Jump to content


The Most OP tanks in the game at the moment?

OP vehicle balance

  • Please log in to reply
76 replies to this topic

jylpah #21 Posted 24 April 2020 - 04:11 PM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

View PostWidd1, on 24 April 2020 - 06:03 PM, said:


Ahh no confidence in all those statistics ;)

 

Not taking on this - this time. I know this is a common (and flawed) position of some ;) I will return on this later...


Edited by jylpah, 24 April 2020 - 04:12 PM.

BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


SockRobber #22 Posted 25 April 2020 - 02:24 PM

    Senior Sergeant

  • Player
  • 21499 battles
  • 673
  • [PHNX]
  • Member since:
    12-05-2014

As ever fabulous work jylpah!

 

Recently on Console they 'started' a rebalancing phase and discussed their rebalancing methodology, and posted various graphs. Unfortunately I can't find the post/topic at this very second. But it was basically WR buckets for the tank vs for the tier(&possibly class).

A little similar to the 'dodgy' graphs on Tank Compare that no one understands, even me :teethhappy:(compared vs overall WR not per tier or per tier + class).

 

I also seem to remember similar for Blitz in the hazy past... Has there been anything recently? It might be good to mirror their methodology and have it publicly available for all tanks. It's certainly something I'm interested in re the 'zeitgeist'/TanksData tool.

 

I assume for these you are using 'all time' stats? Do/can you account for the 'new tank bounce' stats see, eg. lots of tanks for established players will have been played/ground in the distant past, and then sold, never to be touched again. Now in the present time, a 'newer' tank comes along and gets played, and the player is better. But also the old tank may have been buffed / nerfed since (directly or indirectly), and so on.

Is there any way to account for that?

 

Talking stats before my breakfast/coffee so excuse any rambling.........


BlitzStars.com - World of Tanks: Blitz player statistics signature

BlitzStars.com / (Blitz) Tank Compare

Player and Tank Statistics for WoT: Blitz


jylpah #23 Posted 25 April 2020 - 06:45 PM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

View PostSockRobber, on 25 April 2020 - 04:24 PM, said:

As ever fabulous work jylpah!

 

Recently on Console they 'started' a rebalancing phase and discussed their rebalancing methodology, and posted various graphs. Unfortunately I can't find the post/topic at this very second. But it was basically WR buckets for the tank vs for the tier(&possibly class).

A little similar to the 'dodgy' graphs on Tank Compare that no one understands, even me :teethhappy:(compared vs overall WR not per tier or per tier + class).

 

I also seem to remember similar for Blitz in the hazy past... Has there been anything recently? It might be good to mirror their methodology and have it publicly available for all tanks. It's certainly something I'm interested in re the 'zeitgeist'/TanksData tool.

 

Talking stats before my breakfast/coffee so excuse any rambling.........

 

Thanks!

 

WG has said they use mainly the performance of 55-65% players to balance the tanks. They post those WR / damage chart for every update to their Discord to #dev-answers channel:

 

 

The issue with their approach is that the graphs do not take into consideration differences in the player population that might cause bias in some cases. E.g. Sheridan players are better in average than BatChat players (but not so much to explain the amount of WR difference in the graph).

 

SockRobber, on 25 April 2020 - 04:24 PM, said:

I assume for these you are using 'all time' stats? Do/can you account for the 'new tank bounce' stats see, eg. lots of tanks for established players will have been played/ground in the distant past, and then sold, never to be touched again. Now in the present time, a 'newer' tank comes along and gets played, and the player is better. But also the old tank may have been buffed / nerfed since (directly or indirectly), and so on.

Is there any way to account for that?

 

My graph was only for battles made during 6.8. I ran the stats fetching during each update and calculate a diff between the (cumulative) snapshots retrieved from WG API.

 

For this graph I selected only those players with more than 30 battles during the update 6.8 AND the sum of battles before 6.8 and during the 6.8 has to be more than 60. This way I filter out players who either played too few battles and got lucky/unlucky and players who just started the grind.

 

R code:

stats.update.tier[ (WR.tier >.55) & (WR.tier<.65)&(all.battles + all.battles.old > 60) & (all.battles > 30) & (type %in% c('Light Tank', 'Medium Tank';)), 
                   .(Tank=unique(name), WR.tank = mean(WR)*100, WR.tier=mean(WR.tier)*100, WR.delta = mean(WR - WR.tier)*100, Avg_Dmg = mean(avg_dmg), Players = uniqueN(account_id)), 
                   by=tank_id][order(WR.delta)]

 

This part is filter/row select

 

R has lovely syntax, doesn't it ?


BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


SockRobber #24 Posted 25 April 2020 - 08:54 PM

    Senior Sergeant

  • Player
  • 21499 battles
  • 673
  • [PHNX]
  • Member since:
    12-05-2014

View Postjylpah, on 25 April 2020 - 06:45 PM, said:

 

Thanks!

 

WG has said they use mainly the performance of 55-65% players to balance the tanks. They post those WR / damage chart for every update to their Discord to #dev-answers channel:

 

 

The issue with their approach is that the graphs do not take into consideration differences in the player population that might cause bias in some cases. E.g. Sheridan players are better in average than BatChat players (but not so much to explain the amount of WR difference in the graph).

 

 

My graph was only for battles made during 6.8. I ran the stats fetching during each update and calculate a diff between the (cumulative) snapshots retrieved from WG API.

 

For this graph I selected only those players with more than 30 battles during the update 6.8 AND the sum of battles before 6.8 and during the 6.8 has to be more than 60. This way I filter out players who either played too few battles and got lucky/unlucky and players who just started the grind.

 

R code:

stats.update.tier[ (WR.tier >.55) & (WR.tier<.65)&(all.battles + all.battles.old > 60) & (all.battles > 30) & (type %in% c('Light Tank', 'Medium Tank';)), 
                   .(Tank=unique(name), WR.tank = mean(WR)*100, WR.tier=mean(WR.tier)*100, WR.delta = mean(WR - WR.tier)*100, Avg_Dmg = mean(avg_dmg), Players = uniqueN(account_id)), 
                   by=tank_id][order(WR.delta)]

 

This part is filter/row select

 

R has lovely syntax, doesn't it ?

 

Ah yeah; the natural, and problematic (for the stats), attraction of better players to better tanks!

Seems your method - and perhaps the console team doing the bucketed version - is a lot better.

 

Not sure I can adequately come up with something for BlitzStars; while I really like the data and these kinds of stats, the interest is pretty niche as far as the site goes I think, and it also takes a lot of computing power on the server - in a relative sense - when in general I like to keep anything happing on there to under 10ms if possible.

As ever, a Q of what I have time to work on that will benefit the most people the most (or that I really, really want to work on ;) ).

 

But still, something along these lines would be a great addition to the zeitgeist or the 'tanks' page....

 

We'll see! Got a couple of other things in the pipeline at the minute.

 


BlitzStars.com - World of Tanks: Blitz player statistics signature

BlitzStars.com / (Blitz) Tank Compare

Player and Tank Statistics for WoT: Blitz


Corpsetaker #25 Posted 25 April 2020 - 09:00 PM

    Senior Sergeant

  • Player
  • 22137 battles
  • 506
  • [TAOW]
  • Member since:
    09-07-2016
The Type 62 is one of the best tanks in the game.

jylpah #26 Posted 25 April 2020 - 11:00 PM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

View PostCorpsetaker, on 25 April 2020 - 11:00 PM, said:

The Type 62 is one of the best tanks in the game.

 

I have rediscovered it. It is fantastic. Speed + DPM.


BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


Guard_Of_Justitia #27 Posted 25 April 2020 - 11:01 PM

    First Sergeant

  • Player
  • 32163 battles
  • 1,041
  • [HIALL]
  • Member since:
    03-10-2016

View Postjylpah, on 25 April 2020 - 06:45 PM, said:

 

Thanks!

 

WG has said they use mainly the performance of 55-65% players to balance the tanks. They post those WR / damage chart for every update to their Discord to #dev-answers channel:

 

 

The issue with their approach is that the graphs do not take into consideration differences in the player population that might cause bias in some cases. E.g. Sheridan players are better in average than BatChat players (but not so much to explain the amount of WR difference in the graph).

 

 

My graph was only for battles made during 6.8. I ran the stats fetching during each update and calculate a diff between the (cumulative) snapshots retrieved from WG API.

 

For this graph I selected only those players with more than 30 battles during the update 6.8 AND the sum of battles before 6.8 and during the 6.8 has to be more than 60. This way I filter out players who either played too few battles and got lucky/unlucky and players who just started the grind.

 

R code:

stats.update.tier[ (WR.tier >.55) & (WR.tier<.65)&(all.battles + all.battles.old > 60) & (all.battles > 30) & (type %in% c('Light Tank', 'Medium Tank';)), 
 .(Tank=unique(name), WR.tank = mean(WR)*100, WR.tier=mean(WR.tier)*100, WR.delta = mean(WR - WR.tier)*100, Avg_Dmg = mean(avg_dmg), Players = uniqueN(account_id)), 
 by=tank_id][order(WR.delta)]

 

This part is filter/row select

 

R has lovely syntax, doesn't it ?

