Jump to content


Real stats repository


  • Please log in to reply
59 replies to this topic

gallabru #21 Posted 15 February 2019 - 10:19 AM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

Thanks Jylpah.

 

I have had a look at the evolution of WR of the players with battle count. Good news : it increases. Bad news : slowly. Might be out of experience or discouragement. Or newer players are worse and will remain while older ones were always better and remain better. It is interesting to notice a bunch of very good players (70% WR) in the low battle count players... who could they be ?

This might explain why the average of the WR of the players I was seeing on both accounts was different. This also mean that the WR distribution width I have plotted above was slightly overestimated. Once corrected for the variation with battle count, the width is 5.6%.

 

I have plotted together the data of my main account (orange) and alt account (blue). The blue dots are drawn on top, so none is hidden by orange ones. Both accounts have seen players with any number of battles, but I have added a vertical line at 5000. You can draw conclusions by yourself, I guess.

 

I understand better also why it is more difficult to play on my main account than on my alt. Players are better. Simple.

 

 

Oh, and yes, I have seen one player with nearly 100 000 battles and a WR below 30%. That would deserve some sort of reward imo.


Edited by gallabru, 20 February 2019 - 02:19 PM.

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


jylpah #22 Posted 15 February 2019 - 08:15 PM

    First Sergeant

  • Player
  • 20852 battles
  • 4,820
  • [AFK]
  • Member since:
    08-05-2014

View Postgallabru, on 15 February 2019 - 12:19 PM, said:

Thanks Jylpah.

 

I have had a look at the evolution of WR of the players with battle count. Good news : it increases. Bad news : slowly. Might be out of experience or discouragement. Or newer players are worse and will remain while older ones were always better and remain better. It is interesting to notice a bunch of very good players (70% WR) in the low battle count players... who could they be ?

This might explain why the average of the WR of the players I was seeing on both accounts was different. This also mean that the WR distribution width I have plotted above was slightly overestimated. Once corrected for the variation with battle count, the width is 5.6%.

 

I have plotted together the data of my main account (orange) and alt account (blue). The blue dots are drawn on top, so none is hidden by orange ones. Both accounts have seen players with any number of battles, but I have added a vertical line at 5000. You can draw conclusions by yourself, I guess.

 

I understand better also why it is more difficult to play on my main account than on my alt. Players are better. Simple.

 

 

 

Oh, and yes, I have seen one player with nearly 100 000 battles and a WR below 30%. That would deserve some sort of reward imo.

 

If I understood the graph correctly it displays WR and battle count of the players your two accounts have met in the battlefield, right? It could be that 1) your 2.8k account has just recently been elevated to the main queue or 2) WG is slowly introducing more experienced players to the newbies or 3) sometimes there are too few players so the MM just merges the two queues. #1 or #3 are my guesses.

 

 



gallabru #23 Posted 15 February 2019 - 10:38 PM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

Right. 

I have always seen players with high battle counts (there is a plot with when I saw them). I started this account in the same time (2016) as my main and let it dormant until recently. So maybe I get a different treatment, but the battles were recorded after 5.5. Still, I clearly see more newer players on this account. I have the time stamp, so I could track when i see more experienced players. But hypothesis 3 definitely is part of it. 


My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


Mjr_Eazy #24 Posted 16 February 2019 - 08:26 AM

    First Sergeant

  • Player
  • 30215 battles
  • 6,699
  • [BT_SV]
  • Member since:
    04-26-2017

Amazing work m8, if only we could get one of those claiming to get mostly low tier battles (68% low tier one claimed if I remember correctly) to provide some data, it would make an interesting comparison and maybe show some sot of bias/manipulation, rigging even ;)

 

As you say, your 100k sub 30% is a spectacular player, I’m sure I’d read somewhere before that a serial afk’er gets nearer 37%!


