Jump to content


Why skilled based mm will ruin world of tanks

Skill based mm Mm Skill Will rate Rigged Wg

  • Please log in to reply
59 replies to this topic

NeatoMan #41 Posted Mar 05 2018 - 16:30

    Major

  • Players
  • 25597 battles
  • 17,391
  • Member since:
    06-28-2011

View Postspud_tuber, on Mar 05 2018 - 09:58, said:

Are you implying that current MM isn't skill random?  If so, I've seen plenty of evidence to suggest it is(baring of course the effects of the recording player's and any toon mates s/he has skills).  NeatoMan has the largest data set I've seen recently, but even SoTrue's data strongly suggests skill random. Both of which I'm pretty sure are in that thread you linked. Mesnwhile, no one has ever shown a statistically relevant sample size of data that suggests the game is rigged.

It's not only mine, but it's every single person who has ever kept track of team balance.  They all get the same normal distribution pattern that is typical for random systems; even some of the biggest tinfoil riggers.  No matter what rating system you choose to measure team balance, it will generate the same bell shaped curve.

 

Nobody is saying MM cannot be manipulated.  We're simply saying it's not manipulated the way you think it is.  The expected results of suggested manipulation does not line up with the data or the patent.

 

I can think of many possible ways to manipulate win rates, it doesn't mean they are using them, nor would I have any proof that they are.   e.g. they could manipulate your MM by adding or removing from your tank's MM weight based on your win rate.  This would be a subtle nudging effect over many games rather than the supposedly "obvious" session by session manipulation riggers complain about.   However, I got zero evidence to back it up, and nothing in the patent supports it either, therefore I do not come here spouting it as fact.  It's just a thought, nothing more.  It doesn't make it real.  You'd need some evidence to make that jump.

 

also it looks like krupp found a new term, "logical fallacies".  he keeps using it but has no idea what they are.  He can't even describe which ones we are using and why they are fallacies.


Edited by NeatoMan, Mar 05 2018 - 16:35.


_Gungrave_ #42 Posted Mar 05 2018 - 16:31

    Major

  • Players
  • 41624 battles
  • 15,468
  • [X-OUT] X-OUT
  • Member since:
    12-07-2011

View Postkruppw, on Mar 05 2018 - 15:21, said:

 

​Interesting how the last major thread had pages of math that clearly proved you wrong over and over again. All I see from you is a lot of claims and zero supporting evidence. This is further reiterated by the plethora of logical fallacies.

So far we have only seen the same arguments spammed, proof by assertion and ad nauseam. Each one gets shot down/earlier proof where it was shot down is cited.
​"MM cannot be manipulated"
​"You don't know what SBMM is"
" SBMM would increase que times"
​"SBMM would lower the stats of good players"
​"Current MM is random"

Did I miss any?
Source: http://forum.worldof...n-matchmaking/​

 

I don't bother with statistical analysis like others do because I speak from experience of playing other games that have used skill MM with lackluster or failing results.

 

As for you well you might want to practice what you preach then because I've yet to see you counter any of the arguments with any sort of evidence.



TerrorJoe #43 Posted Mar 05 2018 - 16:36

    Corporal

  • Players
  • 20094 battles
  • 38
  • Member since:
    04-23-2011

View Postspud_tuber, on Mar 05 2018 - 14:58, said:

Are you implying that current MM isn't skill random?  If so, I've seen plenty of evidence to suggest it is(baring of course the effects of the recording player's and any toon mates s/he has skills).  NeatoMan has the largest data set I've seen recently, but even SoTrue's data strongly suggests skill random. Both of which I'm pretty sure are in that thread you linked. Mesnwhile, no one has ever shown a statistically relevant sample size of data that suggests the game is rigged.

 

Further, are you implying that skill balanced MM would not negatively effect the win rate at least of any player better than average?  HiBan's simulation available in the thread you linked shows that skill balanced MM would have the equivalent effect on a SU of swapping one average player for one super bad player on his team and one average player for one very good/borderline unicum player on the opposing team in every single battle that super unicum plays.  Don't you think that will drag his/her WR down?

