AI PROJECT: 4th July 2003: Results discussion

Maximilian I · Jul 4, 2003

please post your opinions, obeservations and comments about the latest ai file here.

thanks

Vulture · Jul 4, 2003

My observations (1939 and 1941)

1939 Italy - Italy is a tad bit too good. Might be more a scenario setup problem though. Italy beats back the UK in Northern Africa consequently. UK doesn't exist anymore from Egypt to the Horn. Solution: Tone down Italians a bit or introduce a passivity command in the AI file
1939 Germany - Perfectly executed series of attacks. Poland is overrun in little over a month. Luxemburg, Belgium and Netherlands follow in short succession. It takes a little longer to make it into France, but eventually they give France the steamroll treatment and set up Vichy. Afterwards Germany executes a perfect Weserübung: Denmark and Norway are annexed. No tweaks needed
1939 Canada - Makes a series of assaults on Southern France in which they take about 6 Vichy provinces. Good but timing is off. Solution: Maybe make a third AI file for the UK, SAF, CAN, NZL and AST with invasion commands and exclude invasion commands from the default War.ai. Trigger this AI file after the war is on for a while. AI should build up forces first
1941 Germany - Has a harder time in the USSR. Front is lucid for little less than a year. USSR and Germany swap provinces around. Eventually they did take Leningrad and Moscow and the Russian front slowly collapsed. No more tweaks needed. Strong AI
1941 USSR - It succeeds in holding out nicely for a year or so and makes attempts to counterattack, even taking 8 provinces in Southern Russia and Northern Romania. Eventually they lose Leningrad and Moscow and are pushed back to the borders of Stalingrad. Solution: Russian Defensive AI should be strenghtened a bit. Defense in dept? More reserves and definately more troops garrisoned in Moscow and Leningrad. It also needs passivity commands towards Manchukwo
1941 UK - Pushes back the Italians quite fast in Northern Africa, which is good. They however struggle very hard and lose too much troops in the assault on Benghazi, even tough Benghazi is surrounded. Solution: Reduce fort in Benghazi and make the UK AI put more ships to blockading Benghazi
1941 Japan - I saved the best till last. The Japan_sea.ai is really good at the moment. Probably the best AI when it comes to invasions. The front in China is lucidly stalled as it is supposed to be. Japan slowly takes more ground. Very slowly. They however take Mongolia and advance into Russia. Solution: higher passivity towards Russia and introduce passivity command towards Mongolia. Japan executes very good invasions. They invaded and took in successive order Wake, the Phillipines, Hong Kong, Malayan Peninsula with Singapore, the Solomon Islands (!!!) and big chunks of Indonesia. Thumbs up to the Japanese invasions. Minor work needs to be done on their garrisoning. Currently they put too little troops in Phillipines which sometimes causes pushbacks and even surrounding. In the end they invade with a bigger army though. Also make it focus more on Singapore. Currently it invades Kota Bahura and sits there for a long while. Solution: Area multiplier and province priority for Manila. Area multiplier and province priority for Kota Bahura and Singapore. Bigger focus on Singapore

unmerged(14753) · Jul 5, 2003

"1939 Italy - Italy is a tad bit too good. Might be more a scenario setup problem though. Italy beats back the UK in Northern Africa consequently. UK doesn't exist anymore from Egypt to the Horn. Solution: Tone down Italians a bit or introduce a passivity command in the AI file"

No please don't do this yet. Here's my uk.ai file which stops this problem while not exposing the main island (Britain) to invasion by Germany too much (in at least 8 handsoff games Germany hasn't been able to successfully invade England)

#
# file: uk.ai
#
# Hearts of Iron
#
# By Trip (modified by Roxxor)
#
# Standard UK AI file pre-rebound
#

# 100 = Total warmonger, 0 = absolute pacifist
war = 50

max_garrison_prop = 0.6
min_garrison_prop = 0.6
strat_redeploy_threshold = 15

influence = {
AFG BEL BHU CHC DEN FIN GRE HOL IRQ LUX NEP NOR OMN
PER POL ROM SAU SPA SPR SWE TIB TUR USA VEN YEM YUG
}

construction = {

max_factor = 0.3

max_AA = 4
AA_batteries = atwar
max_AA_level = 10

coastal_fort = atwar
max_coastal_level = 5

ic_at_war = no
ic_end_year = 1946
}

military = {

relative_build_scheme = no

infantry = 36
cavalry = 0
motorized = 16
mechanized = 0
panzer = 2
paratrooper = 0
marine = 0
bergsjaeger = 0
# 54 %

fighter = 4
strategic_bomber = 4
tactical_bomber = 4
naval_bomber = 4
dive_bomber = 2
transport_plane = 0
torpedo_plane = 2
# 20 %

battleship = 5
carrier = 2
destroyer = 6
cruiser = 6
submarine = 2
transports = 5
# 26 %

