Jump to content


how come you don't take skill/win rate into account in matchmaking


  • Please log in to reply
1967 replies to this topic

SoTrue #1941 Posted Mar 19 2018 - 05:30

    Major

  • Players
  • 33309 battles
  • 3,302
  • [ATTOP] ATTOP
  • Member since:
    04-01-2011

View PostNeatoMan, on Mar 18 2018 - 20:05, said:

Because your SBMM is just balance for balance sake.  It doesn't do a damn thing about what most people are complaining about (blowouts).

 

All your SBMM does is change who wins the game.  You claim it's not about the END RESULT,  However, wins and losses ARE the END RESULT.  Your are exclusively focusing on the END RESULT without even realizing it.

 

 I see all.  ;)

 

 

Ok,

A) It's Sunday night, you should be playing the game right not, now rambling on the forums.

B) Sbmm is not 'balancing for the sake of balancing', it's about creating a level playing field at the start of every battle.  I am a competitive recreational cyclist.  Not a very good one.  But there is a category I am competitive at.  So I race in it.  Can I enter the expert or pro races?  Actually, at the local level, yes I can.  Do I?  Of course not.  Entering a 'competitive' event where I have no chance to win is just stupid.  Same with the crap battles random mm dishes out.  When the battle screen comes up and all the skill is on one team, that's as stupid as me racing the bike pros.  It's not fun for me.  The pros take no joy in beating a random 'fred'.

C)  I will agree that many who complain about random are 'too' focused on blowouts.  But you know my position on that.  There are two kinds of blowouts.  Snowball (what most think of when you say blowout).  And Steamroll.  There is a huge difference between the two.  Snowballs can only occur between two roughly similar teams.  One team gets an early advantage and it snowballs into a blowout.  A steamroll can only occur between a skilled and unskilled team.  The skilled team simply steamrolls over the less skilled team.  The skilled team doesn't 'create an early advantage', it's built into the mix by the MM.  By putting all the skill on one team, MM has decided the winner.

D)  This goes back to the previous post, where you try and claim 'sometimes the less skilled team wins' - as if to say the rare exception somehow eliminates the 99% of the times where the lesser skilled team loses.  Once in a blue moon, the lessor skilled team does manage to win.  It's almost always down to the skilled players on the skilled team dropping the ball.  Let's look at an example.  In the battle below, these are the teams:

                   my team        enemy team

blue                   0                       2

green                 1                       1

yellow                3                       5                    above this line are skilled players

orange               7                       6                    below this line are unskilled players

red                     4                       1

You can clearly see my team has (4) skilled players, and the enemy has (8) skilled players, included 2 blue unicums.   They should have won this easy, and most of the time they do.  Why didn't they then?  Their unicums made bad plays.  They are human, it happens.  To be clear, it's rare, but it does happen.  The ARL 44 went valley with the lemmings, and the Thunderbolt camped back of map to snipe.  Had they both gone city, like they should have, they would have rolled the city easy.

But look, just because this happens once in a blue moon, that DOES NOT NEGATE the vast majority of the times where the more skilled team rolls over the less skilled team.  I mean, come on, they didn't even scout mid.  Their heavies all went valley.  The skilled team threw this away.

 



EmperorJuliusCaesar #1942 Posted Mar 19 2018 - 05:36

    Major

  • Players
  • 37052 battles
  • 5,741
  • [EOR] EOR
  • Member since:
    03-16-2014

View PostSoTrue, on Mar 18 2018 - 19:39, said:

View PostNeatoMan, on Mar 17 2018 - 12:31, said:

But whenever the underdog wins it cannot possibly be a crap game.  Also, draws and games that come down to the wire cannot be considered crap games either.  Therefore a large portion of those games you label crap are in fact not crap.

 

You claim it's not all about XVM, yet you keep retreating to that 40% value which is based solely on XVM.  You have absolutely no other identifier of crap other than what XVM tells you.  Your OCD with XVM does not deserve to be imposed onto rest of the player base.