 

​NeatoMans data set was debunked and called out with evidence of outliers and bias multiple times in that very thread. SoTrue's data does not suggest random at all, actually the contrary. The term "Rigged" is actually highly subjective. Do you really want to go there?

​Implying? It was proven in that very thread.

​1.) If the current MM was Random, given the current amount of sample size available (IE: the game is not new and the MM hasn't really changed recently), the average WR would be hovering very close to 50%. It does not.

​2.)  You are clearly still stuck on the baseline of everyone stats from the current MM, which is already proven as NOT entirely random.  ****IF***** a player is actually a "Good" player, or "Unicum", or whatever label you which to use to represent a certain stat bracket, then in a "fair" matchmaking system (see SB/Swap) they would NOT see much deviation from their current stats (and label).   Ergo, if John Doe is actually a skilled player, forcing them into fair matches does not reduce said skill. If this Unicum John Doe is forced to have more bad players on his team than before, he will simply have to carry in a fashion that should be relevant to said label and stats. Because a "fair" matchmaking system like SB/SWAP aggregates equally on both sides of the spectrum.... just as said John Doe would have to carry more than before in some matches, there would be the same equivalent requirements on the opposing team and its member(s).
​If Uncium John Doe's WR gets "dragged" down, this is nothing more than proof said WR was clearly a misrepresentation of what said player can actually achieve in a fair, unmanipulated environment.

Edited by kruppw, Mar 05 2018 - 16:39.


TerrorJoe #44 Posted Mar 05 2018 - 16:37

    Corporal

  • Players
  • 20094 battles
  • 38
  • Member since:
    04-23-2011

View Post_Gungrave_, on Mar 05 2018 - 15:31, said:

 

I don't bother with statistical analysis like others do because I speak from experience of playing other games that have used skill MM with lackluster or failing results.

 

As for you well you might want to practice what you preach then because I've yet to see you counter any of the arguments with any sort of evidence.


Translated:

​"I don't deal with evidence, I just spout logical fallacies. The burden of proof is never on me. My mouth and hands are fact generators"


Edited by kruppw, Mar 05 2018 - 16:37.


NeatoMan #45 Posted Mar 05 2018 - 16:47

    Major

  • Players
  • 25597 battles
  • 17,391
  • Member since:
    06-28-2011

View Postkruppw, on Mar 05 2018 - 10:36, said:

​NeatoMans data set was debunked and called out with evidence of outliers and bias multiple times in that very thread. SoTrue's data does not suggest random at all, actually the contrary. The term "Rigged" is actually highly subjective. Do you really want to go there?

oooh.  look he learned another new word... "debunked".  

 

How can it be debunked when every single person who has ever kept track gets the same normal distribution pattern?  That's called confirmation.  Learn your science.



da_Rock002 #46 Posted Mar 05 2018 - 16:56

    Major

  • -Players-
  • 7097 battles
  • 2,801
  • Member since:
    11-24-2016

Skilled matchups haven't ruined it yet.   Well, only 40% of it, that is, and those are skill-UNMATCHED-up.

 

Yeah, the skill-unmatched battles account for around 40% of the battles, leaving 60% of the rest as "filter-matched-up".     When you filter out the impurities in water, you have clean water, don't you.


 

You've been playing WoT for quite some time with 60% of your battles SBMM'ed.     WG really didn't mean for that to happen, but it does every day. 


 

If you'd like to know what the game would be like with SBMM cleaning up 100% of the battles, just think back on all those somewhat longish, somewhat even-outcome battles were/are like.


 

SBMM ain't half bad, is it.



spud_tuber #47 Posted Mar 05 2018 - 17:03

    Major

  • Players
  • 49994 battles
  • 6,060
  • Member since:
    08-26-2013

View Postda_Rock002, on Mar 05 2018 - 09:56, said:

Skilled matchups haven't ruined it yet.   Well, only 40% of it, that is, and those are skill-UNMATCHED-up.

 

Yeah, the skill-unmatched battles account for around 40% of the battles, leaving 60% of the rest as "filter-matched-up".     When you filter out the impurities in water, you have clean water, don't you.


 

You've been playing WoT for quite some time with 60% of your battles SBMM'ed.     WG really didn't mean for that to happen, but it does every day. 