}

technology = {

endgoal = {}
preference = {}

electronics = 3.0
armor = 1.0
infantry = 2.0
industry = 4.0
rocket = 0.2
nuclear = 0.2
heavy_aircraft = 2.0
light_aircraft = 3.0
naval = 2.0
submarine = 0.2
artillery = 1.0
naval_doctrine = 2.0
land_doctrine = 0.1
air_doctrine = 2.0

}

garrison = {

# OVERSEAS DISTRIBUTION
home_peace_cap = 25 # When we are at peace, 8 divisions in England should suffice.
defend_overseas_beaches = yes
area_multiplier = {
989 = 2.0 # Gibraltar
1100 = 2.0 # Port Said

# Some really low prio areas:
168 = 0
225 = 0
233 = 0
260 = 0
268 = 0
270 = 0
272 = 0
274 = 0
275 = 0
276 = 0
308 = 0
309 = 0
414 = 0
461 = 0
469 = 0
1035 = 0
1057 = 0
1283 = 0
1310 = 0
1331 = 0
1332 = 0
1340 = 0
1348 = 0
}

# PRIORITIES:
beach_1 = 61 # Beach level 1
beach_2 = 85 # Beach level 2
beach_3 = 99 # Beach level 3
capital = 150 # Our capital
human_border = 100 # Ignored for allied human players
war_target = 100 # The next country targeted for attack by this AI

opposing_alliance = 100 # For neutrals, all alliances are "opposing"
claim_threat = 100 # If we are neutral, countries with claims on us get this
unstable_border = 50 # Bordering countries at war with someone else get this

key_point_prio_mult = 0

# Specific provinces
province_priorities = {
498 = 200 # Coventry
501 = 200 # Plymouth
502 = 200 # Southampton
503 = 200 # Portsmouth
1091 = 100
1099 = 100
1100 = 100
}
}

front = {
recklessness = 0 # 0-3
enemy_reinf_days = 6
reserve_prop = 0.0
panic_ratio_vs_ai = 1.000
panic_ratio_vs_human = 1.000
base_attack_odds = 3.000
min_attack_odds = 2.500
max_attack_odds = 4.000
reinforce_odds = 1.000
withdraw_odds = 0.8
distrib_vs_ai = even
distrib_vs_human = reactive
}

bomber = {

#defensive fighter ai factors

defensiveIC = 5.0
recentlyBombed = 7.0 #protect anything bombed.
protectOil = 3.0
protectCoal = 2.0
protectSteel = 2.0
protectRubber = 1.0

#Evaluation factors for determening which landarea to be strategically bombarded.

areaSize = 2.0 #size of area.
areaDistance = 1.0 #distance
adjacency = 1.0 #if its neighbor to us.

#Evaluate which province to get a bombing mission.

targetIC = 5.0 #How important are IC to target..
avoidAA = 1.0 #how bad AA are.
targetCoal = 2.0
targetSteel = 2.0
targetOil = 4.0
targetRubber = 2.0
occupied = 15.0 #Subtract for provinces which are occupied, ie bomb germany ahead of france as UK.

}

admiral = {
target = {
504 = 1001
510 = 1001
516 = 1001
591 = 1001
593 = 1001
597 = 1001
888 = 1001
892 = 1001
894 = 1001
899 = 1001
904 = 1001
909 = 1001
1008 = 1001
1010 = 1001
}
}

Maximilian I · Jul 5, 2003

I agree with Roxxor here. the UK AI has not yet been improved and why should we prune the working italian ai already after the first try?

