Jump to content


Playing without xvm has been just so dissappointing, especially with sbmm... sigh...


  • Please log in to reply
57 replies to this topic

SimplyPzB2 #41 Posted May 22 2019 - 20:41

    Sergeant

  • -Players-
  • 105 battles
  • 248
  • Member since:
    05-26-2016

View Post_Tsavo_, on May 22 2019 - 20:34, said:

 

I don't like it, I've explained why I don't like it.  I'm not testing it or gathering data on it, others have done enough of that already.

 

But you play a lot, so your data would really help.  You seem to post a lot and care a lot about the game, I find it strange you're not willing to do a least a little data mining to make the game a little better..

_Tsavo_ #42 Posted May 22 2019 - 21:29

    Major

  • Beta Testers
  • 45220 battles
  • 19,261
  • [BRVE] BRVE
  • Member since:
    02-16-2011

View PostSimplyPzB2, on May 22 2019 - 14:41, said:

 

But you play a lot, so your data would really help.  You seem to post a lot and care a lot about the game, I find it strange you're not willing to do a least a little data mining to make the game a little better..

 

Mostly because I forget to.  I have replay files but not much else in terms of spare time currently.  I tried but usually end up doing it for a day or two then just forget.  

TheManFromKekistan #43 Posted May 22 2019 - 21:42

    Major

  • -Players-
  • 28 battles
  • 2,549
  • Member since:
    02-03-2017
I love xvm for the ui and mods it allows but don't use the stat part. It's just too depressing to load in and see nearly all of your team in deep deep red and the other side populated with barely any players lower than yellow. Might as well go in blind and hope for the best instead of knowing over half your team is about to die in under three minutes since it kinda kills the will to fight.. 

SquishySupreme #44 Posted May 22 2019 - 23:32

    First lieutenant

  • -Players-
  • 3116 battles
  • 568
  • Member since:
    10-29-2018
(to the OP)
It's players like you that go drown themselves or YOLO because their little app says that they can't win.

I use no mods.  I have my eyes, my gun, and my trigger finger.  Every game is different and I play as hard as I can.  The odds make no difference at all on how I play or would play.

Edited by SquishySupreme, May 22 2019 - 23:33.


SimplyPzB2 #45 Posted May 22 2019 - 23:47

    Sergeant

  • -Players-
  • 105 battles
  • 248
  • Member since:
    05-26-2016

View Post_Tsavo_, on May 22 2019 - 21:29, said:

 

Mostly because I forget to.  I have replay files but not much else in terms of spare time currently.  I tried but usually end up doing it for a day or two then just forget.  

 

Try this.  Keep a pencil and pad of paper next to your computer.  Pick 1-3 tanks you play most.  Then create a log for each tank.  Simply write down what the difference in skill is, win or loss, and what tier battle it was.  (differnce in skill means how many more/less skilled players your team has.  So if your team has 4 more skilled players you'd write +4).  Here's a couple samples:

 

-2  W 9  (My team had two less skilled players, we won, it was a tier 9 battle)

0   L  7    (Teams had same number of skilled players, we lost, it was a tier 7 battle)

+4  W  8  (My team had four more skilled players, we won, it was a tier 8 battle).

 

Note on the xvm rainbow skill metric, I consider yellow and up as 'skilled'.  Orange and below to be unskilled.

 

You'll be surprised how quick and easy this is to do.  Do it for 200-500 battles for each tank.  At the end, you enter it into excel and you have a data set.  I've found the following:

 

-1, 0, +1      =  crapshoot, it's anywones game.

-3,-2,+2,+3  = noticable disadvantage, advantage.  Not a guaranteed loss/win, but the less skilled team has their work cut out for them

-4, +4 or greater = Almost a lock for loss/win.  Very hard to overcome this disparity.  (Not 100% impossible, but highly unlikely).

 

Keep in mind, this is a pretty rough gauge of team skill.  It does not factor in top tier unicum vs bottom tier unicum, or unicum in medium vs unicum in arty, etc etc.  But, while rough, it's still very functional.

-

View PostTheManFromKekistan, on May 22 2019 - 21:42, said:

I love xvm for the ui and mods it allows but don't use the stat part. It's just too depressing to load in and see nearly all of your team in deep deep red and the other side populated with barely any players lower than yellow. Might as well go in blind and hope for the best instead of knowing over half your team is about to die in under three minutes since it kinda kills the will to fight.. 

 

I feel this pain.  I would suggest running the stats though.   I find it very usefull to know up front that I need to farm damage hard before my team gets rolled.  Or I need to farm damage hard before the enemy team evaporates.

-

View PostSquishySupreme, on May 22 2019 - 23:32, said:

(to the OP)
It's players like you that go drown themselves or YOLO because their little app says that they can't win.

I use no mods.  I have my eyes, my gun, and my trigger finger.  Every game is different and I play as hard as I can.  The odds make no difference at all on how I play or would play.

