AI responsiveness

[Tantum eruditi sunt liberi]

Played just under 50 turns with KH. Awesome game so far. But, unlike EBI, where KH was incredibly challenging due to Makedonia breathing down your neck, they don't do anything. They sent a couple reinforcements during the siege of Korinthos, but nothing serious and they were not offensive at all. Epieros and Makedon just sort of sit around building stacks and getting poor with their armies un-garrisoned. I'm about to wipe Epeiros out, since they refuse to attack me or accept a ceasefire, no matter what I do. They can have three stacks lying around, but if I attack a city, they'll ignore it.

Roma and Qarthadastim haven't done anything, nor has the Seleukids and Ptolomies. On a positive note, Pergamon has advanced on the Seleukids in Asia Minor.

So, is this working as intended with the lethargic AI? Is the summer release going to make adjustments? I'm playing 2.03b on medium.

[z3n]

"No matter what I do"

How much gold have you offered them? Have you offered them a city?
Anyway, as has been said before- if the AI is just being silly use force diplomacy.



Since the summer release is further away- yes it will probably have adjustments to it.

I'm pretty sure you're either a) blitzing and or b) playing rather well in terms of stack management/fronts of war. When I tried to blitz athens as makedonia I was in fact attacked by epeiros after leaving pella relatively open. But the CAI is WIP one step at a time because I do have several other fronts of interest to work on (as do all the other team members).

I'm very curious about your remarks though 'Roma and Qarthadastim haven't done anything, nor has the Seleukids and Ptolomies'

Here's a recent test run of mine and the factions you mentioned have clearly done something.. Can you be more specific as to what they haven't done in your playthrough?
http://i.imgur.com/fYu5LAD.jpg
http://i.imgur.com/cFKd1sY.jpg

Here is a comparison screen so you can see the minimap compared to the above ones^
http://i.imgur.com/iesECdi.jpg

[Thuycidides]

The point is that blitzing shouldn't be possible when Makedonia and Epeiros have multiple stacks (though it is a clever way to blame the player for the AI's passivity). I have observed exactly the same thing as the OP when I was testing the KH reforms. Between the two hellenic kingdoms Makedonia is especially pusillanimous.

[z3n]

The point is that blitzing shouldn't be possible when Makedonia and Epeiros have multiple stacks (though it is a clever way to blame the player for the AI's passivity)

First this is not to say I am the only guy with more than one task or the most tedious and time consuming task. Many others on our team have the same issue hence this thread for example.

Well it's a lot more of a complex issue than you might realize. I am also doing the BAI/pathfinding/cohesion alongside of the CAI's foundation code and diplomacy. Additionally, I have work in real (as do other team members) alongside my own school, for which I have several final assignments due this week. Therefore I can say that I'm pleased with the progress made thus far, thanks to the feedback such as this. However- it isn't necessarily easy or simple to design anything AI related.

To get into the technical part of things. As far as my testing has showed. Four formulas play into effect.
(att_str*att_str_modifier)/def_str
att_def_strength_ratio < (ideal_str_ratio*str_limit_weak)
att_def_strength_ratio > (ideal_str_ratio*str_limit_strong)
ideal_str_ratio = > 1.0 (invade_min_priority/invade_max_priority*1.0=ideal_str_ratio)


Three other main factors come into play- the various invasion strategies, the varied invasion decisions and their respective invasion priorities, the various invasion factors (i.e. the fact that the CAI often only calculates attacks based on one stack, though my take on the various formulas has currently helped in that respect).

Thus far I have been able to make an RTW style aggressive CAI (albeit with it's various problems) I believe however that other various complaints would soon come (i.e. I was besieged by an army smaller than mine!). That is the reason I am currently in the preliminary stage of testing before its release of the various mathematical equations and trying to find their respective results. At times CA has written code which was re-purposed or it modifies on top of pre-existing equations which in turn is confusing due to the lack of documentation provided by CA in certain areas of their code. Thus- while it has been relatively slow progress I believe we do with feedback- stand a chance of improving the CAI to a reasonable degree beyond hopefully pre-existing creations (i.e. CA's)

[alin]

Well it's a lot more of a complex issue than you might realize. I am also doing the BAI/pathfinding/cohesion alongside of the CAI's foundation code and diplomacy. Additionally, I have work in real (as do other team members) alongside my own school, for which I have several final assignments due this week. Therefore I can say that I'm pleased with the progress made thus far, thanks to the feedback such as this. However- it isn't necessarily easy or simple to design anything AI related.

And we are also pleased z3n, there have been major improvements to the AI since you joined our team and we're happy to see such good progress.

I'm sure you'll find the right balance in the end, even with all those obstacles provided by CA.

[General WVPM]

...
Thus far I have been able to make an RTW style aggressive CAI (albeit with it's various problems) I believe however that other various complaints would soon come (i.e. I was besieged by an army smaller than mine!).
...

In my Pergamon 2.03b playthrough I noticed the AI moving in a few troops to one of my trade routes and slowly reinforce that army, probably to prepare for a siege (I never let it come to that )
So I don't think that is a real concern at the moment.

[z3n]

Still WIP.. but this is where it stands right now.

[Tantum eruditi sunt liberi]

Ok..."No matter what I do with reason". I get the AI can be silly with diplomacy. I was under the impression it wasn't EBI silly. I see that it's much much better, as Makedonia sued for peace after the loss of Korinthos. It's clearly in Epeiros's best interests to have a ceasefire and our relations are good. I offered 5k-ish with like 2k a turn for 10 turns with variants around there. Having one (or two) towns left and two enemies looming on your borders while you're bankrupt would make that a reasonable (if not very generous) deal. But, I get that the AI diplomacy can be silly. No biggie.

I am certainly not blitzing. I build up an army over many turns and did not engage either power meaningfully for over 20 turns. I attacked with an equal or lesser force with balanced armies (comprised of a reasonable number of levies). B is likely. Obviously, I play with optimal "fronts of war in mind". But, either power could have rolled me or put me in serious trouble at any time with the size of their armies (I had one stack or so around the siege of Korinthos, they each had at least 3). When I attacked Epieros, they had multiple supporting stacks to the north and west (this is well after Korinthos). Rather than lift the siege, they moved around or back and forth not engaging anyone (even Makedonia), so I eventually took the city. Neither side has attacked me during the 50+ turns of the campaign except to sally at the last minute for the two cities that I had took from them.

Maybe I have overstated the lack of action with the rest of the big powers, as I do not have fog of war lifted, but Roma still hasn't pushed towards Rhegion (though, they kicked Epieros out of Italy, at least) and the Ptolomies and Selukids haven't moved much besides losing to Pergamon in Asia Minor. With the observed behavior from my closest neighbors, I thought the general pattern of stalemate that I was seeing could have been related.

[QuintusSertorius]

Also bear in mind, this has all been tweaked again in 2.03c, I slipped another update to diplomacy and BAI in when I was fixing the issues with the descr_strat. So your experience may be slightly different (or a lot different) on that front.