Jump to content


A Statistical Demonstration of why Statistics are Unreliable


  • Please log in to reply
11 replies to this topic

b0d #1 Posted 04 March 2016 - 03:11 PM

    Colonel

  • Player
  • 19589 battles
  • 3,909
  • Member since:
    05-22-2015

Hi

Here is a veteran players WN8; 1144

Here is the same players current WN8 - 870 on a new account after a short break and geographical relocation of thousands of miles.

That first value is offset by the majority of games played in a scout tank with a much lower Wn8 value than all other classes.  Most other values, HT and TD's (TD's were played more seriously after 8-10K games and that WN8 number reflects that increase in experience).  This "play style" in scout tanks that is responsible for the low WN8 in that class brings down the average value that does not reflect the players experience.  In other words the average WN8 is a weak statistic and skewed by playstyle of one class of tank.

The player is now new and improved on the new account, a "better" player overall, without a shadow of doubt - reflected by increased % to win compared to old account.

So why the lower WN8?

Because all statistics are subject to bias.  Of course the person relying on the statistic as an argument will rarely tell you this.



Joggaman #2 Posted 04 March 2016 - 03:33 PM

    Major General

  • Player
  • 22737 battles
  • 5,926
  • [WJDE] WJDE
  • Member since:
    05-23-2011

WN8 basically measures damage. 

Shoot and do 2-3 damaging shots with your scout each battle, and WN8 will rise drastically. 

I think my WN8 record in a battle is 29k with type64. 


Edited by Joggaman, 04 March 2016 - 03:35 PM.


Eaglax #3 Posted 04 March 2016 - 03:36 PM

    Lieutenant Сolonel

  • Player
  • 18973 battles
  • 3,481
  • Member since:
    01-12-2012

hm..normal people would have a higher WN8 on new accounts, since they don't make beginner mistakes anymore...

funny that "arguments" like yours only come from people with bad WN8:popcorn:

 

 



b0d #4 Posted 04 March 2016 - 04:12 PM

    Colonel

  • Player
  • 19589 battles
  • 3,909
  • Member since:
    05-22-2015

View PostDocPurple, on 04 March 2016 - 03:36 PM, said:

hm..normal people would have a higher WN8 on new accounts, since they don't make beginner mistakes anymore...

funny that "arguments" like yours only come from people with bad WN8:popcorn:

 

 

 

yes strange isnt it?

Maybe its because your mind is narrow.  It has to do with the amount of time I spent at tier 1 on this account.  Hence the wn8 is only partially related to effectiveness and contains a significant amount of bias.

The source of this bias is financial, nothing to do with effectiveness (I do agree stats indicate a measure of effectiveness in this context).  The stats are a tool for the developer to encourage people to spend money (spend money = wn8 will increase).  There are probably more sources of bias other than this.  Fund me and I will prove it to you.

Also have not considered what I mentioned about the data from the other account, and just regurgitate other peoples opinion.

All these shitty arguments about stats and their claimed usefulness never ever address the weaknesses of stats,.

There is no such thing as luck.


Edited by b0d, 04 March 2016 - 04:20 PM.


b0d #5 Posted 04 March 2016 - 04:25 PM

    Colonel

  • Player
  • 19589 battles
  • 3,909
  • Member since:
    05-22-2015

View PostJoggaman, on 04 March 2016 - 03:33 PM, said:

WN8 basically measures damage. 

Shoot and do 2-3 damaging shots with your scout each battle, and WN8 will rise drastically. 

I think my WN8 record in a battle is 29k with type64. 

 

So WN8 is a most unreliable indicator of someones performance if performance is not exclusively related to damage.  Funny how people go on that it is.

Joggaman #6 Posted 04 March 2016 - 04:40 PM

    Major General

  • Player
  • 22737 battles
  • 5,926
  • [WJDE] WJDE
  • Member since:
    05-23-2011

View Postb0d, on 04 March 2016 - 04:25 PM, said:

 

So WN8 is a most unreliable indicator of someones performance if performance is not exclusively related to damage.

Yes. 

Old discussion. 



Joggaman #7 Posted 04 March 2016 - 04:45 PM

    Major General

  • Player
  • 22737 battles
  • 5,926
  • [WJDE] WJDE
  • Member since:
    05-23-2011

That said, I have yet to see a 60%+wr player with low WN8...

You draw the conclusion. 



b0d #8 Posted 04 March 2016 - 04:59 PM

    Colonel

  • Player
  • 19589 battles
  • 3,909
  • Member since:
    05-22-2015

Oh sorry.  Why do people beat on about it then.  Thanks for helping me learn to ignore them!

I must not take it personally.

I must not take it personally.

I must not take it personally.

I must not take it personally.