Vulture · Jul 5, 2003

Okido... I'm too lazy to check all the differences between the normal UK.ai and yours Roxxor. Care to explain the differences? And where they are?

unmerged(5627) · Jul 6, 2003

In my single observation game so far, Japan has had a very easy time in China, quickly pushing them back into central China and only getting slowed down around those provinces north of Vietnam and that border Sinkiang. I'm not sure why Japan had such an easy time, but I could see that the Nationalist Chinese were putting too many troops in reserve on garrison duty. Garrison units wouldn't attempt to advance against the enemy or to support troops in battle unless Japanese units were right next to them.

Germany invaded Poland fine, same with France, moving through Belgium and the Netherlands to reach Northern France. After Vichyification, Germany, Italy and the rest of the Axis allies dow'ed Greece while the UK installed a puppet government in Yugoslavia.

Meanwhile, in Finland the Soviets seemed to be having a harder than expected time. The Finns didn't provide much of a defense, but the Soviet units just didn't attack. Even when it looked like they had the advantage in numbers by alot they wouldn't attack and they only slowly advanced into Finland. They also only attacked with small armies while having larger armies seemingly roam about. Perhaps they had low org? It just seems like that Soviet invasion was sloppy, although they did end up annexing Finland after a long battle in Turku where 4 Soviet Units battled 6 remaining Finish divisions, all while some 30+ odd Soviet divisions directly adjacent to the province just sat there.

The game crashed shortly thereafter, possibly from a Romanian annexation of Yugoslavia.

unmerged(10750) · Jul 7, 2003

One quick observation -

Haven't been able to run any tests on 1.05 yet, but there is a large discrepancy in the human_border AI preference.

Some nations have it set to zero, some at 150, some at 210.

While this makes no difference in hands-off games, obviously, it could lead to some erratic AI behavior when one is actually playing. In more than one case, the human_border variable changes when a country switches AIs.

Perhaps a discussion is needed as to the optimal setting of that variable. I like to set it at zero - I find the AI actually performs better when they're not trying to pile a bunch of units on my border. Of course, I try not to use exploits or gamey tactics...

Vulture · Jul 7, 2003

Originally posted by czolgosz
One quick observation -

Haven't been able to run any tests on 1.05 yet, but there is a large discrepancy in the human_border AI preference.

Some nations have it set to zero, some at 150, some at 210.

While this makes no difference in hands-off games, obviously, it could lead to some erratic AI behavior when one is actually playing. In more than one case, the human_border variable changes when a country switches AIs.

Perhaps a discussion is needed as to the optimal setting of that variable. I like to set it at zero - I find the AI actually performs better when they're not trying to pile a bunch of units on my border. Of course, I try not to use exploits or gamey tactics...

Open a thread for it maybe

unmerged(14753) · Jul 8, 2003

I put my discussion of the above uk.ai file in the UK and commonwealth thread.

unmerged(14393) · Jul 8, 2003

Originally posted by czolgosz
One quick observation -

Haven't been able to run any tests on 1.05 yet, but there is a large discrepancy in the human_border AI preference.

Some nations have it set to zero, some at 150, some at 210.

While this makes no difference in hands-off games, obviously, it could lead to some erratic AI behavior when one is actually playing. In more than one case, the human_border variable changes when a country switches AIs.

Perhaps a discussion is needed as to the optimal setting of that variable. I like to set it at zero - I find the AI actually performs better when they're not trying to pile a bunch of units on my border. Of course, I try not to use exploits or gamey tactics...

This is a very good point. If you play for example Italy you can turn the tide of the whole Eastern front by just placing your troops at the Romanian border and all of the sudden the Russians deploy extremely many troops where your troops are, leaving other areas virtually undefended.

Maximilian I · Jul 8, 2003

Originally posted by czolgosz
One quick observation -

Haven't been able to run any tests on 1.05 yet, but there is a large discrepancy in the human_border AI preference.

