The only thing that should be lowered really is the land percentage of German to Russian troops in the Declare war event, imo. In EIR, I actually decreased the number of troops required in "Garrison" at the Soviet border, because the system is set up to create more historical level of troops, but that should be irrelevant in the vanilla game.
In EIR, the countries that may be in the Axis at the start of Barbarossa is far less certain, and might even include Japan, so, the trigger is set up to include a-historical alliance paterns and so I also created a complex formula for the number of Allies the German's have at the time it wants to attack, as well as land percentage.
I would say no to adding any AI specific surprise event. I actually agree with the concept of the "surprise event" in theory, but would choose it over the GDE malus that the SU gets at the begining of Barbarossa. But, changing the whole GDE structure would have an impact on the strategic abilities of the SU prior to the German attack, in particular in situations where a human player chooses a-historical early war options, and so on. So, unless the whole system were to be re-written from top to bottom, I think it would be best to leave it alone at this point, since we have no idea how such a change would impact the game overall.
First of all one should remember that the SU "AI" is already nerfed at the start of Barbarossa, with very high passivity, which acounts for some of the "surprise" effect you are all ready talking about.
Secondly, there might be some simple changes to the German Barbarossa ai, which could be made very simply without tampering with the event system.
For example (German Barbarrossa "AI" for the first year is set up with the following front conditions in the "GER_Russia.ai" file):
Code:
front = {
recklessness = 1 # 0-3
distrib_vs_ai = offensive
distrib_vs_human = reactive
base_attack_odds = 1.1
min_attack_odds = 1.0
max_attack_odds = 2.0
enemy_handicap = { }
passivity = { }
}
One could modify these as follows to make the German AI even more aggressive:
Code:
front = {
recklessness = 2 # 0-3
distrib_vs_ai = offensive
distrib_vs_human = reactive
base_attack_odds = 1.0
min_attack_odds = 0.9
max_attack_odds = 2.0
enemy_handicap = { }
passivity = { }
}
One could also try "op_offensive" instead of "offensive".
Thirdly, it is a little known fact that the German 1942 AI for Russia is nerfed, I assume in order to pacify the AI in order to ensure more historical results. In other words the German AI is basically programmed to succeed only in 1941, or never, even if it does well.
From the "front" conditions from the "GER_Russia_Second_Year.ai":
Code:
###################################
# Ground Combat, Leaders and Garrisons
###################################
front = {
recklessness = 1 # 0-3
distrib_vs_ai = reactive
distrib_vs_human = reactive
enemy_reinf_days = 6
reserve_prop = 0.1
panic_ratio_vs_ai = 2.5
panic_ratio_vs_human = 2.5
base_attack_odds = 1.3
min_attack_odds = 1.1
max_attack_odds = 3.0
reinforce_odds = 1.2
withdraw_odds = 0.6
enemy_handicap = { }
passivity = { }
}
These could be changed to make the German AI more aggressive in 1942:
Code:
###################################
# Ground Combat, Leaders and Garrisons
###################################
front = {
recklessness = 1 # 0-3
distrib_vs_ai = offensive
distrib_vs_human = reactive
enemy_reinf_days = 6
reserve_prop = 0.1
panic_ratio_vs_ai = 2.5
panic_ratio_vs_human = 2.5
base_attack_odds = 1.2
min_attack_odds = 1.0
max_attack_odds = 3.0
reinforce_odds = 1.2
withdraw_odds = 0.6
enemy_handicap = { }
passivity = { }
}
Starting with those kinds of changes and then tuning the AI from there is a much simpler and subtle approach than changing the event system.
The problem is with resources. The USSR has so much of everything that just by improving the USSR AI, it was made almost impossible for the German AI to go east. Remember that it doesn't know about brakthroughs and encirclements, and better builds don't do squat once it runs out of manpower. If you want a better Barbarossa, you need to give the German AI some kind of an artificial head start. This is what I meant in my previous post.
Actually it does know about breakthroughs and encirclements, and there are parameters which tell the ai how agressive it should be in closing pockets. It breaks through because it can be told to focus on the weak spots in the line, in order to create breakthrough, which then creates pockets, which the AI will then try and close.
And no, they were not teleported to the redeployment pool - they were GONE. I guess the game handles an ally reclaiming control of his territory exactly the same as an actual enemy force seizing it. :-\
