Jump to content


MM is a Ridiculous Joke


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

Enzo_Gorlami #61 Posted 26 June 2017 - 01:36 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

View PostSpartacusDiablo, on 26 June 2017 - 02:22 AM, said:

I've noticed that when I play really bad my win rate tends to decline. I've also noticed that when I play well my win rate tends to increase.

Strange...

 

 

:trollface:



Enzo_Gorlami #62 Posted 26 June 2017 - 02:19 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

View PostCptCheez, on 24 June 2017 - 08:55 PM, said:

 

You're right, it is.  MM, please don't put this guy on my team ever.

 

It already is random.

 



Enzo_Gorlami #63 Posted 26 June 2017 - 06:34 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

View PostCptCheez, on 25 June 2017 - 04:51 PM, said:

 

Actually, that's completely wrong.  Any statistician will tell you that a random matchmaker will MORE often result in unbalanced games than balanced games.  Look at this post of mine from 2 years ago, where I explain exactly why math backs up the fact of a random matchmaker in this game:

 

 

 

It still has the effect of setting you up for a 85-90 percent chance of a loss by putting you and maybe one other with 5 43'ers. These are the games I get. Very likely, you have the benefit of not drawing those assignments. With random assignments for a tier, at least you are not set up for failure. 

cheasesteak #64 Posted 26 June 2017 - 06:49 PM

    First Sergeant

  • Players
  • 18503 battles
  • 3,993
  • [COD-R]
  • Member since:
    11-15-2014

I don't think you really are getting what CptCheez is getting at.  It is pretty clear from even his quick and dirty experiment that random variation will give you lopsided games.  That is one of the things you cited as evidence that things were NOT random.  You said: "My point was that if it were really random for a tier, we would not have lopsided games where one side is stacked and one side is a joke​" but as the good Captain showed, this is not the case. 

 

 


 

 

Overpriced Lackey to the Barons of Entrenched Corporate Greed


Enzo_Gorlami #65 Posted 26 June 2017 - 07:06 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

View Postcheasesteak, on 26 June 2017 - 06:49 PM, said:

I don't think you really are getting what CptCheez is getting at.  It is pretty clear from even his quick and dirty experiment that random variation will give you lopsided games.  That is one of the things you cited as evidence that things were NOT random.  You said: "My point was that if it were really random for a tier, we would not have lopsided games where one side is stacked and one side is a joke​" but as the good Captain showed, this is not the case. 

 

 

 

He makes a case for where random assignments would result in more lopsided games. Maybe so, however, most people are in the average median, with a WR of from 30 to 70-80 pool to choose from, not 0-100. MM remains a black box routine, and I stand by my earlier statements. 

noobest_noob #66 Posted 26 June 2017 - 07:28 PM

    Sergeant

  • Players
  • 11 battles
  • 462
  • Member since:
    09-26-2016

okok, thanks WG

I need to stop right now...maybe a week :arta:

 

 

 



40_Percenter #67 Posted 26 June 2017 - 07:54 PM

    Senior Sergeant

  • Players
  • 29689 battles
  • 677
  • [III-R]
  • Member since:
    12-30-2016
I'm sorta with the OP.  I got savaged yesterday.  I had two 10 game losing streaks in one day.  It was the worst day of tanking ever.  Nothing I did changed my luck.
Nobody likes 40 Percenters.  

CptCheez #68 Posted 26 June 2017 - 08:00 PM

    This Space Intentionally Left Blank

  • Players
  • 15398 battles
  • 6,798
  • [III]
  • Member since:
    07-08-2014

View PostEnzo_Gorlami, on 26 June 2017 - 03:06 PM, said:

He makes a case for where random assignments would result in more lopsided games. Maybe so, however, most people are in the average median, with a WR of from 30 to 70-80 pool to choose from, not 0-100. MM remains a black box routine, and I stand by my earlier statements. 

 

I never said anything about WR in that dice post.  Not a single time.  My arbitrary scale of "1-10" rates their *skill level*, not their WR.  And it is absolutely true that there are people playing this game from all skill levels, from 1-10.  Or 1-100, whatever scale you want to use.

 

Oh and "average median" is a nonsensical phrase.  Those two words mean absolutely nothing when used together like that.  You can talk about average WR or median WR, but "average median" is not a thing.


Edited by CptCheez, 26 June 2017 - 08:01 PM.

"When the going gets tough and the stomach acids flow, 
The cold wind of conformity is nipping at your nose.
When some trendy new atrocity has brought you to your knees
Come with us we'll sail the Seas of Cheese."


Enzo_Gorlami #69 Posted 26 June 2017 - 08:33 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

View PostCptCheez, on 26 June 2017 - 08:00 PM, said:

 

I never said anything about WR in that dice post.  Not a single time.  My arbitrary scale of "1-10" rates their *skill level*, not their WR.  And it is absolutely true that there are people playing this game from all skill levels, from 1-10.  Or 1-100, whatever scale you want to use.