Some nations have it set to zero, some at 150, some at 210.

yeah, it was changed for major nations only.

unmerged(14856) · Jul 8, 2003

Comments to the AI files 0407:

czech.ai
One of the most underestimated countries in the european theatre. In the '36 scenario they start with the 9th largest army in the world!
There should be 3 czech AIs: peacetime/joined_axis/joined_allies
Since we know that the czech.ai will face problems with production due to lack of rubber and oil the file should reflect this.
Producing arm/mot/mec is just stupid when the AI can't trade oil if their life depended on it.
There where still NAVAL techs in the tech list
Sending a proposition to Vulture *click*
Minors
Are building IC far too late, they should stop in 1941 at the latest...
Are building marines???
Wastes IC on building armor and researching armor techs...
Are wasting IC on building NAVAL techs which they will never get a chance to use...
Are wasting IC on electrical/industrial research, put it in as preferences instead NEVER do it randomly.
Many nations like Ireland/Iraq/etc have 0 in building armor but still have values for researching armor tech...
Many nations which have no domestic oil production are building mec/mot/arm which will cripple their countries IC by stealing the Oil needed for conversion to rubber.
This means that ALL of them will die if they end up in a war (No world trade).
There should be a preference list of techs that only targets INF/CAV so that it makes sense for smaller countries.

IF you want more just say so...

Exel · Jul 9, 2003

Originally posted by Spoonist
IF you want more just say so...

Keep it coming.

Maximilian I · Jul 9, 2003

Originally posted by Exel
Keep it coming.

seconded.
keep it coming. any input = good.

mog · Jul 10, 2003

Hello.

So far I've only used a few of these changes, mainly the building preferences. I've seen the USSR, for example, build a lot of bombers of varying types, in quantity more than my Luftwaffe. However, the problem that I see with the ai is that it only uses its air force in dribs and drabs - 2 squadrons here, 3 there, or even just singly. If they could be programmed to use them in larger groups, say minimum of 6 with fighter cover, I'm sure that they would be far more effective.

Btw, I witnessed a USA invasion of Vichy North Africa last night - over 40 divisions and 80+ bombers lol. But because they fritter away their air force in small numbers, I've beaten the invasion off. But it was a big and pleasant surprise to see it

unmerged(16695) · Jul 10, 2003

I ran the 0407 ai pack last night, starting in 1939, on normal/normal.

Germany steamrolled poland, annexing them on Sept 30. in november, they stormed through BeNeLux and got the vichy event in early spring, 1940. the belgians did retake antwerp briefly, not sure where those guys came from, but were quickly dispatched again.

the most surprising thing was germany started barbarosa in june, 1940, and romania joined the axis shortly thereafter.

the SU immediately took Lwow, and that has changed hands 3-4 times now. germany is making very slow advances through the ukraine and the baltic states. i had to stop and go to bed, but as of Nov 1940, the lines were pretty stable on the russian steppes.

it's interesting to see how aggressive german ai has been this game. no one else in the world seems to be fighting at this point- nothing moved in africa, the pacific is as calm as it's namesake.

unmerged(16020) · Jul 10, 2003

My results posted on the minor thread on some of South America:

The AI project Argentine AI is odd. It does keep 1 unit in one of the low infra provinces along the Chilean border, as well as a tank unit in San Sebastion, with the rest of the units being in Buenos Aires and none in Mar Del Plata.

AI Project Chilean AI has their economy collapse quickly, and for some reason starts building coastal defenses and anti-aircraft guns. Chilean AI needs to be constantly improving all of their provinces, and to even keep their 40 to 25 IC drop up, they'll need a few steel to coal trades.

Bolivian AI (Default one I believe) improves IC, which is the worse thing possible to do to that poor country. Drops them from 19 to 7 IC if you start improving IC since their economy can support exactly 18, not even 0.1 more.

Brazil AI spends over half their troops to guard the low infra provinces next to the "opposing alliance" in French and British Guyana. It would accomplish much more to instead just focus on the beaches, which would guard the only real vulnerable place should the French decide to attack out of Guyana, not to mention guarding more important areas.

unmerged(14856) · Jul 11, 2003

Originally posted by Maximilian I
seconded.
keep it coming. any input = good.

Here are some more comments:

Most smaller garrison AIs have the key_point_prio_mult set to 0. This should be _TOP_ priority when you in the game will get annexed if that province falls.
This will also lead to historically correct behavior. The minors will fight decisively in what they regard as their most valued areas.
All minor nations should have this set to
key_point_prio_mult = 201
Please vulture?
The front AIs enemy_reinf_days value seem to be set to the default regardless of the size of the nation.
For countries like USSR, China, Canada, Brazil, etc the 6-8 days are worthless, they must have values of 20-40 to be able to make the counter move on time.

romania.ai
Romania has a domestic oil supply. They should use this and build motorized troops.
Again they are building
panzer = 5
but are only giving armor tech a low priority
armor = 0.1
it would be better if those armor would be motorized inf.
Also 10% cavalry is too large a number for ROM.
The Construction ai is only allowed to build aa in one province at a time, change this to 3.

