Andy my experience has been that the AI doesn't buy back LCUs. I switch over about once a week and re-buy units manually for the AI. Maybe it only buys back certain units and leaves others? It does however buy back all destroyed Air Groups (other than ship board search plane groups). I will try the new 'nasty' script in my next play through. I do have one question though-how many times will the AI attempt to take the same target before it realizes that A) It needs to bring more firepower to win the fight or B) The target isn't worth it anymore? I currently hold Wake Island, and the Japanese have tried 3 invasions so far-and lost the entire attacking force all 3 times. Each invasion has less chance of success due to strengthened defenses, yet the AI doesn't bring a stronger force to compensate. Is there a cut-off -either based on date or losses, where the AI 'decides' to abandon the invasion? Another rather interesting feature is the fact that the marshal Island betty force has been basically gutted due to constant unescorted raids, yet they keep coming in 3s and 4s to the slaughter, even though each attack is turned back once one A/C is destroyed.
Occasionally a CV group will show up and put my fighters in their place (about every 10 days or so actually), and the first CV battle of the war occurred with the sinking of one Japanese CVL for no losses to my side, which brings into question one further question. Does the AI have the ability to interpret intel? In the example above, my CVs had been on station 1 Hex north of Wake Island for 3 days supporting a supply convoy, and the CV fighters had been involved in a fight against enemy bombers each day. Yet when the Japanese CVs arrived on the scene, they were still set to attack the AF first, allowing me first shot at their carriers. I would have expected the Carriers to either attempt to attack my CVs first, or to withdraw (they were heavily outnumbered-3 US CVs vs 2 Jap CVLs), instead I sank one CVL and heavily damaged the other one, which I could have sank but instead I ordered my force to disengage and allow the other CVL to return home.
So i guess what I am asking is this:Are there algorithms the AI uses to assign both a value to a target and, using available intel, a force sufficient to capture said target? A single SNLF unit is not going to be enough to take Wake anymore, I have the original defenders plus a group of Combat engineers there now, dug in with plenty of supplies. So the question becomes is Wake Is worth it? If the AI has assigned a high value (weight) to the Island, then a force capable to take it should be formed. If not, then invasions should be cancelled, rather than continuing to throw small forces at the problem and allow them to be destroyed.
Please don't take this post as flaming or condescending at all, that is not the purpose of it. I enjoy the game very much, and appreciate all the effort that went into making it, as well as the continued support by the developers. I only hope that by providing feedback we can continue to enhance the game experience, as the computer gaming industry for mature-gamers is slowly dying to be replaced with the newest GTA or COD clone. Matrix continues to be on of the few companies devoted to serious gaming, and one that I will continue to support for years to come. If there is any help I can provide, don't hesitate to ask.