 

Just wondering, given the Maus is quite high, do you suppose WG will nerf it? And if so how?


 

 


jylpah #28 Posted 26 April 2020 - 10:40 AM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

View PostGuard_Of_Justitia, on 26 April 2020 - 01:01 AM, said:

 

Just wondering, given the Maus is quite high, do you suppose WG will nerf it? And if so how?


Well. It has been one of the most competitive tanks at tier X for long time. The recent Tier X HT HP buff just made tanks like Maus & IS-4 stronger. 
 

WG doesn’t not balance tanks based on pure performance, but they also have their “strategic plan” behind concerning how much OPness is allowed and to which tanks. Few examples:

 

  1. frequent expensive OP premiums
  2. FV183 has been nerfed to one of the worst performing tanks due to its “over-popularity” and impact on meta
  3. HT buffs through the tiers (med prammo nerf, HT HP buff, buffed-through-the-roof-Tiger1-Tiger2-VK36H, etc) to make the game more friendly to low-skilled players. 

BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


jylpah #29 Posted 30 April 2020 - 01:36 PM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

6.9 Stats are out

 

Tier X

 

Tier IX

 

Tier VIII

 

 

Tier VII

Tier VI

Tier V

 

 

Statspadders, please go and buy KV-220 Beta-Test when it next time emerges to stores. It is the most OP tank in the game with a margin. :popcorn:


BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


Jukkis74 #30 Posted 30 April 2020 - 02:15 PM

    Senior Sergeant

  • Player
  • 37483 battles
  • 986
  • [AFK]
  • Member since:
    06-29-2014
T8 looks horribly pay-to-win with OP premiums. Sad. Maybe I should go and take my KV-220 out for a drive :-)

llo1 #31 Posted 30 April 2020 - 03:03 PM

    First Sergeant

  • Player
  • 32157 battles
  • 7,717
  • [1RTR]
  • Member since:
    06-27-2014

Sorry jylpah, KV220T OP!!..,?

not in my hands, or my games.....always bottom tier, alway full of useless players, and the GUN!!!

Doesnt pen, doesn’t cause enough alpha, and is penned by every other tank in the tier/game........

4 recent games, 4 bottom tier, 4 bottom tier, 3 5 shots and only 1 pen, and every hit on me pens!......

That tells me it is not OP for me!


Edited by llo1, 30 April 2020 - 03:14 PM.


jylpah #32 Posted 30 April 2020 - 03:33 PM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

View Postllo1, on 30 April 2020 - 05:03 PM, said:

Sorry jylpah, KV220T OP!!..,?

not in my hands, or my games.....always bottom tier, alway full of useless players, and the GUN!!!

Doesnt pen, doesn’t cause enough alpha, and is penned by every other tank in the tier/game........

4 recent games, 4 bottom tier, 4 bottom tier, 3 5 shots and only 1 pen, and every hit on me pens!......

That tells me it is not OP for me!

 

I am really sorry to hear that. I guess it does not help that the other 1335 KV-220 Beta-test players reach on average 9.6% higher WR in the tank vs their other Tier V tanks?

 

Nerd alert :)

 

 


BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


Replicant_2 #33 Posted 30 April 2020 - 04:45 PM

    First Sergeant

  • Player
  • 13546 battles
  • 1,383
  • [HRG]
  • Member since:
    10-08-2018

Fantastic work jylpah, makes me even happier that I spent some of my free bing vouchers on the K-91 :P:izmena:

 

my WR in the KV-220 is just under 74%, a bit up from the average 53% for tier V (including the KV), again, given its price it was such a well spent couple of quid :trollface:

Shame I actually rarely play it.


All those moments will be lost in time, like tears in rain. Time to die.


Tijgerhaai_XIV #34 Posted 30 April 2020 - 06:19 PM

    First Sergeant

  • Player
  • 12554 battles
  • 2,376
  • [ALBA]
  • Member since:
    06-08-2016

The only thing is that this awesome analysis doesn't really work for Light tanks in tiers IX and X, as there are only two options. The figures only compare the performance of a player between US and French Lights. This tells us nothing about where the "balance" is relative to tier, they both could be under-performing or both over-performing. 

 

So the T92E1's high placement on the graph behind the K-91 is misleading, as it only says how much better it has performed against the BatChat. So it could say more about how bad the BatChat is than how good it is; conversely, if the BatChat is performing well that would hide how well the Missile tank is doing. The true relative performance would have to take more into account given the low numbers of tanks, such as the player's average in mediums of the tier indexed to the average of the lights, or something appropriately complex.


