16 replies to this topic

[jylpah]

Hello,

 

I have been working on the scripts (https://github.com/Jylpah/blitz-tools) to add new functionality : analysis of the player population based on the replays. I have not updated the GitHub version yet, but here are the first results based on 82 replays (all T-54 i.e. tier IX and played in late August - early September). The population data does not include myself, only the teammates and enemies. Actually the script also excludes ones' platoon buddies if any. 

 

The Win rate here is not the players global WR, but WR and all other stats (including battle count) in the tier of tank he played.

 

Win rate     | Allies        | Enemies       | TOTAL
0% - 35%     |     1 ( 0.2%) |     3 ( 0.5%) |     4 ( 0.4%)
35% - 40%    |    16 ( 3.3%) |    28 ( 4.9%) |    44 ( 4.1%)
40% - 45%    |   104 (21.1%) |   122 (21.3%) |   226 (21.2%)
45% - 50%    |   152 (30.9%) |   187 (32.6%) |   339 (31.8%)
50% - 55%    |   134 (27.2%) |   142 (24.7%) |   276 (25.9%)
55% - 60%    |    60 (12.2%) |    67 (11.7%) |   127 (11.9%)
60% - 65%    |    19 ( 3.9%) |    16 ( 2.8%) |    35 ( 3.3%)
65% -        |     6 ( 1.2%) |     9 ( 1.6%) |    15 ( 1.4%)

 

Avg. Dmg.    | Allies        | Enemies       | TOTAL
0 - 250      |     1 ( 0.2%) |     0 ( 0.0%) |     1 ( 0.1%)
250 - 500    |     0 ( 0.0%) |     3 ( 0.5%) |     3 ( 0.3%)
500 - 1000   |    73 (14.8%) |   101 (17.6%) |   174 (16.3%)
1000 - 1500  |   182 (37.0%) |   252 (43.9%) |   434 (40.7%)
1500 - 2000  |   185 (37.6%) |   161 (28.0%) |   346 (32.5%)
2000 - 2500  |    45 ( 9.1%) |    50 ( 8.7%) |    95 ( 8.9%)
2500 - 3000  |     5 ( 1.0%) |     7 ( 1.2%) |    12 ( 1.1%)
3000 -       |     1 ( 0.2%) |     0 ( 0.0%) |     1 ( 0.1%)

 

Battles      | Allies        | Enemies       | TOTAL
0k - 1k      |   151 (30.7%) |   197 (34.3%) |   348 (32.6%)
1k - 3k      |   156 (31.7%) |   191 (33.3%) |   347 (32.6%)
3k - 5k      |    64 (13.0%) |    76 (13.2%) |   140 (13.1%)
5k - 10k     |    76 (15.4%) |    66 (11.5%) |   142 (13.3%)
10k - 15k    |    23 ( 4.7%) |    24 ( 4.2%) |    47 ( 4.4%)
15k - 25k    |    17 ( 3.5%) |    16 ( 2.8%) |    33 ( 3.1%)
25k - 50k    |     5 ( 1.0%) |     4 ( 0.7%) |     9 ( 0.8%)
50k -        |     0 ( 0.0%) |     0 ( 0.0%) |     0 ( 0.0%)

 

I am planning to run the analysis for newer replays + to different tiers once I have time to fix few bugs there. I also might change the basis of stats to global stats vs cumulative stats. I also try to find a way to do comparison of the player population at different tiers and between now and year ago.

Edited by jylpah, 17 December 2019 - 05:53 AM.

[minitelrose]

The return of Jylpah the great !

[Mjr_Eazy]

Nice one m8, that’s better than my cruddy hack on my fork....

I look forward to it’s push to git

[jylpah]

Mjr_Eazy, on 17 December 2019 - 04:05 AM, said:

Nice one m8, that’s better than my cruddy hack on my fork....

I look forward to it’s push to git

 

Soon, soon. I am using a database setup on my end so now I need to debug the script that it works also without the database setup...

[gallabru]

Nice to see you back here. This is an interesting way of tackling the playerbase distrubution.