 

If you'd like to know what the game would be like with SBMM cleaning up 100% of the battles, just think back on all those somewhat longish, somewhat even-outcome battles were/are like.


 

SBMM ain't half bad, is it.

I've had longish, somewhat even outcome battles in what you would consider unbalanced skill wise matches.  I've also had fast blowouts in matches you would consider very skill balanced.  Everyone has.  Just look at NeatoMan's data for a more detailed view of just how little effect removing significantly unbalanced matches would have on the blowout rate.(hint, not a lot)



NeatoMan #48 Posted Mar 05 2018 - 17:11

    Major

  • Players
  • 25597 battles
  • 17,391
  • Member since:
    06-28-2011

View Postspud_tuber, on Mar 05 2018 - 11:03, said:

I've had longish, somewhat even outcome battles in what you would consider unbalanced skill wise matches.  I've also had fast blowouts in matches you would consider very skill balanced.  Everyone has.  Just look at NeatoMan's data for a more detailed view of just how little effect removing significantly unbalanced matches would have on the blowout rate.(hint, not a lot)

referencing my data won't help.  they think it was spawned by the antichrist.



Fractured_Raptor #49 Posted Mar 05 2018 - 17:14

    First lieutenant

  • -Players-
  • 20087 battles
  • 999
  • [-RISK] -RISK
  • Member since:
    05-28-2016

Why skilled based mm will ruin world of tanks for stat padders

 

FTFY. Free of charge.



SoTrue #50 Posted Mar 05 2018 - 17:16

    Major

  • Players
  • 28171 battles
  • 3,302
  • Member since:
    04-01-2011

View Postkruppw, on Mar 05 2018 - 07:36, said:

 

​NeatoMans data set was debunked and called out with evidence of outliers and bias multiple times in that very thread. SoTrue's data does not suggest random at all, actually the contrary. The term "Rigged" is actually highly subjective. Do you really want to go there?

​Implying? It was proven in that very thread.

​1.) If the current MM was Random, given the current amount of sample size available (IE: the game is not new and the MM hasn't really changed recently), the average WR would be hovering very close to 50%. It does not.

​2.)  You are clearly still stuck on the baseline of everyone stats from the current MM, which is already proven as NOT entirely random.  ****IF***** a player is actually a "Good" player, or "Unicum", or whatever label you which to use to represent a certain stat bracket, then in a "fair" matchmaking system (see SB/Swap) they would NOT see much deviation from their current stats (and label).   Ergo, if John Doe is actually a skilled player, forcing them into fair matches does not reduce said skill. If this Unicum John Doe is forced to have more bad players on his team than before, he will simply have to carry in a fashion that should be relevant to said label and stats. Because a "fair" matchmaking system like SB/SWAP aggregates equally on both sides of the spectrum.... just as said John Doe would have to carry more than before in some matches, there would be the same equivalent requirements on the opposing team and its member(s).
​If Uncium John Doe's WR gets "dragged" down, this is nothing more than proof said WR was clearly a misrepresentation of what said player can actually achieve in a fair, unmanipulated environment.

 

Yeah, your point in #2 is always overlooked by the random trolls.  They all whine that sbmm would add more bad players to a good players team.  In comparison to random, sbmm would only 'add' bad players HALF the time.  The other HALF of the time good players would be added to the good players team.  Don't know why they can't grasp this.

spud_tuber #51 Posted Mar 05 2018 - 17:18

    Major

  • Players
  • 49994 battles
  • 6,060
  • Member since:
    08-26-2013

View PostNeatoMan, on Mar 05 2018 - 10:11, said:

referencing my data won't help.  they think it was spawned by the antichrist.

ROFL, yeah....I'm beginning to get that impression.  Never mind that you've the best backed up data I've ever seen with screenshots and/or replays on demand.  But hopefully any as yet uncorrupted minds reading the thread will be more open to reality. 

 

Honestly, for my own sanity, and the sake of avoiding more *headdesk* and *facepalm* than I already have, I think I just need to block the worst of the crazies.



spud_tuber #52 Posted Mar 05 2018 - 17:26

    Major

  • Players
  • 49994 battles
  • 6,060
  • Member since:
    08-26-2013

View PostSoTrue, on Mar 05 2018 - 10:16, said:

 

Yeah, your point in #2 is always overlooked by the random trolls.  They all whine that sbmm would add more bad players to a good players team.  In comparison to random, sbmm would only 'add' bad players HALF the time.  The other HALF of the time good players would be added to the good players team.  Don't know why they can't grasp this.

