Difficulty level questions

[1stvermont]

1]when choosing the difficulty level of AI, does increasing the difficulty simply give more "cheats" better cohesion/seeing better etc?

2] What would the AI setting be so that it has no cheats?


3] What would be best AI performance, while not simply giving to many cheats?


thanks.

[Cardinal Ape]

The difficulty level gives the AI a bonus to movement speed, cohesion recovery, increased production, reduced command point penalties, and free replacements.

Playing at 'standard' the AI does not receive cheats, I think.

Most people seem to set the AI detection level to a lower setting. If the AI can see too far through the fog of war it will cause it to make some wonky moves - most commonly when it rushes toward Pittsburgh in '61.

[ifailmore]

try low detect ai make it less like a mongol horde runnign around once they got through a hole in your defense lines. I play on leutnant ai, med trafic med delay, historical atrition to both ai and player, easy supply off, historical forces,give more time, all AI behavior,veteran activation, MEDIUM aggressiveness, +1 activ bonus

BUT im having more fun using Hard aggressiveness ( I just feel theirs more battles as the ai tries a lil bit harder) BUT with no bonus to ai activation so its based on their general stats I think the way activation occurs

[1stvermont]

Thanks for both responses. The manual says to give extra AI detect bonus to help it strategically that was my plan for next game. But you both seem to say no, so i should give no bonus to detection than? it will make ai better this way?


what does med delay have with AI? does that effect performance at all?.



another thing that i was unsure on is the activation system for generals, what is this suppose to represent historically if anything?


thanks.

[1stvermont]

would it help to have AI as north with more line of sight or less fog of war so they can attack better?

than AI as south less so they dont run crazy?

[Captain_Orso]

The affects of giving Athena a greater advantage with FOW will be the same for both sides; she will be more likely to attack locations further behind from the front lines.

[ArmChairGeneral]

I think of the difficulty settings this way: Detection bonus gives her more information to base decisions on, Difficulty (Rank) makes her soldiers better at operational tasks (like marching and maintaining cohesion ) and Activation makes her generals more responsive and more able to execute the orders she gives. No setting will change her combat rolls directly, and no setting gives her more resources or troops to work with, except for the highest Rank setting, which increases the size of her replacement pool somewhat (though she still has to find a way to pay for them out of the normal level of resources). You can increase the size of the force pools, but this affects both sides, not just the AI, so isn't a "difficulty" setting per se.

Activation:

The activation rules are designed to abstractly model the friction that happens between when orders are given and when/if they are executed. It represents (among other things) leaders who are unprepared, unwilling, slow to get moving, or simply unable to execute their orders in a timely manner, as well as unexpected minor delays from roads being washed out, having to wait for the supply wagons to come through, etc. There are some other mechanisms for modeling some of these things, like the delay setting, but activation is tied to the leadership qualities of the general so that good ones are able to overcome these normal and expected problems as they arise in the field. In many ways, the strategic rating (which activation rolls are based on) is the most important of a general's three stats.

In my experience at least some of the problems with the AI being passive are related to inactivity (especially for the Union). The AI does not do as good a job at managing activity levels as the human can, so in many cases she is issuing the right orders, but they simply can't be executed because her leaders are inactive or fixed in place. When I play with the hard activation settings I especially like to give her an activation boost: she is going to suffer more from her troops being fixed in place than a human will since she cannot anticipate whether a leader she gives an order to is likely to be fixed the following turn, and can't work around any specific incidents as well as humans can.

Detection:

Athena does sometimes have a problem with getting carried away when she can see too far into the back ranks due to Detection bonus. This (IMO) is partly because she values some objectives (Pittsburgh and Knoxville are the most notorious examples) more highly than she should. When she can see far enough into the back-lines she can see when distant regions she values highly are lightly defended and will often make ill-considered lunges toward them thinking her opponent has made a mistake. This is annoying, and some recommend not giving her a high detection bonus because she can easily fritter away corps sized stacks pursuing impossible-to-achieve objectives. Personally, I find that with high Detection she is enough better at everything else that I give it to her anyway and find ways to work around this problem by keeping sufficient garrisons to discourage her (she looks at the PWR rating to determine this) or just role-playing it and forgoing taking full advantage of the error.

Is this ideal? No. Should this well-documented and sometimes crippling issue be addressed at some point by either the developers or the community so that the AI can reap the full benefits from a high Detection? Of course. Nonetheless, high Detection DOES result in an increased ability to respond to and "anticipate" the human player's moves, and allows her to keep up with a human's far superior capability for recon, resulting in more effective tactical behaviors IN MOST SITUATIONS. She will be much better at detecting weak points in defenses too. She is already pretty good at making humans pay for glaring mistakes, so enhancing her ability to see when those mistakes are made definitely makes her a tougher opponent. (She just needs her definition of "mistake" tweaked slightly).

"Difficulty" (Rank):

This setting is somewhat misnamed, as it is not the only, or even necessarily the best, way to make the AI better, but does give her the most concrete and direct mechanical advantages of the three settings. She moves faster, and recovers cohesion it more quickly after movement or battle (and loses less cohesion to movement in the first place since that is a function of how many days the force is marching). It definitely does not give her direct bonuses to combat stats, although of course having more cohesion when she arrives at the battlefield does affect the outcomes of battles indirectly. Unless I am totally mistaken, and please correct me if you have evidence of this, the Difficulty setting does not give her more troops or more resources, it just makes her troops better at doing things.

[FightingBuckeye]

ArmChairGeneral,

I think you really captured just about everything as far as the settings go. But, I do believe that the AI gets some sort of production bonus at the higher levels. I've seen the South's economy almost rival the Union's. And the Union AI gets so much stuff it's almost stupid.

[Cardinal Ape]

At the highest level the AI receives a 100% bonus to the production of money, conscripts, and war supplies. 50% bonus at the second highest rank.

[1stvermont]

Thanks for all the info guys.

[ArmChairGeneral]

Thanks for the correction guys!