 


llo1 #35 Posted 30 April 2020 - 06:26 PM

    First Sergeant

  • Player
  • 32157 battles
  • 7,717
  • [1RTR]
  • Member since:
    06-27-2014

View Postjylpah, on 30 April 2020 - 03:33 PM, said:

 

I am really sorry to hear that. I guess it does not help that the other 1335 KV-220 Beta-test players reach on average 9.6% higher WR in the tank vs their other Tier V tanks?

 

Nerd alert :)

 

 

Are you sure the graphs are showing best to the left, as for me these seem to be ‘tother way around!

Have you got your axis label correct and not ticked “data in reverse”............



jylpah #36 Posted 30 April 2020 - 06:38 PM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

View Postllo1, on 30 April 2020 - 08:26 PM, said:

Are you sure the graphs are showing best to the left, as for me these seem to be ‘tother way around!

Have you got your axis label correct and not ticked “data in reverse”............

:teethhappy:


BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


jylpah #37 Posted 30 April 2020 - 06:43 PM

    First Sergeant

  • Player
  • 27141 battles
  • 5,682
  • [AFK]
  • Member since:
    08-05-2014

View PostTijgerhaai_XIV, on 30 April 2020 - 08:19 PM, said:

The only thing is that this awesome analysis doesn't really work for Light tanks in tiers IX and X, as there are only two options. The figures only compare the performance of a player between US and French Lights. This tells us nothing about where the "balance" is relative to tier, they both could be under-performing or both over-performing. 

 

So the T92E1's high placement on the graph behind the K-91 is misleading, as it only says how much better it has performed against the BatChat. So it could say more about how bad the BatChat is than how good it is; conversely, if the BatChat is performing well that would hide how well the Missile tank is doing. The true relative performance would have to take more into account given the low numbers of tanks, such as the player's average in mediums of the tier indexed to the average of the lights, or something appropriately complex.


Nope. It compares the players average in they other tier IX tanks - all vehicle classes included. This is IMHO better way of doing the comparison than what I used earlier that was tier average including the tank in question. Since T92E1 is so popular and much-played tank, players' WR in it is closer to their overall WR in all IX tanks (since large share of the tier average WR was made of battles played with the very same tank).

 

Someone knowledgeable of R could have pointed this out earlier ;)


Edited by jylpah, 30 April 2020 - 08:22 PM.

BlitzAnalysiz.com  In-depth analysis of WoT Blitz game statistics

Jylpah @EU  Jylpah @Asia  Blitz-tools  


Tijgerhaai_XIV #38 Posted 30 April 2020 - 09:58 PM

    First Sergeant

  • Player
  • 12554 battles
  • 2,376
  • [ALBA]
  • Member since:
    06-08-2016

View Postjylpah, on 30 April 2020 - 07:43 PM, said:


Nope. It compares the players average in they other tier IX tanks - all vehicle classes included. This is IMHO better way of doing the comparison than what I used earlier that was tier average including the tank in question. Since T92E1 is so popular and much-played tank, players' WR in it is closer to their overall WR in all IX tanks (since large share of the tier average WR was made of battles played with the very same tank).

 

Someone knowledgeable of R could have pointed this out earlier ;)

Ok, my bad, from my quick glance at you posted immediately before these charts I thought it was comparing WR to tier+class WR. 

 

Completely my error, please pretend I said nothing, like everyone else that has me blocked ^^


 


llo1 #39 Posted 30 April 2020 - 10:09 PM

    First Sergeant

  • Player
  • 32157 battles
  • 7,717
  • [1RTR]
  • Member since:
    06-27-2014
Going back to the OP tanks.......the red ones that I always seem to find in my games!

Debbie_Does_Dallas #40 Posted 01 May 2020 - 08:03 AM

    First Sergeant

  • Player
  • 72209 battles
  • 1,938
  • [HE-49]
  • Member since:
    06-13-2016
Imo there are no OP tanks, only noob friendly vehicles. The smasher will help a 'pond skater' achieve a reasonable result as opposed to the same player driving something like a T71, yet a T71 in the hands of an excellent light tank player could be considered OP. It's my experience that a tank like the bulldog is almost a liability on my team due to it having very little going for it, but once in a while you'll see one driven by a skilful madman...normally if I'm in a smasher I salivate at the site of a red bulldog but I once got my pants pulled down and thoroughly spanked by an excellent exponent of this vehicle...thus making on this occasion the vehicle OP because of his ability.

And when I reach the gates of heaven, Saint Peter I will tell..."one more soldier reporting sir, I've served my time in hell".

verse found among possessions of dead american GI on omaha beach. Normandy 1944.






Also tagged with OP, vehicle balance

1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users