The admiral ai has orders to patrol the whole black sea which will lead to the annihilation of the navy, this should be limited to their own coast only.
870 = 200 # West Black Sea
867 = 200 # Central Black Sea

Suggestion:

Code:

military = {
	
	relative_build_scheme 	= no
	
	infantry = 		42
	cavalry =		5
	motorized = 		30
	mechanized =		2
	panzer = 		2	#Unless given tech their tanks will be useless
	paratrooper = 		0
	marine = 		0
	bergsjaeger = 		2	#ROM starts with MTN so they should build some
#					83 %
	fighter = 		6
	strategic_bomber = 	0
	tactical_bomber = 	0
	naval_bomber = 		0
	dive_bomber = 		5	#instead of TAC for low tech
	transport_plane = 	0 
	torpedo_plane = 	2
#					13 %
	battleship = 		0
	carrier = 		0
	destroyer = 		3	#ROMs navy was more a coast guard
	cruiser = 		0	
	submarine = 		0
	transports = 		1	#Who would they invade?
#					4 %
}

technology = {

	endgoal = {}
	preference = {}

	electronics = 		0.001	#Minimum	
	armor = 		0.1
	infantry = 		2.0
	industry = 		1.0	#Lots of godies early
	rocket = 		0	#No
	nuclear = 		0	#No
	heavy_aircraft =	0.001	#Minimum
	light_aircraft =	0.001	#Minimum
	naval = 		0.001	#Minimum
	submarine = 		0.001	#Minimum
	artillery = 		3.0	#Here is where all the bonuses are for INF and MOT
	naval_doctrine =	0.001	#Minimum
	land_doctrine = 	0.1
	air_doctrine = 		0.001	#Minimum
}

Canada_Peace.ai

Why would a peaceful Canada build an airforce? They don't even have the range to attack between the provinces in Canada...

Don't build MOT/ARM/MEC when at peace if you don't have a domestic oilsource, you'll only hurt production.

suggestions:

Code:

strat_redeploy_threshold 	= 30	#The provinces in Canada are HUGE, and in the front they should never redeploy
construction = {	#Keep the others as they are these two I'd like to change	
max_factor			= 0.25	#Four provinces of 16 need upgrading
ic_at_war			= no	#Why would they try to build IC at war when they don't have the oil?
}
technology = {

	endgoal = {}
	preference = {}

	electronics = 		0.001	#Without preferences this is useless, if you set preferences then you can increase this again
	armor = 		0.001	#Minimum they are still at peace
	infantry = 		2.0
	industry = 		4.0	#They need the conversion techs to keep up production
	rocket = 		0
	nuclear = 		0
	heavy_aircraft = 	0.001	#Minimum
	light_aircraft = 	0.001	#Minimum they are still at peace
	naval = 		0.001	#Minimum not enough return on investment 
	submarine = 		0.001	#Minimum
	artillery = 		2.0	
	naval_doctrine = 	0.5	#2.0 is way too high it's better to have low org & run then to fight badly
	land_doctrine = 	1.0
	air_doctrine = 		0.001	#Minimum
}

Canada_War.ai

The airforce should only be defensive fighters, they will not be able to deploy their force to other theatres anyway. Remove the bombers.
Put this into destroyers instead, something which they will need.

I still think that since Canada doesn't have an oil supply they are building too much MOT/MEC/ARM.
IT's better to build INF with the oil going to full IC production.

The garrison ai should defend the overseas beaches.

