shiny new tank. thats for normal mode 

because panther 2 has worse dpm, but i'm not sure if its gotten buffed. 

And the t-34-2 has a way worse gun, and horrible armour. t-34-3 has wz-112 turret armour, and has like a 3.6 second aim time with a 122mm cannon


Unicum, Don't Unicum

Yesterday is history, tomorrow is a mystery, but today is a gift. That is why it is called the present.

The more you take, the less you have

 


whatzup22 #1910 Posted 05 November 2021 - 02:23 PM

    First Sergeant

  • Players
  • 4043 battles
  • 1,555
  • Member since:
    03-25-2017

View PostDroodles_Little_Noodle, on 05 November 2021 - 10:10 PM, said:

because panther 2 has worse dpm, but i'm not sure if its gotten buffed. 

And the t-34-2 has a way worse gun, and horrible armour. t-34-3 has wz-112 turret armour, and has like a 3.6 second aim time with a 122mm cannon

you just said these premium tanks are normal. no theyre not?? Which brings me back to my point that becuz most premiums are better than the tech tree, ratings should be limited to only techtree cuz fk premiums 



_Crusader6_ #1911 Posted 05 November 2021 - 02:27 PM

    St. Javelin: NLAW we Trust

  • Players
  • 91042 battles
  • 20,514
  • [III]
  • Member since:
    12-08-2014

View Postwhatzup22, on 05 November 2021 - 09:23 AM, said:

you just said these premium tanks are normal. no theyre not?? Which brings me back to my point that becuz most premiums are better than the tech tree, ratings should be limited to only techtree cuz fk premiums 


WG should just focus on releasing balance premiums...

 


I hate Annihilator spammers...  
Tank Hoarder: 526 tanks in Garage:  531/540 Played Tanks Aced
 
430 extra garage slots too - lucky me. 
 
I need more tanks...
 
    Wallet Warrior: Loyal Original M60 owner
 

 


Shaymin_Loves_You #1912 Posted 05 November 2021 - 03:07 PM

    First Sergeant

  • Players
  • 7213 battles
  • 1,582
  • Member since:
    02-07-2021

View Post_Crusader6_, on 05 November 2021 - 07:27 AM, said:


WG should just focus on releasing balance premiums...

 


They're making a few new balanced premiums, THE RAMPANZER, and of course, THE SUPERRRRRRHELLLCAT

 

Yes the stats will most likely be broken on the SUPPERRR HELLCATTT. But I have never seen someone actually play the tank well. 

 

also, tanks with like shovels on the front, should do more damage when ramming because memes


Edited by Droodles_Little_Noodle, 05 November 2021 - 03:17 PM.

Unicum, Don't Unicum

Yesterday is history, tomorrow is a mystery, but today is a gift. That is why it is called the present.

The more you take, the less you have

 


JefeMac #1913 Posted 05 November 2021 - 03:19 PM

    Foro Fuego Jefe (Jubilado)

  • Players
  • 11655 battles
  • 2,470
  • [ROTI]
  • Member since:
    11-22-2014

View Post_Crusader6_, on 05 November 2021 - 02:27 PM, said:


WG should just focus on releasing balance premiums...

 

Without STUPID consumables !!!!!!



 

“And yet when you hear people say ‘conspiracy theory’ that’s designed to pluck this little part of your brain that says, well, that thing’s not true,” said Attkisson. “And I always keep an open mind and say, that crazy thing that they say is a conspiracy theory may well have some truth in it.”
 


_Crusader6_ #1914 Posted 05 November 2021 - 03:32 PM

    St. Javelin: NLAW we Trust

  • Players
  • 91042 battles
  • 20,514
  • [III]
  • Member since:
    12-08-2014

View PostJefeMac, on 05 November 2021 - 10:19 AM, said:

Without STUPID consumables !!!!!!


It is my opinion that all of the super consumables be removed.

   Balance tanks properly - then if WG wants more diversity - have more tanks - and make variants with certain tradeoffs available.

 

 


I hate Annihilator spammers...  
Tank Hoarder: 526 tanks in Garage:  531/540 Played Tanks Aced
 
430 extra garage slots too - lucky me. 
 
I need more tanks...
 
    Wallet Warrior: Loyal Original M60 owner
 

 


whatzup22 #1915 Posted 05 November 2021 - 03:35 PM

    First Sergeant

  • Players
  • 4043 battles
  • 1,555
  • Member since:
    03-25-2017

View Post_Crusader6_, on 05 November 2021 - 11:32 PM, said:


It is my opinion that all of the super consumables be removed.

