104 replies to this topic

[BlitzStats]

Updated tank-compare with 3.5 data including the new Lunar New Year (or whatever it is) reward tanks.

Edited by BlitzStats, 25 January 2017 - 06:16 PM.

[BlitzStats]

Tank Compare got a little update today.

 

Previously the 'BlitzStars Data' showed the data used in the 'tank averages' calculations (a 40k random sample of players' career statistics, not specifically BlitzStars users).

Now the BlitzStars Data section shows BlitzStars tracked player's statistics for the previous 90 days. This should provide more of a 'rolling commentary' on tank performance, although will be skewered higher than the previous data.

 

I've also added some graphs to show individual tank performances versus the player's overall performance (shows 90day and career).

As a basic guide to reading the graphs: If players consistently have lower winrates in a tank than their overall winrate, the orange and red lines will be lower than the 'baseline' blue line (which shows where a winrate equal to the player's overall winrate would fall).

The area data shows the % of players at a given winrate.

In the example below, we can see RU 251 players generally have a lower winrate in the RU251 than they do in as an average of all their tanks. We can see that T71 players generally do worse also, specifically better players (data further right) do even worse - no doubt due to the extremely low DPM and carry ability of the tank when it was released. The Type 62 shows another story - the tank is fairly strong for all, but excels in the hands of a player whose overall winrate is also high (ie. better players, generally speaking). 

 

X Axis: Players' overall winrate (account)

Y Axis: Players' winrates in specific tank (2nd Y axis/area data: % of players - ie number of players at specific winrate)

 

[RageKicker]

Nice, thanks for the work you do!

[BlitzStats]

Updated with data from 3.6 - new French line and the various other vehicle changes.

[BlitzStats]

Update for 3.8

 

So I've updated the Tank Compare tool with 3.8 data and equipment. It hasn't had much testing so if you encounter any issues please report them in this thread (or email or twitter or whatever...).

[CC_Sly]

Pinned for ease of access.

[cheasesteak]

Very nice.  I like the ability to see the different equipment options side by side.  Unless I am using it wrong, you only have 4 sets of equipment offered.  Binos and the level 2 mobility options might be good additions.

[BlitzStats]

cheasesteak, on 05 May 2017 - 04:36 PM, said:

Very nice.  I like the ability to see the different equipment options side by side.  Unless I am using it wrong, you only have 4 sets of equipment offered.  Binos and the level 2 mobility options might be good additions.

 

Several of the new equipment items do not directly have an affect on any of the stats shown - so those are missing.

 

3 (I believe) are missing that *do* have a direct effect on shown stats - namely optics, 'Improved Control' and 'Engine Accelerator'.

 

I'm not showing the latter two currently because they both primarily effect engine hp and traverse. And traverse is borked (WoT PC formula to calculate traverse does not seem to work for Blitz - and there's no working Blitz formula that I'm aware of).

 

So I'm not quite sure how to deal with them for now. Not much point adding them when one of the main stats involved is not calculated correctly. (Tank Compare shows the 'base' traverse number as reported by the client, no crew skills or equipment modifiers used).

 

If I, or someone, comes up with a way to get accurate traverse calculations working that would change.

 

Optics - not currently shown/enabled as it breaks the UI/ two banks of four displaying equipment, and is super easy to do in your head if you want to (10%...).

[little_brother_of_war]

CC_Sly, on 05 May 2017 - 10:36 AM, said:

Pinned for ease of access.

 

thanks, Sly!

[Doc_01_]

Game is a rigged piece of garbage

[BlitzStats]

Tank Compare update for 3.9

 

I've updated Tank Compare to use 3.9 game data.

 

Ammo changes seem to be included - though I haven't checked the figures myself in-game.

 

As always, feedback and bugs can be reported here or via usual channels.

[BigBrigham]

You, sunshine, are a legend! +1

Edited by BigBrigham, 26 May 2017 - 11:44 PM.

[BarbierDeTVils_NA]

BlitzStats, on 24 May 2017 - 02:42 PM, said:

Tank Compare update for 3.9

 

I've updated Tank Compare to use 3.9 game data.

 

Ammo changes seem to be included - though I haven't checked the figures myself in-game.

 

As always, feedback and bugs can be reported here or via usual channels.

 

Very good site 

gg to you 

 

I hope see it in French traduction 

 

careful some of your datas are wrong on crédit coefficient 

[ElderMaxson]

.

Edited by DirtFriend, 19 June 2017 - 09:01 AM.

[1991Syclone]

Late to the party here, but could you add a section at the bottom that shows the 7 day rolling stats fora  given tank.  This might give a better idea of what it will take to Ace a tank (although I think you'd need to include the XP amount in addition to the damage done), and I'm not sure if the API publishes that.

 

Ohhhh, and how about the tanks with the minimum average damage per type/tier.  If someone had a mission to get a 1st class or ace a tank, it might help to know which ones are the lowest performing so someone would have an idea of where to start.