 

Um, yes.. even when the underdog wins it can still be a 'crap' game.  I've done it myself on occasion.  My team is horrible, but I still squeak out a win.  It's almost entirely luck though.  Or a major heavy carry.  Just because a grossly mismatched battle ends 'close', doesn't negate the fact that it was grossly mismatched.  That's bs logic right there.  Heck, just this week in the NCAA, the number 16 upset the number 1.  But it's the first time in NCAA history.  Does that suddenly mean every time the 16 seed has a good chance to upset the number 1 seed?  Of course not.  For all eternity, the number 1 seed is going to have more than good odds to win against the number 16 seed.  You are trying to prove something wrong using a tiny exception.  That is why you are wrong.

-

Also, my 'range' of 40-60% is not 'based on' xvm.  All xvm does is show me the relative skill of each team - and THAT SKILL RATING IS REAL.  That one team has 8 greens, and the other team has 2 - that's all random mm's fault.  And it's a REAL skill imbalance.  To deny otherwise is to deny reality.


 

 

"To deny otherwise is to deny reality."

 

They seem to be quite good at doing just that.



Nixeldon #1943 Posted Mar 19 2018 - 08:42

    Major

  • Players
  • 60879 battles
  • 2,218
  • Member since:
    10-30-2011

View PostEmperorJuliusCaesar, on Mar 18 2018 - 22:11, said:

That's not what we've said at all, not even once.

That is exactly what has been said over and over. Truebudha has done it on several occasions, he is about to do it again. All he is concerned about is matching colors.

 

View PostSoTrue, on Mar 18 2018 - 22:39, said:

Um, yes.. even when the underdog wins it can still be a 'crap' game.  I've done it myself on occasion.  My team is horrible, but I still squeak out a win.  It's almost entirely luck though. 

Underdog/balanced stomps in under 4 minutes don't fit your categories so they are irrelevant?

 

View PostSoTrue, on Mar 18 2018 - 23:30, said:

But look, just because this happens once in a blue moon, that DOES NOT NEGATE the vast majority of the times where the more skilled team rolls over the less skilled team. 

Stronger teams don't "roll over" weaker teams the vast majority of time. You can't even produce a significant blowout correlation. The only thing you can say definitively is the better teams should win more often.



EmperorJuliusCaesar #1944 Posted Mar 19 2018 - 09:20

    Major

  • Players
  • 37052 battles
  • 5,741
  • [EOR] EOR
  • Member since:
    03-16-2014

View PostNixeldon, on Mar 18 2018 - 23:42, said:

That is exactly what has been said over and over. Truebudha has done it on several occasions, he is about to do it again. All he is concerned about is matching colors.

 

Underdog/balanced stomps in under 4 minutes don't fit your categories so they are irrelevant?

 

Stronger teams don't "roll over" weaker teams the vast majority of time. You can't even produce a significant blowout correlation. The only thing you can say definitively is the better teams should win more often.

 

Keep your head buried firmly buried in the sand if you so choose.  The rest of us see the truth, and see that it is the biggest problem of this game.

o7o7o7o7o7o7o7o7o7o7o7o7 #1945 Posted Mar 19 2018 - 09:24

    First lieutenant

  • -Players-
  • 424 battles
  • 865
  • [ITDED] ITDED
  • Member since:
    01-23-2015

98 pages of I WANT A FREE WIN BUTTON later... yall are not closer to getting it.

 

SERB laughs at you. praise serb. 



Nixeldon #1946 Posted Mar 19 2018 - 10:01

    Major

  • Players
  • 60879 battles
  • 2,218
  • Member since:
    10-30-2011

View PostEmperorJuliusCaesar, on Mar 19 2018 - 03:20, said:

Keep your head buried firmly buried in the sand if you so choose.  The rest of us see the truth, and see that it is the biggest problem of this game.

I realize your feels are more important than dealing with facts. Your idea of truth is a religious truth. You have zero evidence but faith. 



SoTrue #1947 Posted Mar 19 2018 - 10:04

    Major

  • Players
  • 33309 battles
  • 3,302
  • [ATTOP] ATTOP
  • Member since:
    04-01-2011

View PostNixeldon, on Mar 18 2018 - 23:42, said:

That is exactly what has been said over and over. Truebudha has done it on several occasions, he is about to do it again. All he is concerned about is matching colors.

