LargeSlowTarget
Posts: 4443
Joined: 9/23/2000 From: Hessen, Germany - now living in France Status: offline
|
Originally posted by Pasternakski: [I]My needs are few - just the ability to order my subordinates NOT to do obnoxiously ridiculous things. I don't want to micromanage air strikes, I just want to set some rules of engagement as a good theater commander would do. (...) I want to be able to say, "Keep your aircraft away from Rabaul, it's too dangerous up there, but be on the lookout for attempts to resupply bases to the west and southwest."[/I] I agree. The design decision to leave the naval attack target selection to the AI is fine with me - the only problem is that those AI decisions are sometimes very questionable. I don't question 'errors' and blunders like misidentifying targets or escorts getting separated (these things did happen IRL and it is well modelled in UV, although a few more messages during combat resolution might help to explain what has happened). I question decisions to attack low-value targets far away that pose no immediate threat (e.g. MSW at Rabaul) time and again despite high costs (CAP, fatigue, morale, weather, operational losses) but little success - and even worse, to do this when at the same time high-value targets directly threatening my bases are close and abundant (e.g. invasion TF off GG)! [Yes, might be a weather issue, but how realistic is this: "We can't find the invasion TF in our backyard because it's socked in, so we better take off to hit this MSW at Rabaul". Yeah, sure.] I have nothing against high-risk missions accepting heavy casualties as such - but only if the price is worth it. When my bombers attack an invasion TF on its way to GG for example, find the TF covered by LRCAP and suffer heavy losses - so be it. This is war and we have to seek combat and accept losses. But I want my AI subordinates to realize that it is silly to risk destruction of entire squadrons just to attack small fry at a place like Rabaul. One or two sharp defeats of my bombers should be enough to teach my commanders that the bomber not always gets through, and they should change targets and tactics accordingly and leave this place alone until fighter escort is available. Unfortunately my commanders don't seem to learn from defeat. [B] So what we need are better AI subordinates. I don't ask for the ability to target individual TFs or to assign target priorities myself. I'm asking for a more intelligent AI target selection to avoid suicidal attacks against low-value/high-risk targets far away while high-value/most-likely-less-risky targets nearby are ignored.[/B] Now this is easy to demand, but may be hard to implement. I'm not a programmer, but to incorporate a coherent concept of 'calculated risk' and 'cost-benefit analysis' into the AI target selection sounds difficult to me. One way around it might be limited player target control: restrict naval attack missions to 'at sea only' by default - if you want to allow naval attacks into the base hex, you would have to use the 'port attack' or a new 'attack ships in port' option as secondary mission. But let's keep it in proportion, the lack of player naval target selection or the sometimes poor AI decisions does not make UV unplayable. To give some credit were it's due: IMO UV is the best game to date in this genre and the Matrix staff deserve highest praise for the game, the exemplary player support and the continuous patching efforts. I for my part am more than willing to pay for an add-on if it includes certain enhancing features - target priority selections for air and subs, being able to designate patrol zones for subs, interception for surface TFs, and a 'pure CAP/stand down'-option, to name the most important ones.
_____________________________
|