Jump to content


Tier 8 matchmaking stats


  • This topic is locked This topic is locked
6 replies to this topic

alyster #1 Posted 06 August 2017 - 05:45 PM

    Staff Sergeant

  • Beta Tester
  • 38425 battles
  • 354
  • [BB4] BB4
  • Member since:
    10-28-2010

I just got bored and made some notes for myself and now I figured to share it as well about the tiers MM gives u in tier 8 tanks. 

 

How I collected data:

  • Wrote down in the beginning of every game which tier game it was.
  • Only played tier 8s straight up, no higher or lower inbetween.
  • Didn't play any preferential MM tanks. Mainly played Object 416 and Pershing, with 1 Scorpion game as well I think.  

 

Totally I played 42 games and I'm bit tired of tier 8s so I won't climb up to 100 games, although I may add to this statistics later on. 

 

Odd thing is that I never got a single top tier 5/10 matchmaking out of 42 games?!?? I assume the 4 games top tier in 3/5/7 is because when you've been bottom tier for too long you will get some top tier games lol according to WG. 

 

My skills in probability and math are not the best so I couldn't compare the upper graph to to what should be probable average of each column. So I narrowed it down to the most annoying 3/5/7. Orange is actual games out of 25 games in 3/5/7 format. Number is given as a percentage. Pink is How probable are you to get bottom/mid/top tier (7/15*100=X)

 

 

To be honest when I started counting I expected alot worse. That being said tier 8 matchmaking sucks. Bring back the old MM. The 3 top tiers in the team are soooo [edited]precious and if they are bit .... tomato  then it's pretty hard to make up for their lack of participation in game when you're tier 8 most of the time. 

 

_______________

PS. If someone wants to add to these numbers so that the total ammount of games played only in tier 8 would be bigger, feel free to do so. Just would be better not to play any other tiers inbetween tier 8 games. 

 


Edited by alyster, 06 August 2017 - 05:49 PM.


SABAOTH #2 Posted 06 August 2017 - 05:47 PM

    Major

  • Player
  • 36437 battles
  • 2,913
  • [-133-] -133-
  • Member since:
    08-28-2011

NIce charts, you're hired.

 

On a serious note: yes is a bit better but still sucks big time. :(



NUKLEAR_SLUG #3 Posted 06 August 2017 - 05:58 PM

    Lieutenant

  • Player
  • 27494 battles
  • 1,947
  • Member since:
    06-13-2015
So what you're saying here is that on average the new MM works perfectly well and all the crap posts about always being bottom tier are just that.

alyster #4 Posted 06 August 2017 - 06:02 PM

    Staff Sergeant

  • Beta Tester
  • 38425 battles
  • 354
  • [BB4] BB4
  • Member since:
    10-28-2010

View PostNUKLEAR_SLUG, on 06 August 2017 - 06:58 PM, said:

So what you're saying here is that on average the new MM works perfectly well and all the crap posts about always being bottom tier are just that.

Not really. The second graph shows that mid tier MM is underwhelming and bottom tier is too common.



Balc0ra #5 Posted 06 August 2017 - 06:06 PM

    Field Marshal

  • Player
  • 64257 battles
  • 15,369
  • [WALL] WALL
  • Member since:
    07-10-2012
Well let's see if the new game mode that will pull more tier X's to it will fix the tier 8 MM a bit more as some suspect it might do. As of now, the few tier 8's I play "most with limited MM", sees more equal tier games then 5-10 or 3-5-7 "Eu2". And that's not ideal for the IS-6 etc. As I'm currently grinding... or trying to grind a new crew with it.

narHhallas #6 Posted 07 August 2017 - 08:43 AM

    Sergeant

  • Player
  • 24273 battles
  • 257
  • [BLUTE] BLUTE
  • Member since:
    08-12-2011
Ein guter Anfang.
Allerdings bezweifle ich, dass die Stichprobengrosse von 42 ausreichend ist wenn du Verteilungen hast wie beispielsweise 3/45 als toptier in 3:5/7.

Außerdem wird vermutlich auch die aktuelle Tier Verteilung auf dem Server eine Rolle spielen welche templates erzogen werden (hatte beim alten MM Einfluss und ich denke das wird hier nicht anders sein, da es sonst zu zu langen Wartezeiten kommen kann). Das bedeutet die Statistik muss hier auch über einen längeren Zeitraum geführt werden.

Ich kann die Grafiken leider nicht sehen, hast du hier auch Berechnungen zur Plausibilität angestellt? Falls ja auf welchem konfidenzniveau?

Edited by narHhallas, 07 August 2017 - 08:44 AM.


Asklepi0s #7 Posted 07 August 2017 - 08:49 AM

    Moderator

  • Moderator
  • 5641 battles
  • 839
  • [WG_M] WG_M
  • Member since:
    01-23-2017
That is a good thread indeed,how about share it in our pinned Match Making thread.
Thread is closed.





1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users