I must not take it personally.

I must not take it personally.

I must not take it personally.

I must not take it personally.

I must not take it personally.

I must not take it personally.



Eaglax #9 Posted 04 March 2016 - 05:07 PM

    Lieutenant Сolonel

  • Player
  • 18973 battles
  • 3,481
  • Member since:
    01-12-2012

View Postb0d, on 04 March 2016 - 04:12 PM, said:

 

yes strange isnt it?

Maybe its because your mind is narrow.  It has to do with the amount of time I spent at tier 1 on this account.  Hence the wn8 is only partially related to effectiveness and contains a significant amount of bias.

The source of this bias is financial, nothing to do with effectiveness (I do agree stats indicate a measure of effectiveness in this context).  The stats are a tool for the developer to encourage people to spend money (spend money = wn8 will increase).  There are probably more sources of bias other than this.  Fund me and I will prove it to you.

Also have not considered what I mentioned about the data from the other account, and just regurgitate other peoples opinion.

All these shitty arguments about stats and their claimed usefulness never ever address the weaknesses of stats,.

There is no such thing as luck.

 

lol....WN8 is not an official WG rating :facepalm:, and can't be influenced directly by WG...

the official WG-rating is crap and no one gives a [edited] about it, your "financial" argument is just plain BS. Edit: if WG would care about "skill-ratings", they would implement a skill-matchmaking....

 

ofc WN8 and every other rating has it's downsides (no spotting damage ect.), and it will never be 100% accurate, but to say "WN8 is a most unreliable indicator of someones performance" is, again, complete BS. And what a coincidence, that players with high WN8 have high winrates.....and, also again, it's not the "people" who "beat on about it", it's noobs like you, who don't like to be the ones with bad ratings, and instead of trying to get better, they go the easy way and try to find "arguments" to make those ratings look obsolete, and just like all your low WN8-predecessors, you fail too:medal:


Edited by DocPurple, 04 March 2016 - 05:08 PM.


b0d #10 Posted 04 March 2016 - 06:04 PM

    Colonel

  • Player
  • 19589 battles
  • 3,909
  • Member since:
    05-22-2015

View PostDocPurple, on 04 March 2016 - 05:07 PM, said:

 

lol....WN8 is not an official WG rating :facepalm:, and can't be influenced directly by WG...

the official WG-rating is crap and no one gives a [edited] about it, your "financial" argument is just plain BS. Edit: if WG would care about "skill-ratings", they would implement a skill-matchmaking....

 

ofc WN8 and every other rating has it's downsides (no spotting damage ect.), and it will never be 100% accurate, but to say "WN8 is a most unreliable indicator of someones performance" is, again, complete BS. And what a coincidence, that players with high WN8 have high winrates.....and, also again, it's not the "people" who "beat on about it", it's noobs like you, who don't like to be the ones with bad ratings, and instead of trying to get better, they go the easy way and try to find "arguments" to make those ratings look obsolete, and just like all your low WN8-predecessors, you fail too:medal:

 

Well thanks for pointing that out I am grateful.

You still have not talked about the point I made in my original post.  Please try and do so.

In my case WN8 has proven itself to be a "most unreliable indicator".  I know many "high WN8" players who play worse than me, farming damage in isolation of a victory.



Eaglax #11 Posted 04 March 2016 - 06:23 PM

    Lieutenant Сolonel

  • Player
  • 18973 battles
  • 3,481
  • Member since:
    01-12-2012

View Postb0d, on 04 March 2016 - 06:04 PM, said:

 

Well thanks for pointing that out I am grateful.

You still have not talked about the point I made in my original post.  Please try and do so.

In my case WN8 has proven itself to be a "most unreliable indicator".  I know many "high WN8" players who play worse than me, farming damage in isolation of a victory.

 

I can't do that genius, if you don't tell me your 1144 account name....

pls post some names of people with far better WN8, and you think that they play worse than you.:popcorn: I bet you find a reason to not post them...

 

and I said already, your "point" from the first post is just a attempt to say your stats are not that bad...and sorry, they are, even in non-scouts tanks, such as the KV1 you have bad stats (842WN8, same as your account WN8...but sure blame it on the non-rewarding-scout battles................). Saying that scouts ruin WN8 is complete BS, also because it's "easy" to farm WN8 with them...All your "arguments" are just a try-hard-excuse for your bad stats....just accept that you are just on the edge to average and live with that....or learn more about the game and get better, but don't post the usual WN8-is-nothing-worth



caucainis #12 Posted 12 March 2016 - 10:40 AM

    Private

  • Player
  • 14599 battles
  • 3
  • Member since:
    01-03-2013
 test http://wotlabs.net/sig_dark/eu/pomidoras5454/signature.png




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users