Jump to content

TGIF: scenario list, discussion, and house rules


Volcano
 Share

Recommended Posts

Can someone please send me the AAR and report .htm from the game when it is finished?  I would like to use it for a project I'm working on, and will also share some of those results with the group as I glean them.  Thank you!

Link to comment
Share on other sites

Following up, I would like to sanity check some statistical analysis I am doing on the report files, so any feedback here will be valuable, many thanks in advance.  As an aside, the Mongoose report looks weird... was that maybe a game that crashed, or is there a reason a table wouldn't be generated for what unit killed which?

 

Regardless, I have these stats for the "Jackal" game report, I'm interested if the participants think these are accurate.  These are generated by pulling the table from the report that attributes who shot who, filtered for the column "Destroyed" having an "X," and broken down into three target (what was destroyed, not whom) categories -- Tanks, PCs, and personnel.

 

For those not familiar with summary statistics, the count variable is how many times an event (in this case, hitting a vehicle with Damage marked as "X") happened, and the other numbers are range in meters.

"mean" is the average

min - smallest range in m

max - largest range in m

std - standard deviation (a measure of the spread from the mean)

...and the others are quartiles, to give a sense of the grouping.

 

Statistics for Report from:  .\reports\Brush Fire-Jackal-NG-4264.sce_8_06-25-21_22_47_58.htm

Summary statistics on Kill ranges in meters for Tank targets:
count      51.000000
mean     1357.725490
std       975.547704
min        73.000000
25%       705.000000
50%       933.000000
75%      2326.000000
max      3862.000000

 

Summary statistics on Kill ranges in meters for PC targets:
count      17.000000
mean      714.470588
std       419.394224
min        42.000000
25%       505.000000
50%       773.000000
75%       833.000000
max      1905.000000

 

Summary statistics on Kill ranges in meters for personnel targets:
count      155.000000
mean       543.451613
std       1209.463502
min         35.000000
25%        119.500000
50%        278.000000
75%        482.500000
max      10421.000000
 

Are statistics like this interesting?  What else might be interesting, e.g., what vehicle/unit made the max kill range?  Who had the most kills?  Anything else?  And am I reinventing the wheel and overlooking more obvious ways to glean this information?  Inquiring minds want to know...

 

 

Link to comment
Share on other sites

GG on Friday.

 

So here are the stats from that day, with some future details coming.  Based on feedback from Dark, I filtered out artillery hits/kills, because they were creating massive outliers (think 12km+) and moving the means too much.

 

Is it interesting to highlight (here) things like who had the longest distance kill, and in what platform?  It might also help me sanity check accuracy of these figures given extra eyes and egos involved.

 

Statistics for Report from:  .\reports\Civil War Mykonos-4265.sce_12_07-02-21_19_16_12.htm without artillery kills

Summary statistics on Kill ranges in meters for Tank targets:
count     158.000000
mean      477.107595
std       621.043293
min        36.000000
25%       232.750000
50%       306.000000
75%       306.000000
max      3211.000000
 

Summary statistics on Kill ranges in meters for PC targets:
count     105.000000
mean      582.333333
std       756.883878
min        85.000000
25%       162.000000
50%       205.000000
75%       745.000000
max      3699.000000
 

Summary statistics on Kill ranges in meters for personnel targets:
count     216.000000
mean      343.226852
std       414.587404
min         3.000000
25%        73.000000
50%       209.000000
75%       322.000000
max      2150.000000
 

Link to comment
Share on other sites

27 minutes ago, Valleyboy said:

So here are the stats from that day, with some future details coming.  Based on feedback from Dark, I filtered out artillery hits/kills, because they were creating massive outliers (think 12km+) and moving the means too much.

 

Think it may be interesting to narrow things down further. So for example with Kills on tanks he may be interesting to see this broken down further into "what" caused the kill ie Infantry (rpg), Infantry ATGM, Tank, APC etc

 

Also you may want to check the totals... I'm not convinced there were 158 kills in that mission (I could be wrong). Are you sure you aren't getting hits on things which are already dead?.

Link to comment
Share on other sites

1 hour ago, DarkAngel said:

Think it may be interesting to narrow things down further. So for example with Kills on tanks he may be interesting to see this broken down further into "what" caused the kill ie Infantry (rpg), Infantry ATGM, Tank, APC etc

 

Also you may want to check the totals... I'm not convinced there were 158 kills in that mission (I could be wrong). Are you sure you aren't getting hits on things which are already dead?.

I think you hit it on the nose, I think it for sure is getting "non-kill" impacts on killed vehicles mixed in and confusing the numbers.  So there's an initial challenge (and thank you for that first sanity-check observation), is there a way in that table to filter for a specific metric to indicate whether it is the actual "kill hit," or is the solution to try to sort chronologically and then disqualify all hits after that on that specific entity?  E.g., "T-72 1/1" gets hit with a X on Destroyed for X event, and then after that, all other instances are ignored based on timestamp?

Link to comment
Share on other sites

51 minutes ago, Valleyboy said:

is there a way in that table to filter for a specific metric to indicate whether it is the actual "kill hit," or is the solution to try to sort chronologically and then disqualify all hits after that on that specific entity?  E.g., "T-72 1/1" gets hit with a X on Destroyed for X event, and then after that, all other instances are ignored based on timestamp?