Edited by 1991Syclone, 19 June 2017 - 12:08 PM.

[BlitzStats]

BarbierDeTVils_NA, on 15 June 2017 - 09:44 PM, said:

Very good site 

gg to you 

 

I hope see it in French traduction 

 

careful some of your datas are wrong on crédit coefficient 

 

Thank you.

If you would like to provide a French translation feel free to get in touch with me and I'll send you the relevant details. All the translations from English are provided by community members.

The credit coeff figures are also community provided - no 'official' numbers exist - and have to be entered manually. Thus likely to 'fall behind' sometimes and not be 100% accurate. They are a 'best guess' (certainly within a few % by all accounts).

 

1991Syclone, on 19 June 2017 - 12:07 PM, said:

Late to the party here, but could you add a section at the bottom that shows the 7 day rolling stats fora  given tank.  This might give a better idea of what it will take to Ace a tank (although I think you'd need to include the XP amount in addition to the damage done), and I'm not sure if the API publishes that.

 

Ohhhh, and how about the tanks with the minimum average damage per type/tier.  If someone had a mission to get a 1st class or ace a tank, it might help to know which ones are the lowest performing so someone would have an idea of where to start.

 

7 day rolling stats are potentially possible.

 

However, I'm not sure you could deduce the things you want to from them.

Regarding masteries - XP is all that matters - and XP data provided by WG includes 'premium' accounts bonus (but not other bonuses), so the numbers get muddled and are somewhat unusable.

Even if this wasn't the case, the 'averages' we could get from the data provided wouldn't particularly be relevant to the 'best' games that the mastery system uses.

Lowest performing tanks probably means 'worst' tanks. I suppose you could argue they might be easier to ace as better players are less likely to play bad tanks, but it's a stretch. If XP averages are low on a tank, that doesn't mean it's easy to go and ace. It, more likely, just means those tanks perform worse. The relative difficulty to ace them would remain approximately the same (this makes various assumptions about the balance of tanks and how easy/difficult a tank is to play, but overall is probably a fair assumption).

[BlitzStats]

Updated with patch 3.10 data

[BarbierDeTVils_NA]

BlitzStats, on 19 June 2017 - 03:15 PM, said:

 

Thank you.

If you would like to provide a French translation feel free to get in touch with me and I'll send you the relevant details. All the translations from English are provided by community members.

The credit coeff figures are also community provided - no 'official' numbers exist - and have to be entered manually. Thus likely to 'fall behind' sometimes and not be 100% accurate. They are a 'best guess' (certainly within a few % by all accounts).

 

I could help you with pleasure for French version.

 

for crédit coefficient, even if it don't really say anything about tank rentability, regarding about repair and ammunitions costs, it is quite simply to calculate it with about 10 real results of battles win or loose. It don't exist official WG numbers but I could give you them and justify them without pb  

 

For example :

 

- Lupus is 150% without mistake

 

- but its repair cost is quitly more important than others premiums tanks tier 7, this cost could be compare to a standard tank. It is about 25% to 40% more if totaly destroyed than another tier 7 premium tank.

 

https://tof.cx/image...d6cc463050e.png

 

https://tof.cx/image...149a031caab.png

 

 

Edited by BarbierDeTVils_NA, 24 June 2017 - 07:01 PM.

[BlitzStats]

BarbierDeTVils_NA, on 24 June 2017 - 06:12 PM, said:

 

I could help you with pleasure for French version.

 

for crédit coefficient, even if it don't really say anything about tank rentability, regarding about repair and ammunitions costs, it is quite simply to calculate it with about 10 real results of battles win or loose. It don't exist official WG numbers but I could give you them and justify them without pb  

 

For example :

 

- Lupus is 150% without mistake

 

- but its repair cost is quitly more important than others premiums tanks tier 7, this cost could be compare to a standard tank. It is about 25% to 40% more if totaly destroyed than another tier 7 premium tank.

 

https://tof.cx/image...d6cc463050e.png

 

https://tof.cx/image...149a031caab.png

 

 

 

If you email me - sockrobber at blitzstars dot com - I will send you the document / details regarding the translation. A french language version would be great.

 

You're absolutely right about the credit coefficient 'issue'. But I'm not sure how we can make a better metric for it, because repair and ammo costs are going to depend a lot on external factors (particularly - quality of player).

[BootlegTampico]

BlitzStats, on 29 June 2017 - 12:40 PM, said:

 

If you email me - sockrobber at blitzstars dot com - I will send you the document / details regarding the translation. A french language version would be great.

 

You're absolutely right about the credit coefficient 'issue'. But I'm not sure how we can make a better metric for it, because repair and ammo costs are going to depend a lot on external factors (particularly - quality of player).

 

I remember you mentioning you'd be interested in camouflage values. I have most Tier Xs done. I can direct you to my post if you'd like.