And the skill balanced proponents ignore that the long term effect is the equivalent of swapping an average player on a Super Unicum's team for a super bad, and an average player on the opposing team for a very good/borderline unicum player in every single battle.  Nor can they say how that is fair, instead dodging the question, or claiming they don't understand the question, claiming that this effect isn't real despite all the evidence to support it.

 

So, I'll ask again, and you'll dodge again:

 

How is it fair in a skill balanced MM that in the  long term, better players get worse teammates and better opponents than they would under a skill random MM?

 

How is it fair in a skill balanced MM that in the long term, good players get worse teammates and better opponents than bad players under that same skill balanced MM?



TerrorJoe #53 Posted Mar 05 2018 - 17:30

    Corporal

  • Players
  • 20094 battles
  • 38
  • Member since:
    04-23-2011

View PostSoTrue, on Mar 05 2018 - 16:16, said:

 

Yeah, your point in #2 is always overlooked by the random trolls.  They all whine that sbmm would add more bad players to a good players team.  In comparison to random, sbmm would only 'add' bad players HALF the time.  The other HALF of the time good players would be added to the good players team.  Don't know why they can't grasp this.

 

​Of course it is. They constantly overlook anything that doesn't fit their narrative. Where do we see that everyday?

​I mean we have people in this very thread that clearly have zero understanding of even basic statistics, acting like a skewed data set with every outlier possible not to mention biased conditional probabilities in the extrapolation is some sort of bible (EX: Neatoman's data). Just a sign of the blind leading the blind.
 

TerrorJoe #54 Posted Mar 05 2018 - 17:32

    Corporal

  • Players
  • 20094 battles
  • 38
  • Member since:
    04-23-2011

View Postspud_tuber, on Mar 05 2018 - 16:26, said:

And the skill balanced proponents ignore that the long term effect is the equivalent of swapping an average player on a Super Unicum's team for a super bad, and an average player on the opposing team for a very good/borderline unicum player in every single battle.  Nor can they say how that is fair, instead dodging the question, or claiming they don't understand the question, claiming that this effect isn't real despite all the evidence to support it.

 

So, I'll ask again, and you'll dodge again:

 

How is it fair in a skill balanced MM that in the  long term, better players get worse teammates and better opponents than they would under a skill random MM?

 

How is it fair in a skill balanced MM that in the long term, good players get worse teammates and better opponents than bad players under that same skill balanced MM?


​Ignored?

​All of what you just regurgitated was directly refuted with my #2 that he was quoting. But you know..... that proof by assertion, ad nauseam...... etc.    How are those logical fallacies working out for you?


​Furthermore, for the record, I would like to just give a helpful hint out there who are currently the cause for great amusement in said discussion.

​If you are actually going to argue that the current MM is not rigged, or rather that it is "random" on the same hand you are arguing that a SB/SWAP MM will make good players stats essentially decrease it is the equivalent of shooting yourself in the proverbial foot as they say.

​If the current MM was random, then said good player shouldn't be seeing such a proportion of matches where they have an unequal amount of similar "good" players on their team compared to the opposing team. Therefore SB/SWAP MM wouldn't be changing the amount of matches said player receives "bad" players on their team nor the amount of bad players compared to the current iteration.

Edited by kruppw, Mar 05 2018 - 17:46.


NeatoMan #55 Posted Mar 05 2018 - 17:44

    Major

  • Players
  • 25597 battles
  • 17,391
  • Member since:
    06-28-2011

View PostSoTrue, on Mar 05 2018 - 11:16, said:

Yeah, your point in #2 is always overlooked by the random trolls.  They all whine that sbmm would add more bad players to a good players team.  In comparison to random, sbmm would only 'add' bad players HALF the time.  The other HALF of the time good players would be added to the good players team.  Don't know why they can't grasp this.