   Balance tanks properly - then if WG wants more diversity - have more tanks - and make variants with certain tradeoffs available.

 

 

annhilator: trade off? what trade off?



Shaymin_Loves_You #1916 Posted 05 November 2021 - 03:57 PM

    First Sergeant

  • Players
  • 7213 battles
  • 1,582
  • Member since:
    02-07-2021

View Post_Crusader6_, on 05 November 2021 - 08:32 AM, said:


It is my opinion that all of the super consumables be removed.

   Balance tanks properly - then if WG wants more diversity - have more tanks - and make variants with certain tradeoffs available.

 

 

I think there are a few balanced premiums in the game right now, panther 8,8,  stg, rampanzer, super heel cat, wz-111, turtle mk1, kunz panzer.


Unicum, Don't Unicum

Yesterday is history, tomorrow is a mystery, but today is a gift. That is why it is called the present.

The more you take, the less you have

 


JefeMac #1917 Posted 05 November 2021 - 08:04 PM

    Foro Fuego Jefe (Jubilado)

  • Players
  • 11655 battles
  • 2,470
  • [ROTI]
  • Member since:
    11-22-2014

View PostDroodles_Little_Noodle, on 05 November 2021 - 03:57 PM, said:

I think there are a few balanced premiums in the game right now, panther 8,8,  stg, rampanzer, super heel cat, wz-111, turtle mk1, kunz panzer.


Panther 8,8 is a hidden gem, especially with its terrific credit earning.



 

“And yet when you hear people say ‘conspiracy theory’ that’s designed to pluck this little part of your brain that says, well, that thing’s not true,” said Attkisson. “And I always keep an open mind and say, that crazy thing that they say is a conspiracy theory may well have some truth in it.”
 


Shaymin_Loves_You #1918 Posted 05 November 2021 - 08:06 PM

    First Sergeant

  • Players
  • 7213 battles
  • 1,582
  • Member since:
    02-07-2021

View PostJefeMac, on 05 November 2021 - 01:04 PM, said:


Panther 8,8 is a hidden gem, especially with its terrific credit earning.

I think people see it as bad, i see it as a well balanced tank. 


Unicum, Don't Unicum

Yesterday is history, tomorrow is a mystery, but today is a gift. That is why it is called the present.

The more you take, the less you have

 


JefeMac #1919 Posted 05 November 2021 - 08:09 PM

    Foro Fuego Jefe (Jubilado)

  • Players
  • 11655 battles
  • 2,470
  • [ROTI]
  • Member since:
    11-22-2014

View Post_Crusader6_, on 05 November 2021 - 03:32 PM, said:


It is my opinion that all of the super consumables be removed.

   Balance tanks properly - then if WG wants more diversity - have more tanks - and make variants with certain tradeoffs available.

 

 

Honestly - why freaking nerf tanks after/during testing them, then turn right around and give them some moronic consumable boost? 
 

Did you play any Burning Games or Uprising Cru? Tanks like the Emil were ABSOLUTE BEASTS in the 2nd/3rd life with the added speed and damage boosts that come with 2nd/3rd lives. It’s not like they don’t know how to balance characteristics or increase/decrease them….


Edited by JefeMac, 05 November 2021 - 08:14 PM.


 

“And yet when you hear people say ‘conspiracy theory’ that’s designed to pluck this little part of your brain that says, well, that thing’s not true,” said Attkisson. “And I always keep an open mind and say, that crazy thing that they say is a conspiracy theory may well have some truth in it.”
 


JefeMac #1920 Posted 05 November 2021 - 08:10 PM

    Foro Fuego Jefe (Jubilado)

  • Players
  • 11655 battles
  • 2,470
  • [ROTI]
  • Member since:
    11-22-2014

View PostDroodles_Little_Noodle, on 05 November 2021 - 08:06 PM, said:

I think people see it as bad, i see it as a well balanced tank. 

Totally concur. 



 

“And yet when you hear people say ‘conspiracy theory’ that’s designed to pluck this little part of your brain that says, well, that thing’s not true,” said Attkisson. “And I always keep an open mind and say, that crazy thing that they say is a conspiracy theory may well have some truth in it.”
 





4 user(s) are reading this topic

0 members, 4 guests, 0 anonymous users