The invasion ai should also have british beaches as targets. This means that if the germans have done sealion Canada will try to retake it.
Increase the 'enemy' value to at least 4, and adjecentenemy to 2, it's an invasion of germany we are talking about.

suggestions:

Code:

strat_redeploy_threshold 	= 30	#The provinces in Canada are HUGE, and in the front they should never redeploy
max_garrison_prop		= 0.1	#Canada should trust in the aid of USA
construction = {	
	max_factor			= 0	#Canada doesn't have the oil for both increased IC and troops
	ic_end_year			= 1939
}
technology = {

	endgoal = {}
	preference = {}

	electronics = 		0.001	#Without preferences this is useless, if you set preferences then you can increase this again
	armor = 		0.001	#They don't have the IC to spare, they should rely on people sending tech
	infantry = 		3.0
	industry = 		1.0
	rocket = 		0
	nuclear = 		0
	heavy_aircraft = 	0.001	#Minimum
	light_aircraft = 	0.001	#Minimum they are still at peace
	naval = 		0.001	#Minimum since you are not building any naval forces
	submarine = 		0.001	#Minimum
	artillery = 		3.0	#This is where most of the bonuses for INF is.
	naval_doctrine = 	0.001	#Minimum since you are not building any naval forces
	land_doctrine = 	1.0
	air_doctrine = 		0.001	#Minimum
} 
	country_priorities = {
			USA = -1 #add this line so that they wont guard the border with the US
			}

Yugoslavia.ai

The military ai should at least build 1% destroyers and 1% submarines to get the historical feel to it, and shouldn't build so many MOT since they don't have the oil to support them during wartime.

The garrison ai is just crazy. Since it is a relative scale it's pointless to set all priorities to 50, you could equally set all to 0. Instead number them in the order you wish the ai to make the priority, like 50-51-52-53-etc. In country_priorities it doesn't help to write in every neighbour, stick to one or two as a maximum. The value to tweak here is the unstable_border value which should have top priority. Also why is the province_priority for beograd set to 0 (???) since the capitol value is in the list of priorities this is already taken care of...

suggestions:

Code:

influence = { ITA AUS GRE HUN ROM BUL }	#Italy and Bulgaria was missing and austria was in twice

construction = {
	max_factor		= 0.7
	max_AA		= 3	#So that they actually can achieve some airdef
	AA_batteries 	= yes
	max_AA_level	= 2
	coastal_fort	= yes
	max_coastal_level = 2
	max_land_level	= 2
	fort_borders = { GER }	#Will only happen after anschluss didnt include italy beacause of historic ties
	ic_at_war		= no	#should concentrate on building troops
	ic_end_year		= 1940	#Can't support too many IC
	force_ic_until	= 1937
}
technology = {

	endgoal = {}
	preference = {}

	electronics = 		0.001	#Without preferences this is useless, if you set preferences then you can increase this again
	armor = 		0	#since they don't build them
	infantry = 		4.0
	industry = 		0.001	#Minimum will never have the IC to get going
	rocket = 		0
	nuclear = 		0
	heavy_aircraft = 	0.001	#Minimum new models are too far away
	light_aircraft = 	0.001	#Minimum new models are too far away
	naval = 		0.001	#Minimum since you are not building any naval forces
	submarine = 		0.001	#Minimum
	artillery = 		3.0	#To get the INF soft+
	naval_doctrine = 	0.001	#Minimum since you are not building any naval forces
	land_doctrine = 	1.0
	air_doctrine = 		0.001	#Minimum
}

PS
IF anyone wonders why I have the 0.001 instead of 0 in the tech trees is because I for historical reasons wouldn't want to remove them totally.
DS

PB-DK · Aug 26, 2003

hi guys... first post here, so please dont flame me too much

i have been tinkering with the ai files for some time now, so i would like to post some off my results:

Max and minimum Garrison:
most countries should have a min garrison of 0.1...
Max garrison: well i have played around with it... seems to me that a value of 0.5 is the max a value of around .3 for the most aggresive countries seems like a good compromise...

Max factor:
well, been doing a bit buzzing on this... M.F. of 0.5 for the peace AI's and those countries that where in 'a industrial expansion face' e.g. early germany, USA for the most part...
a max factor of 0.2 for when countries are on the defensive,
countries like canada, that does not have the resources to expand thier industrial base should have a MF of 0.0000 (ZERO) at all times
this allows countries to 'rapidly' build up their IC base while still balancing growth vs units'n research

priorities: *ARGH NEEDS EDIT

*
a priority of 0 on Human borders works great! then it nolonger deploys 30 divs on your boarder when you are playing a small country... !!!
The wartarget priority... i have set most of my ai's on 150-200 this works fine as the ai starts to deploys divisions on the boarder (sometimes just not in time before the declaration of war

