Leadership, Research, HQs and AI files.
Hi All,
I have been quietly pottering away working on these AIs
Making progress
Now came to this:
Code:
leader_ratio = {
land_field_marshal =0.083
land_general =0.166
land_lt_general =0.333
sea_grand_admiral =0.083
sea_admiral =0.166
sea_vice_admiral =0.333
air_marshal =0.083
air_general =0.166
air_lt_General =0.333
}
And got a bit stuck – what were the correct values for each nation?
Then hit upon the idea of using the TGW leadership files to deduce these ratios
Have done that now and was able to generate a spreadsheet which has the ‘correct’ leadership ratios for each nation.
So, for example
Code:
British Leaders Number Ratio
land_field_marshal 1 .003
land_general 16 .052
land_lt_general 33 .108
mj gens 99 # Does not matter
sea_grand_admiral 0 .000
sea_admiral 14 .046
sea_vice_admiral 11 .036
rear admirals 76 # Does not matter
air_marshal 0 .000
air_general 0 .000
air_lt_General 0 .000
air mj gens 56 # Does not matter
Now, do you think this is a fair way to do this?
My big concern is that almost no nation has any ranking Air Generals. What do you think?
How does this part of the AI interact with the “Historical Promotions” in the leader files? Anyone know?
Now, the next part was figuring out how many HQs each nation should build.
The problem here is that in order to build HQs you have to build proportionately less of something else (or maybe everything else) and there is a big IC cost associated with HQs (maybe we will change that?).
So I decided to try figuring out what proportion of Army Officers were of Field Marshal or General Rank and build that proportion of HQs.
I got some interesting results.
Based on that
Code:
HQs as a proportion of Army units?
British 11%
Austria Hungary 20%
Germany 24%
France 19%
Russia 28%
Now that seems a little top heavy! But is it?
Does this represent the overall bureaucracy in the system?
HQs are not just HQs. It represents the whole chain of logistics and Command IMHO.
After all, that would mean that the AI constantly wasted resources on producing ‘HQs’ – but the ESE at the front would be excellent!
It would certainly slow the war down – and it might lead to a stalemate on the western front perhaps?
I am interested in your opinions. My initial thought was to ‘cap’ the number of HQs at 5% What do you think?
Next came this:
Code:
technology = {
endgoal = { }
preference = { }
ignore = { }
armor = 0
infantry = 0
industry = 0
aircraft = 0
naval = 0
land_doctrines = 0
secret_weapons = 0
air_doctrines = 0
naval_doctrines = 0
}
It occurred to me here that this was a political thing and again I could use the leader lists.
The force with the most leaders would have the most political clout and would get the biggest research budget?
Fair assumption?
Then I further divided up this based on the assumption that high ranking leaders (F.Ms and Generals) would be more interested in Doctrines while lower ranking Generals (Lt Gens and Mj Gens) would be more interested in development.
I weighted the results to reflect the fact that Higher Ranking officers tend to get their way more often too.
So, for Britain I get this:
Code:
armor = 0 # Unknown. Different AI to execute Plan 1919?
infantry = 0.40
industry = 0 # To be decided
aircraft = 0.14
naval = 0.24
land_doctrines = 0.13
secret_weapons = 0 # To be decided
air_doctrines = 0.00
naval_doctrines = 0.10
So? Fair Approach?
The problem here is that as almost no nation has any ranking air officers almost no-one does any Air Doctrine research. Will have to tinker with that.
Now obviously, all these values can be changed when the AIs are tested – but as a starting approach to the problem – what do you think?
Obviously we can not discuss every value to the nth degree so is this a good blanket method?
Anyway, while you all discuss this I’m off to start recording target provinces for the various nations and AIs.