 

Underdog/balanced stomps in under 4 minutes don't fit your categories so they are irrelevant?

 

Stronger teams don't "roll over" weaker teams the vast majority of time. You can't even produce a significant blowout correlation. The only thing you can say definitively is the better teams should win more often.

 

I Showed data for over 2000 battles.  It clearly showed more skilled teams won more often.

Nixeldon #1948 Posted Mar 19 2018 - 10:07

    Major

  • Players
  • 60879 battles
  • 2,218
  • Member since:
    10-30-2011

View PostSoTrue, on Mar 19 2018 - 04:04, said:

I Showed data for over 2000 battles.  It clearly showed more skilled teams won more often.

 

No one is arguing any different.

NeatoMan #1949 Posted Mar 19 2018 - 12:37

    Major

  • Players
  • 28180 battles
  • 20,556
  • Member since:
    06-28-2011

View PostSoTrue, on Mar 19 2018 - 04:04, said:

I Showed data for over 2000 battles.  It clearly showed more skilled teams won more often.

So True:  ""It's not about blowouts, it's about winning or losing"

Emperor:  "It's not about winning or losing, it's about blowouts" 

 

 It's like Abbot and Costello here.

 

Block Quote

 Sbmm is not 'balancing for the sake of balancing', it's about creating a level playing field at the start of every battle

but you just said it is in the same sentence.

 

Block Quote

I am a competitive recreational cyclist.  Not a very good one.  But there is a category I am competitive at. 

 league play is not the same as balance.  It doesn't even come close to being a similar substitute.

 

Block Quote

One team gets an early advantage and it snowballs into a blowout

 And yet somehow these more highly skilled teams that get the early advantages always take twice as long to finish the job....:sceptic:   Yeah right.

 

It's like the tomato claiming "I frequently kill 2-3 tanks every game"  not understanding that a 0.5 kpg average means he obviously doesn't.  Same here.   Averages really do tell a story.

 

If these stacked teams are reaching the snowball point earlier, then their games on average will be over sooner.  They have to contend with bigger maps and leftover good players just as much as regular teams do, so if they get the ball rolling earlier, then their games on average should also end that much earlier.

 

But when I look at the data this doesn't happen.  Their blowouts last just as long as any other blowout, regardless of team balance.  They don't become faster until you get to the extremely unbalanced games (which occur nowhere near 40% of the time).



_Tsavo_ #1950 Posted Mar 19 2018 - 13:20

    Major

  • Beta Testers
  • 46222 battles
  • 19,713
  • [BRVE] BRVE
  • Member since:
    02-16-2011

View PostSoTrue, on Mar 19 2018 - 04:04, said:

 

I Showed data for over 2000 battles.  It clearly showed more skilled teams won more often.

 

Such as they should.

SkunkButt #1951 Posted Mar 19 2018 - 13:29

    Captain

  • Players
  • 191 battles
  • 1,459
  • Member since:
    05-27-2011

SBMM would only work one way... and that's not to separate the players into 'camps' of skill, but rather ensure that both teams had the same core stat (call it 'personal rating', or even adopt XVM into the core game).

Take the same 30 players we get now and spread them across both teams, rather than the way it does now - grab the top 15 skilled players and put them in the 'right' column, and toss the other 15 in the 'left' column.  If you are an 'average' player (sub 52% overall WR) then you typically end up on the left, while the 52% and higher too often end up on the right.  The handicap is built in so that better players can get more enjoyment out of the game, and so keep playing (and putting in $$$ to spam their premium ammo).


Edited by SkunkButt, Mar 19 2018 - 13:35.


CavScout19D #1952 Posted Mar 19 2018 - 13:41

    Major

  • Players
  • 24434 battles
  • 4,773
  • Member since:
    04-24-2011

View PostSoTrue, on Mar 19 2018 - 04:04, said:

View PostNixeldon, on Mar 18 2018 - 23:42, said:

That is exactly what has been said over and over. Truebudha has done it on several occasions, he is about to do it again. All he is concerned about is matching colors.