 

Oh and "average median" is a nonsensical phrase.  Those two words mean absolutely nothing when used together like that.  You can talk about average WR or median WR, but "average median" is not a thing.

 

MM does use WR and other statistics to place players, from what I have seen. Your dice roll analogy cannot be used in comparison to MM's programmed selection criteria. Until I see the code and rationale behind MM's game selection criteria, we can not begin to make a comparison,.All we have so far are the end results to look at.  I have been programming for over 35 years, and I can make educated guesses regarding the code utilized in MM. 

Edited by Enzo_Gorlami, 26 June 2017 - 08:34 PM.


__Crusader6__ #70 Posted 26 June 2017 - 08:42 PM

    BANNED

  • Players
  • 56146 battles
  • 10,317
  • [III]
  • Member since:
    12-08-2014

View PostEnzo_Gorlami, on 26 June 2017 - 03:33 PM, said:

 

MM does use WR and other statistics to place players, from what I have seen. Your dice roll analogy cannot be used in comparison to MM's programmed selection criteria. Until I see the code and rationale behind MM's game selection criteria, we can not begin to make a comparison,.All we have so far are the end results to look at.  I have been programming for over 35 years, and I can make educated guesses regarding the code utilized in MM. 

 

No. 

 

Don't quit your day job. 

 

 


Tank Hoarder: 352 tanks in Garage:  355/355 aced (AMX 30B, Mk1 and T49A repo),    wallet warrior.  Loyal M60 owner
I think 5.5 was good for the game - But I also want my Kenny OP nerfed

 


CptCheez #71 Posted 26 June 2017 - 08:44 PM

    This Space Intentionally Left Blank

  • Players
  • 15398 battles
  • 6,798
  • [III]
  • Member since:
    07-08-2014

View PostEnzo_Gorlami, on 26 June 2017 - 04:33 PM, said:

I have been programming for over 35 years, and I can make educated guesses regarding the code utilized in MM. 

 

So you claim, yet you don't know the difference between a "median" and an "average"...

 

MM absolutely does not use "WR and other statistics" to place players.  How is it that people are still making this ridiculous statement when WG has flat out told us otherwise in multiple places, from multiple different sources?

 

Their Wiki

Their News Page

The original Producer's blog

 

3 different WG sources.  All stating that the Blitz matchmaker does NOT use "player's statistics".  It's written right there in black and white.  There's no mistaking what they've said.  And yet you're still doing it.  If you can read 3 different things that say they don't use "player's statistics", but yet your "educated guesses" still include player's statistics, well...that says something about your education.


Edited by CptCheez, 26 June 2017 - 08:46 PM.

"When the going gets tough and the stomach acids flow, 
The cold wind of conformity is nipping at your nose.
When some trendy new atrocity has brought you to your knees
Come with us we'll sail the Seas of Cheese."


Enzo_Gorlami #72 Posted 26 June 2017 - 08:50 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

View PostCptCheez, on 26 June 2017 - 08:44 PM, said:

 

So you claim, yet you don't know the difference between a "median" and an "average"...

 

MM absolutely does not use "WR and other statistics" to place players.  How is it that people are still making this ridiculous statement when WG has flat out told us otherwise in multiple places, from multiple different sources?

 

Their Wiki

Their News Page

The original Producer's blog

 

3 different WG sources.  All stating that the Blitz matchmaker does NOT use "player's statistics".  It's written right there in black and white.  There's no mistaking what they've said.  And yet you're still doing it.  If you can read 3 different things that say they don't use "player's statistics", but yet your "educated guesses" still include player's statistics, well...that says something about your education.

Since you have trouble understanding it, let me google it for you. 

Meanmedian, and mode are three kinds of "averages". ... The "mean" is the "average" you're used to, where you add up all the numbers and then divide by the number of numbers. The "median" is the "middle" value in the list of numbers.

 

I have seen enough of MM that it is not what you think it is. But please don't let me stop you for being WG's MM apologist. 



Enzo_Gorlami #73 Posted 26 June 2017 - 08:52 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

View Post__Crusader6__, on 26 June 2017 - 08:42 PM, said:

 

No. 

 

Don't quit your day job. 

 

 

 



BakalovBoy #74 Posted 26 June 2017 - 08:56 PM

    First Sergeant

  • Players
  • 17935 battles
  • 1,029
  • [O_C_S]
  • Member since:
    06-08-2016
How is this thread even still going ?

So what if MM sucks. WG will not do anything about it and ignore all our complaints anyways.