547146227.png

 


jylpah #25 Posted 16 February 2019 - 08:29 AM

    First Sergeant

  • Player
  • 20852 battles
  • 4,820
  • [AFK]
  • Member since:
    08-05-2014

View Postgallabru, on 16 February 2019 - 12:38 AM, said:

Right. 

I have always seen players with high battle counts (there is a plot with when I saw them). I started this account in the same time (2016) as my main and let it dormant until recently. So maybe I get a different treatment, but the battles were recorded after 5.5. Still, I clearly see more newer players on this account. I have the time stamp, so I could track when i see more experienced players. But hypothesis 3 definitely is part of it. 

 

Just a question. Are those both account playing the same tiers?

gallabru #26 Posted 16 February 2019 - 09:48 AM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

Yes same average tier : 60 days is 7.3, carrier is 5.8. 

Must be some atavism.  


My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


gallabru #27 Posted 16 February 2019 - 09:52 AM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016
Mjr, maybe once you use replays the rigging stops. How can you know ? ;)

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


Mjr_Eazy #28 Posted 16 February 2019 - 10:39 AM

    First Sergeant

  • Player
  • 30215 battles
  • 6,699
  • [BT_SV]
  • Member since:
    04-26-2017

View Postgallabru, on 16 February 2019 - 09:52 AM, said:

Mjr, maybe once you use replays the rigging stops. How can you know ? ;)

 

Lol, true...

Maybe cos we’ve posted we’re looking for it, they switched off the Winning and Losing Queues for us, who knows ;)


547146227.png

 


gallabru #29 Posted 20 February 2019 - 02:17 PM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

More of a gimmick. Average WR of the players as a function of hour in the day. Combined data of both my accounts week day and week-ends. There are actually not many data (54 at most between 20:00 and 21:00), so it is difficult to go further. The variations must not be really significant. 

All the plots on my alt have been updated with 158 battles. Nothing new.

The WR distributions of reds and green on both accounts have been normalized to the number of players for an easier comparison.

Here is the new correlation of the WR difference of the teams between consecutive battles on my alt after 158 battles. 0 means impossibility to predict the WR difference at battle N after battle the reference battle (number 0). 1 means same WR difference, -1 opposite WR difference. Dotted line = 95% confidence.


Edited by gallabru, 21 February 2019 - 11:12 AM.

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


Mjr_Eazy #30 Posted 20 February 2019 - 05:40 PM

    First Sergeant

  • Player
  • 30215 battles
  • 6,699
  • [BT_SV]
  • Member since:
    04-26-2017
It would be interesting to see if you win more than your WR at any particular times of the day, I’m convinced I win more between 8 and 11 in the morning...

547146227.png

 


gallabru #31 Posted 20 February 2019 - 09:32 PM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

Hahaha, this was the idea. See if there were time where I could noob out and still win. 

I know I am losing more from 11pm on (continental time). 

The thing with winning/losing is that it is also related to me (tiers, tired, pissed...). 

But yes, I could try to correlate some factors to my WR.  


My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


gallabru #32 Posted 21 February 2019 - 12:10 PM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

I thought I had found something. Alas...

I have plotted how the the WR difference changes from one battle to the other. I was looking for some sort of compensation : "you have had a tough battle, let's make the next easier". Ok many may think that the rule is more : "you have had a tough battle, let's make the next more unfair".

 

Oh oh, I saw that in my battles !

 

But then I generated random number with the same distribution as the one of the players. Gives the same thing. So it seems more related to random numbers.


Edited by gallabru, 21 February 2019 - 12:11 PM.

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


gallabru #33 Posted 10 April 2019 - 03:46 PM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

Just pasting here what I put in the 'top tier / bottom tier' thread.

On my alt account. I have started a new 100 battle survey after 5.9... we will see.

and on my main account

I guess the statistics per tier are too low to conclude definitely.


Edited by gallabru, 10 April 2019 - 03:52 PM.

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


minitelrose #34 Posted 11 April 2019 - 06:52 AM

    First Sergeant

  • Player
  • 76333 battles
  • 8,008
  • [AFK]
  • Member since:
    07-06-2014
Keep the good work !