 

Underdog/balanced stomps in under 4 minutes don't fit your categories so they are irrelevant?

 

Stronger teams don't "roll over" weaker teams the vast majority of time. You can't even produce a significant blowout correlation. The only thing you can say definitively is the better teams should win more often.

 

I Showed data for over 2000 battles.  It clearly showed more skilled teams won more often.

 

Read much?



SoTrue #1953 Posted Mar 19 2018 - 22:50

    Major

  • Players
  • 33309 battles
  • 3,302
  • [ATTOP] ATTOP
  • Member since:
    04-01-2011

View PostNeatoMan, on Mar 19 2018 - 03:37, said:

So True:  ""It's not about blowouts, it's about winning or losing"

Emperor:  "It's not about winning or losing, it's about blowouts" 

 

 It's like Abbot and Costello here.

 

but you just said it is in the same sentence.

 

 league play is not the same as balance.  It doesn't even come close to being a similar substitute.

 

 And yet somehow these more highly skilled teams that get the early advantages always take twice as long to finish the job....:sceptic:   Yeah right.

 

It's like the tomato claiming "I frequently kill 2-3 tanks every game"  not understanding that a 0.5 kpg average means he obviously doesn't.  Same here.   Averages really do tell a story.

 

If these stacked teams are reaching the snowball point earlier, then their games on average will be over sooner.  They have to contend with bigger maps and leftover good players just as much as regular teams do, so if they get the ball rolling earlier, then their games on average should also end that much earlier.

 

But when I look at the data this doesn't happen.  Their blowouts last just as long as any other blowout, regardless of team balance.  They don't become faster until you get to the extremely unbalanced games (which occur nowhere near 40% of the time).

 

a) We don't have the player base for bracketed or leagued play.  So sbmm is the best we can do.  And it is exactly relevant.  My cycling balances each field with similar skilled riders.  sbmm would balance the skill between teams.  Level playing field.

b) You can't make the claim that data show a difference in length of blowout play.  You have not tracked ALL the actual blowouts.  As you don't track when a team actually loses.  As shown earlier, it's quite common the game to be 'over' long before the last tank is dead.



Hurk #1954 Posted Mar 19 2018 - 23:16

    Major

  • Players
  • 55852 battles
  • 17,382
  • [KGR] KGR
  • Member since:
    09-30-2012

View PostSoTrue, on Mar 19 2018 - 14:50, said:

 

a) We don't have the player base for bracketed or leagued play.  So sbmm is the best we can do.  And it is exactly relevant.  My cycling balances each field with similar skilled riders.  sbmm would balance the skill between teams.  Level playing field.

b) You can't make the claim that data show a difference in length of blowout play.  You have not tracked ALL the actual blowouts.  As you don't track when a team actually loses.  As shown earlier, it's quite common the game to be 'over' long before the last tank is dead.

you want each battle to be level, which stagnates gameplay and compresses win rate without adding anything. 

the statistical analysis of blowouts already posted on the forums and early in this thread several times show that lopsided teams matter less than game design to blowouts. 

 

http://forum.worldof...erical-results/

blowouts on balanced teams are 22% likely

blowouts on massively unbalanced teams (83/17%) are 34% likely. 

so unbalanced teams only add 12% to the chance of a blowout (a little more than 50% for massively out of balance teams)

 

what you are all missing is that getting rid of the small chance of getting a massively lopsided game, with the intent of avoiding the 12% risk of a blowout is a minimal situation that you would have everyone queue for longer and not get to play as much to solve.

 

I disagree with your goal. I do not believe that getting rid of those rare games is worth messing with the games as a whole.  the current veriety of matches from "desprate struggle" down to "easymode have pie and watch" is a large part of what makes WoT better than all those other MOBAs that constantly force you into specific matchups. 

 

not to mention queue times, that several people are living a pipe dream about. queue times in other very sucessful MOBAs are several MINUTES long, even during prime time. Wots matching is extremely fast. 