If I understand these are the career stats of the players you have seen while playing tbe T54, irrespective of the tier they were playing ?
It is amazing that you saw so many low battle count players. Was it on a reroll account ?

[minitelrose]

gallabru, on 17 December 2019 - 08:50 AM, said:

Nice to see you back here. This is an interesting way of tackling the playerbase distrubution.


If I understand these are the career stats of the players you have seen while playing tbe T54, irrespective of the tier they were playing ?
It is amazing that you saw so many low battle count players. Was it on a reroll account ?

During weekend I played and had many games with ca 5K players

[jylpah]

gallabru, on 17 December 2019 - 09:50 AM, said:

Nice to see you back here. This is an interesting way of tackling the playerbase distrubution.


If I understand these are the career stats of the players you have seen while playing tbe T54, irrespective of the tier they were playing ?
It is amazing that you saw so many low battle count players. Was it on a reroll account ?

 

These are career stats _for the same tier tanks the player played in that battle_. In case of tier IX, the player stats are mix of VIII-X stats. 

 

 

[jylpah]

The replay analysis script has been updated in GitHub (for all those two users :-D)

 

You need to update the blitzutils.py too (just take all updated).

[Mjr_Eazy]

jylpah, on 17 December 2019 - 10:20 PM, said:

The replay analysis script has been updated in GitHub (for all those two users :-D)

 

You need to update the blitzutils.py too (just take all updated).

Nice one m8, I’d have thought you’d have got another user by now but I guess it’s just easier to cry about losses and blame MM

[gallabru]

Thx

[jylpah]

Mjr_Eazy, on 18 December 2019 - 02:04 AM, said:

Nice one m8, I’d have thought you’d have got another user by now but I guess it’s just easier to cry about losses and blame MM

 

I think there are few, but still pretty far from GitHub top 10 downloads

[szkaraj]

jylpah, on 18 December 2019 - 05:36 PM, said:

 

I think there are few, but still pretty far from GitHub top 10 downloads

Hy mate,

How does this GitHub work? I download, install and run while i play, or run it after games? Does it send you data automatickly?

[gallabru]

It is a tool for analysing your replays.
You have to save replays.
You send them as a baych (you can select all) to replays@wotbinspector.com from your ingame interface.
You will receive the replays in your email box
You download them in a folder on your computer and run the python scripts

[szkaraj]

gallabru, on 22 December 2019 - 01:06 PM, said:

It is a tool for analysing your replays.
You have to save replays.
You send them as a baych (you can select all) to replays@wotbinspector.com from your ingame interface.
You will receive the replays in your email box
You download them in a folder on your computer and run the python scripts

Ok. And then I send results to jylpah?

[jylpah]

szkaraj, on 22 December 2019 - 11:49 PM, said:

Ok. And then I send results to jylpah?

 

No please :-)

 

How the scripts (analyze_wotb_replays.py) works?

 

1. First you need to transfer / copy the replay files from your gaming device to a computer. I use iPad and iCloud Drive to transfer the replays to my computer.
2. Then follow the instructions here: https://github.com/Jylpah/blitz-tools/blob/master/README.md

 

Remember, before you can analyze the replays, those needs to be posted to WoTinspector.com using the replay uploader since the analysis script does NOT read the replay files itself, but the JSON files returned by WoTinspector.com.

[jylpah]

Just note to the 3+1 (Szkaraj?) users of the scripts. I recently made a major update and integrated cache database (Sqlite3) into the script. This makes repeated analyses far faster since the script can read cached stats from local database instead of fetching those from WG API.

 

If you have use the blitzutils.py in your own scripts, please note that all the function names have been changed according to Python coding style guide and that there new nice helper functions.

[szkaraj]

jylpah, on 25 December 2019 - 09:35 PM, said:

Just note to the 3+1 (Szkaraj?) users of the scripts. I recently made a major update and integrated cache database (Sqlite3) into the script. This makes repeated analyses far faster since the script can read cached stats from local database instead of fetching those from WG API.

 

If you have use the blitzutils.py in your own scripts, please note that all the function names have been changed according to Python coding style guide and that there new nice helper functions.

Oki, thx. Maybe i give it a try.