Read my guides there

"I want the game to be just how it was launched, no MM limitations for platoons, some unbalanced match once in a while, and friendly fire should come back as well. The only thing they should remove is statistics." -phony1907, 08 July 2016, 05:45pm - 

I want just that, and replay files.

 


KV_007 #35 Posted 12 July 2019 - 10:34 PM

    Private

  • Player
  • 13279 battles
  • 4
  • [ERWB]
  • Member since:
    04-03-2017
Just out of curiosity; have you continued to keep up the statistics since? It would be very interesting to know what it looks like now that some time has passed and more data would be available?

gallabru #36 Posted 15 July 2019 - 07:56 AM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016
I have started to save consecutive games again. My reroll is still under 5k. We will see how things evolve.

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


gallabru #37 Posted 03 September 2019 - 02:02 PM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016

Since last update and MM improvement (6.2), it was interesting to see what happened to the playerbase...

For some reason a sizeable proportion of my replays did not upload correctly on wotbinspector. Wondering if this has to do with jumps in server. Anyway...

Newbie account (<5k battles), tiers 7-9

 

Battles Red average WR Green average WR Red battle count Green battle count
50 48.8% 48.8% 4500 4499

 

Very small sample. Still, only the expected average values.

That hints towards that the 'purgatory' is not the newbie queue and that my feeling that there are many more 40%wr players is only a feeling. I will need to see in more details what the distribution is, though.

 

Ok, here are the WR distributions of green and red teams. Same distributions, centred in 47.3% with a variance of 6.3%. Some noise, but I guess things should get smoother with number of battles.

Posted Image

Thx to @Jylpah for the new program files


Edited by gallabru, 04 September 2019 - 10:19 AM.

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


Chairman_merpug #38 Posted 03 September 2019 - 05:49 PM

    First Sergeant

  • Player
  • 49971 battles
  • 9,322
  • [AFK]
  • Member since:
    03-05-2013
Hmm. So the average hasn't changed.

Denoobing member of the Active Frontline Klan - and proud of it. Be nice to me..https://www.blitzstars.com/player/eu/Chairman_merpug


gallabru #39 Posted 03 September 2019 - 06:29 PM

    Senior Sergeant

  • Player
  • 17782 battles
  • 584
  • [LAGER]
  • Member since:
    09-01-2016
Hmmm,yes you are right, lots of effort to say that yhe average is the average.

But I find interesting to see it that way : there are as many 42% wr players as 52% wr.
However, we often present 42% wr as very bad players, nearly trolls, whereas 52% wr players are seen as ok-ish. Hence a bias in the reading of the team composition.

My team stats in real time  : http://forum.wotblit...ats-repository/

Carlsberg clan family is recruiting. Contact us.


Mjr_Leicht #40 Posted 03 September 2019 - 06:44 PM

    Sergeant

  • Player
  • 5736 battles
  • 304
  • [BT_SV]
  • Member since:
    01-21-2019

Nice one m8, keep up the good work. It will be interesting to see what it looks like with more battles...

Re upload errors have updated your maps.json and tanks.json from gitlab?  If not then do that, it will fix a bunch of errors as scripts need them for details for new maps/tanks,  or you can run the scripts for updating them...

It’s probably also worth updating the scripts themselves as sometimes jylpah has to update them cos of WG changes...

The other causes of errors are:

- players with no stats in WG api, these are either deleted accounts, CC accounts or bots, it will be interesting to see if you are getting lots of these, I’m suspicious about bots...

- sometimes you get errors with WG API if it’s getting hit too hard so they might load next time...

Most likely cause is out of date maps/tanks files...

Oh and WG crapness I usuall have between 5-10% empty replay files from the game :(

HTH


Edited by Mjr_Leicht, 03 September 2019 - 06:49 PM.

560273467.png




7 user(s) are reading this topic

0 members, 7 guests, 0 anonymous users