 

Um, I don't do that (xvm suicide).  That's a waste of an opportunity.  If I'm on a really bad team, I know it's a very good opportunity to farm tons of damage and xp.

-

Also, xvm stats etc, is a tool.  I very much adjust my play based on the skill of my team and the enemy team.  And I win more because of it.  It's really no differenct than using a fully equipped tank, or a high skill crew.  You don't have to use these, but it does help when you do.



_Tsavo_ #46 Posted May 23 2019 - 00:00

    Major

  • Beta Testers
  • 45220 battles
  • 19,261
  • [BRVE] BRVE
  • Member since:
    02-16-2011
that's a lot of work, I run vanilla, and I'm not going to do look up individuals.  I'm here to play a game, not continue my probability and statistics course

SimplyPzB2 #47 Posted May 23 2019 - 00:13

    Sergeant

  • -Players-
  • 105 battles
  • 248
  • Member since:
    05-26-2016

View Post_Tsavo_, on May 23 2019 - 00:00, said:

that's a lot of work, I run vanilla, and I'm not going to do look up individuals.  I'm here to play a game, not continue my probability and statistics course

 

Install xvm for the duration of your experiment.  Litterally takes 5 seconds at the start of each battle to jot down the data.   

NeatoMan #48 Posted May 23 2019 - 01:07

    Major

  • Players
  • 27986 battles
  • 19,924
  • Member since:
    06-28-2011

View PostSimplyPzB2, on May 22 2019 - 18:13, said:

Install xvm for the duration of your experiment.  Litterally takes 5 seconds at the start of each battle to jot down the data.   

pfft... manual collection of data is so 9.0.   Get with the times.  Install a stat collecting mod (such as WoTNumbers).  That way it will be in a consistent format, easily shared, and nobody can question the validity of your data.    ... but, i'm positive you never will.   They never do, precisely because of those three reasons I listed. 



_Tsavo_ #49 Posted May 23 2019 - 02:20

    Major

  • Beta Testers
  • 45220 battles
  • 19,261
  • [BRVE] BRVE
  • Member since:
    02-16-2011

View PostNeatoMan, on May 22 2019 - 19:07, said:

pfft... manual collection of data is so 9.0.   Get with the times.  Install a stat collecting mod (such as WoTNumbers).  That way it will be in a consistent format, easily shared, and nobody can question the validity of your data.    ... but, i'm positive you never will.   They never do, precisely because of those three reasons I listed. 

 

What would you suggest?  I'm not for mods, but if they pull replay files or so, or if you'd like bulk replay files, I'd contribute data.   I'm just not running any mods or doing manual gathering of data.   

RingZero #50 Posted May 23 2019 - 03:50

    Major

  • -Players-
  • 21362 battles
  • 2,447
  • [-G-] -G-
  • Member since:
    05-30-2014
Since the actual username and tank used is well known now, could the OP post the actual replay so we can create our own impression of the game flow instead of having to take your description of it?

SimplyPzB2 #51 Posted May 23 2019 - 07:54

    Sergeant

  • -Players-
  • 105 battles
  • 248
  • Member since:
    05-26-2016

View Post_Tsavo_, on May 23 2019 - 02:20, said:

 

What would you suggest?  I'm not for mods, but if they pull replay files or so, or if you'd like bulk replay files, I'd contribute data.   I'm just not running any mods or doing manual gathering of data.   

 

I'd suggest you ask neato.  I think he knows of a software app that does this sort of thing.  I think he said WotNumbers...   It would be really cool if neato created his own Post and did a tutorial on it.  (You know something more productive than just logging on and trolling... lol...)

SimplyPzB2 #52 Posted May 23 2019 - 07:59

    Sergeant

  • -Players-
  • 105 battles
  • 248
  • Member since:
    05-26-2016

View PostRingZero, on May 23 2019 - 03:50, said:

 

I checked your stats.   You are a very good player.  It would be very helpful to this conversation if we had access to battle results.  Any chance you've been collecting data?  If not, any chance you'd be willing to do so?  At least for a few hundred battles of a couple of your most played tanks?  neato says WotNumbers is easy to use....  We could put this issue to rest with data on a player like you.  All (almost) of your tanks have a win rate of over 60%.  So if we could see how your win rate is affected in only battles in the 40-60% win chance bracket (or similar range) fared, we could truly see if a very good player would be forced to 50%.    



Nixeldon #53 Posted May 23 2019 - 08:42

    Captain

  • Players
  • 59952 battles
  • 1,988
  • [PRTSN] PRTSN
  • Member since:
    10-30-2011

View PostSimplyPzB2, on May 23 2019 - 01:59, said:

I checked your stats.   You are a very good player.  It would be very helpful to this conversation if we had access to battle results.  Any chance you've been collecting data?  If not, any chance you'd be willing to do so?  At least for a few hundred battles of a couple of your most played tanks?  neato says WotNumbers is easy to use....  We could put this issue to rest with data on a player like you.  All (almost) of your tanks have a win rate of over 60%.  So if we could see how your win rate is affected in only battles in the 40-60% win chance bracket (or similar range) fared, we could truly see if a very good player would be forced to 50%.    