Ericmopar #75 Posted 26 June 2017 - 08:56 PM

    First Sergeant

  • Players
  • 3261 battles
  • 2,393
  • Member since:
    04-10-2017

Then how is it, that a person can play for hours and almost every team they are on splits up and wanders, and exactly the opposite happens with the red team, repeatedly for hours on end?

It was like that all day yesterday for me, almost every red team stuck together and wolfpacked, almost every green team did the opposite, they wandered like they had played a handful of matches and did almost nothing but drive into walls etc.

To be blunt the red teams acted like they knew each other. Whether they did or not, I can't say, but that's how they were playing.

It went on like that for hours, and I've witnessed that many times.

I'm getting a feel for it now, and just quit playing for a few hours. If it's still persistent later in the day, then I quit until the next day, when things are almost always better.

 

And the match maker is broke as hell.

I mean seriously, it's a simple thing to see that there are 5 heavies on one team and one or none on the other. It just a matter of the program moving a couple of tanks from one team to the other in the MM. Programmers I've talked to say something like that would be easy to do.

 


Edited by Ericmopar, 26 June 2017 - 09:00 PM.

Old and Treacherous Flying Monkey.

 

 


__Crusader6__ #76 Posted 26 June 2017 - 09:11 PM

    BANNED

  • Players
  • 56146 battles
  • 10,317
  • [III]
  • Member since:
    12-08-2014

Ericmopar, 

 

It's called directing. 

 

All Left <------- Please 

All Right -----> Please 

 

I have them and many others hotkeyed.  

 

40%'ers to Town 

Listen to me I'm a Unicum

 

 

Yet sometimes the team still scatters. 

I had a clown yolo in his IS-6 today - he's a dumb seal clubber out of his element past tier 5...

 

You need to do your best and accept you can't win them all. 

 


Tank Hoarder: 352 tanks in Garage:  355/355 aced (AMX 30B, Mk1 and T49A repo),    wallet warrior.  Loyal M60 owner
I think 5.5 was good for the game - But I also want my Kenny OP nerfed

 


Whiskey_Upsurge #77 Posted 26 June 2017 - 09:11 PM

    First Sergeant

  • Players
  • 34035 battles
  • 1,209
  • [SPUDX]
  • Member since:
    06-27-2014

Personally, I don't know what everyone's on about when they say MM is broken. I get that you may prefer +/-2 matches, that's fine, but the matches we have now and the +/-1 format isn't broken.

 

Over a large enough sample size, we all get teams with better or worse players, teams with well balanced tank types, mates that work well together as a team etc.

 

During that sample however, we'll all get streaks where we can't seem to lose and other streaks where we can't seem to do anything. Neither of them means that the system is flawed. It's just the nature of it. 

 

In the long run, if your WR isn't increasing and your stuck in the mid-40's, it's not the teams. It's you. Simple


Inspiration for WriterDude's forum Signature is my single greatest accomplishment since starting to play Blitz! 

__Crusader6__ #78 Posted 26 June 2017 - 09:20 PM

    BANNED

  • Players
  • 56146 battles
  • 10,317
  • [III]
  • Member since:
    12-08-2014

Everyone gets the same MM.  

 

It's how you adapt that differentiates players.  

 

No one wins them all - but better players will win more. 

 

 

 

 


Tank Hoarder: 352 tanks in Garage:  355/355 aced (AMX 30B, Mk1 and T49A repo),    wallet warrior.  Loyal M60 owner
I think 5.5 was good for the game - But I also want my Kenny OP nerfed

 


Enzo_Gorlami #79 Posted 26 June 2017 - 09:23 PM

    First Sergeant

  • Players
  • 31387 battles
  • 2,015
  • [BELLS]
  • Member since:
    01-03-2017

"Everyone gets the same MM. "

 

And you know this for a fact?


Edited by Enzo_Gorlami, 26 June 2017 - 09:46 PM.


cheasesteak #80 Posted 26 June 2017 - 09:34 PM

    First Sergeant

  • Players
  • 18503 battles
  • 3,993
  • [COD-R]
  • Member since:
    11-15-2014

View PostEnzo_Gorlami, on 26 June 2017 - 09:23 PM, said:

And you know this for a fact?

 

​Good grief!  Look at his periodic damage ratio and then compare to WR.  Then go to wobstars history and look at yours.  There is a correlation.  Look at this for just about any player, and you will see that in general stronger WR correlates to better damage ratio.  This is true when comparing players, and when comparing an individual player's periodic statistics.  There are of course outliers, but the trend is there.

 

Bottom line, play better (i.e. higher damage ratio), win more. 

 

edit - Just to note that WR and DR are likely not fully independent.


Edited by cheasesteak, 26 June 2017 - 09:36 PM.

 

 

Overpriced Lackey to the Barons of Entrenched Corporate Greed





1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users