 

You can open it in Excel (even though its HTML) and then sort / extract that way.

 

That's how our OPs Analysis people do it.

Link to comment
Share on other sites

Is possible to filter the results not by unit type but by the specific unit model?. Something like the number of kills/losses between 2 different units like m1 and t-72 can be very helpful for balancing missions.

Link to comment
Share on other sites

49 minutes ago, Colebrook said:

Is possible to filter the results not by unit type but by the specific unit model?. Something like the number of kills/losses between 2 different units like m1 and t-72 can be very helpful for balancing missions.

Yes, I think so, and that is good feedback.  I need to solve the "What was the actual kill" chronology problem, but I'm assuming right now that the first time something is marked as "Destroyed," that's probably where I need to deem that unit dead and disqualify all other hits afterwards for the file.  Makes for some filtering challenges, but its probably useful code for later in the project anyways.

Link to comment
Share on other sites

9 hours ago, Gibsonm said:

 

You can open it in Excel (even though its HTML) and then sort / extract that way.

 

That's how our OPs Analysis people do it.

Thanks, that is good feedback on how the users engage with the data.  Our hope with this project is to help make it more error-free, standardized, faster, and allow batch processing on large amounts of files in the case where multiple cohorts or classes need to be compared, as one of the main use cases.

Link to comment
Share on other sites

Ok, after some sweat and tears (and Python self-re-education), I have the below, for non-artillery, but not looked at critically yet.  Need some rest before tackling more, as I was still seeing unexpected artillery kills in the other version that I need to sanity-check:
 

Statistics for Report from:  .\reports\Civil War Mykonos-4265.sce_12_07-02-21_19_16_12.htm without artillery kills

Summary statistics on Kill ranges in meters for Tank targets:
count      25.00000
mean     1160.84000
std      1000.68292
min        36.00000
25%       225.00000
50%      1315.00000
75%      2106.00000
max      3211.00000
 

Summary statistics on Kill ranges in meters for PC targets:
count      31.000000
mean      994.645161
std      1001.865512
min        85.000000
25%       174.500000
50%       664.000000
75%      1865.500000
max      3699.000000
 

Summary statistics on Kill ranges in meters for personnel targets:
count      66.000000
mean      346.833333
std       446.305366
min         3.000000
25%        73.500000
50%       187.000000
75%       311.500000
max      2150.000000
 

Link to comment
Share on other sites

Posted (edited)

For Tank targets, max kill user and type:

(Shooter, Unit)    (Shooter, Type)    (Target, Target distance)
3/1/Tank Co (Blue) (pancho)    Tank (M60A3(TTS))    3211.0

 

For PC targets, max kill user and type:

(Shooter, Unit)    (Shooter, Type)    (Target, Target distance)
1/3/Mech Co (Red) (-DK-)    PC (M901 ITV)    3699.0

 

Still working on how to generate these in a more user-friendly format with less manual programming.  Ideally this would be part of an export file and also aggregable.

 

Edit: Added on the actual range

Edited by Valleyboy
Link to comment
Share on other sites

42 minutes ago, DarkAngel said:

I t might be nice also to aggregate PC carried infantry kills into the PC kill if it died with it's infantry on board. 

For those, would they have the same time-stamp in terms of death, game-wise, as the vehicle they were in?

Link to comment
Share on other sites

So, to keep things fresh, I have some new stats to try to tell a story about this past game.  Note that the below are now filtered to exclude artillery and mine kills:

 

Unit and type for max range Tank type kill:
    (Shooter, Unit)      (Shooter, Type)  (Target, Target distance)
4/1/A (Red)  Tank (Challenger 2)                     3355.0
Unit and type for max range PC type kill:
           (Shooter, Unit)      (Shooter, Type)  (Target, Target distance)
2/1/A (Red) (Volcano)  Tank (Challenger 2)                     3422.0

 

Tanks were killed by these unit types:
Tank (Challenger 2)         14
Tank (T-72B1 m.1985)        12
Tank (T-72M m.1980)         11
personnel (Missile team)     8
PC (CV90/30-FI)              7
Tank (Piranha DF90)          5
Tank (T-62 m.1972)           3
personnel (Rifle squad)      1

PCs were killed by these unit types:
PC (CV90/30-FI)             7
personnel (Rifle squad)     7
personnel (Missile team)    5
Tank (T-62 m.1972)          3
Tank (T-72B1 m.1985)        2
Tank (Challenger 2)         2
Tank (Piranha DF90)         2
personnel (LMG team)        1
 

 

Link to comment
Share on other sites

  • Members

16 JUL 2021:

 

TGIF Watan Saraf Ekhlas V9-smaller_4265

We will try this again, we ended up playing a different scenario for various reasons.

 

SPECIAL CONSIDERATIONS:

  • Draft? Yes.
  • Random CO selection? Yes.
  • Minimum # players: 12 (Red (Insurgents) [5] vs Blue (Registani Army) [4] & Green (NDF) [3])

NOTES:

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...