 

Do you seriously think someone is going to disprove math?

SquishySupreme #54 Posted May 23 2019 - 09:18

    First lieutenant

  • -Players-
  • 3116 battles
  • 568
  • Member since:
    10-29-2018

View Post_Tsavo_, on May 23 2019 - 00:00, said:

that's a lot of work, I run vanilla, and I'm not going to do look up individuals.  I'm here to play a game, not continue my probability and statistics course

 

And you know what the interesting part is?  You cannot make any statistical analysis at all of random battle mode.

15 players of different ability randomly thrown in a room and each given one life/die roll.  This is like having 15 randomly *sided* dice(skill levels) of random quantities that add up to 15 total dice and rolling them at once to get a single average.  Then repeating the process with another random set and another random outcome.  Then assigning the highest total from the two groups as the winner of the random "match".

Then repeating the process over again from scratch with all new random people with random skill levels for the next game.
Add in the game's RNG functions for aim and damage, which we don't actually know the exact math behind, and it's a bunch of actually random non-connected data sets with a sample size of 1.

WR is meaningless unless the same players play multiple matches or have multiple lives per match.  Any analysis that you can try to do with the current system is merely an illusion compared to hard stats like kills, damage per game, and accuracy. 

Edited by SquishySupreme, May 23 2019 - 09:20.


SimplyPzB2 #55 Posted May 23 2019 - 09:32

    Sergeant

  • -Players-
  • 105 battles
  • 248
  • Member since:
    05-26-2016

View PostNixeldon, on May 23 2019 - 08:42, said:

 

Do you seriously think someone is going to disprove math?

 

Someone already has....  that's why I want you to do this.  I know you won't be dragged down to 50%.  

View PostSquishySupreme, on May 23 2019 - 09:18, said:

 

And you know what the interesting part is?  You cannot make any statistical analysis at all of random battle mode.

15 players of different ability randomly thrown in a room and each given one life/die roll.  This is like having 15 randomly *sided* dice(skill levels) of random quantities that add up to 15 total dice and rolling them at once to get a single average.  Then repeating the process with another random set and another random outcome.  Then assigning the highest total from the two groups as the winner of the random "match".

Then repeating the process over again from scratch with all new random people with random skill levels for the next game.
Add in the game's RNG functions for aim and damage, which we don't actually know the exact math behind, and it's a bunch of actually random non-connected data sets with a sample size of 1.

WR is meaningless unless the same players play multiple matches or have multiple lives per match.  Any analysis that you can try to do with the current system is merely an illusion compared to hard stats like kills, damage per game, and accuracy. 

 

Exactly.  You can't approach this from the text definition of stats.  Too many variable in addition to variations in player skill.  All the random fanboys preach from on high about how 'statistics' HAVE to work (cuz my textbook says it does).  But in thier hypothetical 10,000 minimum sample size, they don't realize that's likely to be a mix of 200,000 different players.  Each of which is playing a different number of battles (and different tanks, different grinds, and different maps, et al).   So it's not only been proven with actual data, but the realization that their model does not factor in all the variables just goes to prove good players won't be dragged down to 50%.

Nixeldon #56 Posted May 23 2019 - 09:52

    Captain

  • Players
  • 59952 battles
  • 1,988
  • [PRTSN] PRTSN
  • Member since:
    10-30-2011

View PostNixeldon, on May 23 2019 - 02:42, said:

Do you seriously think someone is going to disprove math?

View PostSimplyPzB2, on May 23 2019 - 03:32, said:

Someone already has....  that's why I want you to do this.  I know you won't be dragged down to 50%. 

What?!?

 


Edited by Nixeldon, May 23 2019 - 09:52.


HorrorShow #57 Posted May 23 2019 - 10:36

    Sergeant

  • Players
  • 20007 battles
  • 205
  • Member since:
    06-26-2011
I sometimes use XVM for reasons many will dislike. As a "sky cancer," I sky click platoons with exceptional stats. If you take the advantage (platooning with great players), accept my bombardment.

NeatoMan #58 Posted May 23 2019 - 12:14

    Major

  • Players
  • 27986 battles
  • 19,924
  • Member since:
    06-28-2011

View PostSimplyPzB2, on May 23 2019 - 01:54, said:

I'd suggest you ask neato.  I think he knows of a software app that does this sort of thing.  I think he said WotNumbers...   It would be really cool if neato created his own Post and did a tutorial on it.  (You know something more productive than just logging on and trolling... lol...)

Not much to say.  go to the WoTNumbers website,  download the mod, install it and run it in the background when you play.  Sharing the data is as easy as uploading it to drop box.






2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users