Nixeldon #1955 Posted Mar 19 2018 - 23:22

    Major

  • Players
  • 60879 battles
  • 2,218
  • Member since:
    10-30-2011

View PostSoTrue, on Mar 19 2018 - 16:50, said:

 

a) We don't have the player base for bracketed or leagued play.  So sbmm is the best we can do.  And it is exactly relevant.  My cycling balances each field with similar skilled riders.  sbmm would balance the skill between teams.  Level playing field.

b) You can't make the claim that data show a difference in length of blowout play.  You have not tracked ALL the actual blowouts.  As you don't track when a team actually loses.  As shown earlier, it's quite common the game to be 'over' long before the last tank is dead.

a) Poor comparison, as usual

b) We can track match length.

When does a team actually lose?

What if I post sub 4 minute underdog/ balanced blowouts? Do those count?

Why doesn't Kliphie's blowout post count? 

What if my team is down several tanks and a more than double HP deficit and wins?

 


Edited by Nixeldon, Mar 19 2018 - 23:23.


NeatoMan #1956 Posted Mar 20 2018 - 00:57

    Major

  • Players
  • 28180 battles
  • 20,556
  • Member since:
    06-28-2011

View PostSoTrue, on Mar 19 2018 - 16:50, said:

a) We don't have the player base for bracketed or leagued play.  So sbmm is the best we can do.  And it is exactly relevant.  My cycling balances each field with similar skilled riders.  sbmm would balance the skill between teams.  Level playing field.

"But there is a category I am competitive at.  So I race in it.  Can I enter the expert or pro races?  Actually, at the local level, yes I can.  Do I?  Of course not."

You just described league play, even in your cycling example.    No, SBMM isn't the best we can do.  It's a crappy substitute.

 

Block Quote

b) You can't make the claim that data show a difference in length of blowout play.  You have not tracked ALL the actual blowouts.  As you don't track when a team actually loses.  As shown earlier, it's quite common the game to be 'over' long before the last tank is dead.

 This is the biggest bunch of BS ever.   I have to resort to baby steps again.

 

this is what you are saying happens in my data during every blowout:

 

                                      balanced team         stacked team         

blowout initiated                 4 min                     2 min                      

mop up phase                    1 min                    3 min

total time                            5 min                     5 min

 

How is it that a team with SO MUCH more skill takes twice as long to finish off an obviously weaker team?   How is it that a bunch of normal players suddenly start doing things that stacked teams are incapable of doing?  You honestly expect us to believe that on the same maps and with worse players on the other side, that a team stacked with all the good players suddenly goes potato every single time they get a steamroll going, and fails to accomplish what normal WoT players do every time?    That's what it would take in order to make those two averages equal to each other.

 

Face it....  a blowout is a blowout is a blowout.

 



Nixeldon #1957 Posted Mar 20 2018 - 01:03

    Major

  • Players
  • 60879 battles
  • 2,218
  • Member since:
    10-30-2011

View PostNeatoMan, on Mar 19 2018 - 18:57, said:

Face it....  a blowout is a blowout is a blowout.

 

But XVM says....

StiffWind #1958 Posted Mar 20 2018 - 03:17

    Major

  • -Players-
  • 18071 battles
  • 2,076
  • [MOV] MOV
  • Member since:
    03-15-2017

View Post_Gungrave_, on Mar 18 2018 - 13:54, said:

I'm wondering if I should start calling you guys necromancers because you keep beating this dead horse and the reviving it just so you can beat it to death again.

 

In that vein, you must be fascinated by the occult....you can't seem the keep from popping in and out of the thread to whine about it.

 



scyorkie #1959 Posted Mar 20 2018 - 13:05

    Major

  • Players
  • 74349 battles
  • 3,847
  • Member since:
    10-26-2011

View PostStiffWind, on Mar 20 2018 - 10:17, said:

In that vein, you must be fascinated by the occult....you can't seem the keep from popping in and out of the thread to whine about it.

 

Neither can you. At least he doesn't post drivel.



FrankFitz #1960 Posted Sep 20 2019 - 16:56

    Corporal

  • -Players-
  • 21212 battles
  • 50
  • [PUBBY] PUBBY
  • Member since:
    12-06-2016
TBH some situations in life are one sided. Just bend over and deal with it already!!




3 user(s) are reading this topic

0 members, 3 guests, 0 anonymous users