We can't grasp it because it is completely wrong.  You've seen the win chance shift.  That means a good player gets far more games where the good player is shifted off his team and a bad player is shifted onto his team than the other way around.    You still haven't learned a thing about win chance, and with that statement it appears you've actually regressed in your understanding.



spud_tuber #56 Posted Mar 05 2018 - 17:46

    Major

  • Players
  • 49994 battles
  • 6,060
  • Member since:
    08-26-2013

View Postkruppw, on Mar 05 2018 - 10:32, said:


​Ignored?

​All of what you just regurgitated was directly refuted with my #2 that he was quoting. But you know..... that proof by assertion, ad nauseam...... etc.    How are those logical fallacies working out for you?

Your #2 assumes facts not in evidence, specifically that the current MM isn't skill random. You've repeatedly failed to demonstrate anything of the sort, and  every other data set of any statistically relevant size has strongly indicated that the system is skill random.  This means your claim is the proof by assertion, ad nauseum.

 

Sinply put, your #2  can be filed under a combination of claiming they don't understand the question and claiming the effect isn't real.  

 

Thank you for proving my point once again.



TerrorJoe #57 Posted Mar 05 2018 - 17:48

    Corporal

  • Players
  • 20094 battles
  • 38
  • Member since:
    04-23-2011

View Postspud_tuber, on Mar 05 2018 - 16:46, said:

Your #2 assumes facts not in evidence, specifically that the current MM isn't skill random. You've repeatedly failed to demonstrate anything of the sort, and  every other data set of any statistically relevant size has strongly indicated that the system is skill random.  This means your claim is the proof by assertion, ad nauseum.

 

Sinply put, your #2  can be filed under a combination of claiming they don't understand the question and claiming the effect isn't real.  

 

Thank you for proving my point once again.

 

​Oh you mean like the evidence I cited in this very thread? Where is yours?

​By the way, thank YOU for proving my point:
 

Block Quote

Furthermore, for the record, I would like to just give a helpful hint out there who are currently the cause for great amusement in said discussion.

​If you are actually going to argue that the current MM is not rigged, or rather that it is "random" on the same hand you are arguing that a SB/SWAP MM will make good players stats essentially decrease it is the equivalent of shooting yourself in the proverbial foot as they say.

​If the current MM was random, then said good player shouldn't be seeing such a proportion of matches where they have an unequal amount of similar "good" players on their team compared to the opposing team. Therefore SB/SWAP MM wouldn't be changing the amount of matches said player receives "bad" players on their team nor the amount of bad players compared to the current iteration.     



​For everything else there is MasterCard

     

Edited by kruppw, Mar 05 2018 - 17:49.


NeatoMan #58 Posted Mar 05 2018 - 17:57

    Major

  • Players
  • 25597 battles
  • 17,391
  • Member since:
    06-28-2011

View Postkruppw, on Mar 05 2018 - 11:30, said:

​I mean we have people in this very thread that clearly have zero understanding of even basic statistics, acting like a skewed data set with every outlier possible not to mention biased conditional probabilities in the extrapolation is some sort of bible (EX: Neatoman's data). Just a sign of the blind leading the blind.
I think the forums have finally spawned the anti-Neato.  This is a totally new way of arguing;  wait for the opposition to make a point about your argument, then use that very same point except insert the opposition everywhere it applies to you...     brilliant!   You no longer need to think anymore.

spud_tuber #59 Posted Mar 05 2018 - 18:06

    Major

  • Players
  • 49994 battles
  • 6,060
  • Member since:
    08-26-2013

View PostNeatoMan, on Mar 05 2018 - 10:57, said:

I think this the forums have finally spawned the anti-Neato.  This is a totally new way of arguing;  wait for the opposition to make a point about your argument, then use that very same point except insert the opposition everywhere it applies to you...     brilliant!   You no longer need to think anymore.

That is literally laugh out loud funny.



Nixeldon #60 Posted Mar 05 2018 - 21:10

    Captain

  • Players
  • 56525 battles
  • 1,564
  • Member since:
    10-30-2011

View Postkruppw, on Mar 05 2018 - 10:36, said:

​1.) If the current MM was Random, given the current amount of sample size available (IE: the game is not new and the MM hasn't really changed recently), the average WR would be hovering very close to 50%. It does not.

 

What is the average win rate?




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users