)

key point multi...
leaving your keypoints at the merci of the flow of battle is NOT a very great idea... i have used a KPM of 2 for most of my AI's although i think smaller countries with only a few keypoints should have a higher KPM... up to 5 i think... (still havn't testested this out completely but it seems to work great)

ahh last item for today...
*drumroll* *dadaaaaa*
reserve prop... i have also been testing it... for most countries 0.2 is just below what they should use to be efficient...
i have used a R.P. of 0.4 for the german defensive ai... this seems to work great to simulate the battlegroups (kampfgruppen) particulary late war, when the germans(pure ai (post 41)) start to encounter the massive armies of both russia and usa...
generally i have set it for 0.3 which is just about a third of the armies involved, this does give the ai more units to back up their threatended provinces (e.g. they become more strategically flexible)
low RP's (0.1-.2)should probably be assigned those nations that does not have a lot to defend (one or two keypoints) or that where historically inflexible (dont know about russia in barbarrossa... i guessed they just used thier reinforcements faster than they got them...

anyway... this is just my own experiences from my own editing of the AI files... i dont have much knowledge in creating stunningly new (and creative) AI files... i just like to tinker with them.
(tinker... hmmm mess, fix, mod just a few different words...)
still havn't gotten into much of the units and tech parts of the ai... perhaps some other time...

*note that all values are to be compared to the standard ai files and values *

*drumroll* hope you can use any of it!

unmerged(6537) · Sep 3, 2003

Canadian Research and Production

I quess I have to raise the Canadian National Flag here to discuss what Canada did in WWII.

Spoonist...you are right. In 1936... we did not have much oil, did not produce much iron ore and were short in a few other areas of production that makes Canada look like a production wimp except...

The USA/Canada continental economies and production were then and still are...completely intergrated.
(oh the polititians will swear otherwise but they lie).

We are still and have been, since before WWII, the USA's largest trading partner...bar none.

We want to buy ...they sell...they want to buy... we sell...period.

That means their iron ore, oil came to our mills , our cobalt, nickel , gold ,copper, uranium to theirs.

This {resource poor ???} under-populated country pumped out in WWII this...

ARMOURED VEHICLES

UNIVERSAL CARRIERS 28,992

WINDSOR CARRIERS 5,000

ARMOURED CARS 1,506

LIGHT RECCE. CARS 1,761 [OTTER]

SCOUT CARS 3,255

ARMOURED TRUCKS 3,961

VALENTINE TANKS 1,420

TANK CRUISER RAM 1,948

SELF-PROP 25 PDR SEX-TON 2,122

TANK CRUISER GRIZZLY 189

TANK COMMANDER O.P. 84

TANK AA 20MM QUAD SKINK 3

TOTAL 50,241 Armored Units

Plus

a few ...(31,500) Misc

STATION WAGONS, STAFF CARS

AND 4 X 2 Jeeps

Grand TOTAL = 81,942 Vehicles

Canada Also Produced....

14,231 Hurricane - fighters

430 Lancaster - Bombers

894 Curtis Helldiver - Dive Bombers <- USA design

1,000 Mosquito - Bombers

and about 10,000 other types such as... 5,000 trainers - 674 Catalina flying boats <-USA design

{We needed the trainers as we trained 20,000+ pilots.{

Ships... not as many...8,000 small craft and 1,000 Atlantic naval and convoy vessels.

We were the worlds 4th largest naval power at war end...

And our factories were kept fully working mostly... by women.

You will note that the UK and USA freely shared all their research with us and I mean all.

How much...well...

We provided the US with the uranium for the A bombs and the heavy water for researching it. (still did well into the 60's)

We've had the capacity to produce A weapons since the 50's but thankfully...not the desire... due to this research sharing.

As to moving planes, tanks, armies....long distances fast....we have excellent railroads.

This is not ment to be a flame...just some facts...to brighten up the thread.

AI PROJECT: 4th July 2003: Results discussion

Bierbrauer

Aerandir Eärfalas

Second Lieutenant

Bierbrauer

Aerandir Eärfalas

Captain

Defender of the Indefensible

Aerandir Eärfalas

Second Lieutenant

General

Bierbrauer

First Lieutenant

General

Bierbrauer

Private

Captain

Field Marshal

First Lieutenant

Former Paradox Fan

Second Lieutenant