Is the AI exempt from command delays?

[SteveP]

Not as stupid a question as it may seem. Either the AI is exempt from command delays -- at least in the first turn -- or there is something seriously wrong with my software for both CMBB and CMAK. I have now seen this several times under a variety of FOW conditions: when on attack or in any other situation where the AI wants to move in the first turn, its troops are moving from the very first second of game play, even if green or conscript. I don't know if this has always been true, or if I just noticed it for the first time because I've recently been playing some scenarios in which the two sides start out in full view of each other.

My assumption is that the AI does not have this exemption, but that means I've got a very mysterious thing going on with my system. I might add that I even went to the point of totally reinstalling CMBB and its patches but this odd phenomena persists.

Anyway, I'm stumped and it's driving me nuts. Anyone got any ideas.

[Wicky]

If as you say units on the 1st turn are in sight of each other then they will overide any orders (even your own) and do what's necessary to 'preserve' themselves if they perceive a threat.

Did you also notice if your own vehicles immediately take on a life of their own at variance from your orders? i.e. reverse/fast forward into cover etc.

This is what BFC call tacAI in operation and is applicable to AI controlled forces as well as your own.

[Other Means]

IOW, when you give your units orders you're acting as the commanding officer and therefore the delay in getting the orders to your men and having them react is modelled.

However if the men take action on their own it is immediate and dependant upon the circumstances. This is what's referred to as the "tactical (tac) AI".

So what you're probably seeing is the AI's men tac AI reacting to circumstance as your own would.

The AI that tells its side what to do is the "strategy (strat) AI" and still incurs command delay, AIUI.

[Michael Emrys]

Steve has said several times that the AI doesn't cheat, that it is subject to, and to the same degree as, all the rules, including spotting, command delay, etc. Over the years, players have observed actions that appear at first to contradict that position, but SFAIK, on closer examination that has proven not to be the case.

Michael

[SteveP]

OK guys, your responses are very understandable, but what I am seeing isn't that simple. Here's a case in point:

Russian steppe, open terrain, German defender (me) vs. Russian attacker (AI). Playing EFOW. Russian has nothing but green troops (it's a scenario, so I know exactly what he has and where they are starting from). No experience bonus for the AI. My troops can see the entire Russian setup zone. I give my troops short covered arcs or otherwise insure they don't fire at anyone. So, the Russians can't see me and they aren't getting fired at. Here's what I see: Russian infantry and tanks starting to move immediately (i.e., zero command delay) on turn one. Movements are the usual sort of helter-skelter of AI directed troops, though generally consistent with its path-finding methodology (i.e., looking for a covered route to the flags). Some Russian troops don't move in the first turn, although I can spot them. However, they do move in the first second of the second turn (again implying no command delay).

I have now looked at this in several different scenarios and QBs, and I see it every time. I have, for example, tried it in a scenario in which the AI had extensive cover, but I turned off the FOW in order to see what the AI did. In that case, the AI executed very normal moves (e.g., infantry using move commands through the woods toward my flag) but those moves began on the first second of the first turn -- zero command delay.

As I said in my post, I'm having trouble (apropos of Michael Emrys' point) believing this is what should be happening. On the other hand, I am starting to think that possibly it was necessary to exempt the AI from command delays in order to keep the game moving at a reasonable pace -- since the path-finding activity is inherently time consuming with lots of starting and stopping.

Either that, or my software is screwed up in some perplexing way. Anybody interested in running your own test??

[Wicky]

Steve P,

mmm You might be onto something - I just tested a scenario 'waltzing matildas' and also edited in a couple of sharpshooters, with cover arc, inh overview to watch what happens. I similarly observed that some (not all) the AI (rusky) forces setting off as very promptly from their starting blocks, despite not being under fire or detecting any human controlled enemy threat.

*scratches head*

[SteveP]

Originally posted by Wicky:

Steve P,

mmm You might be onto something - I just tested a scenario 'waltzing matildas' and also edited in a couple of sharpshooters, with cover arc, inh overview to watch what happens. I similarly observed that some (not all) the AI (rusky) forces setting off as very promptly from their starting blocks, despite not being under fire or detecting any human controlled enemy threat.

Thanks for checking this out for me. I had stumbled across this while trying to figure out another odd AI behavior that was giving me fits. Between the two, I was really convinced that something had happened to corrupt my game files -- yet I could not seem to repair the problem. I did finally figure out that the other AI behavior problem was an another Stupid AI Trick -- one that AFAIK has never been reported before.

Anyway, it's just possible that giving the AI this exemption command delay does not fall into the category of what Steve would consider instances of "cheating" by the AI.

[Michael Emrys]

Well then, we now have two documented exceptions to Steve's statement that the rules apply equally to the AI as to a live player. The next question is how far does this go? I.e., does the exception only apply on the first turn, or all though a game? And does it matter which side the AI is playing or what year, etc., etc.

Lastly, if this is a genuine bug, is BFC taking steps to ensure it won't appear in CMx2?

Michael

[aka_tom_w]

very interesting

hmmm

-tom w

[JonS]

Originally posted by Michael Emrys:

Well then, we now have two documented exceptions to Steve's statement that the rules apply equally to the AI as to a live player.

What is the other one?

[Michael Emrys]

One each by SteveP and Wickie.

Michael

[JonS]

Ah, ok. Same thing seen by two different people. I thought you meant two different 'cheats'.

[SteveP]

Originally posted by Michael Emrys:

Well then, we now have two documented exceptions to Steve's statement that the rules apply equally to the AI as to a live player. The next question is how far does this go? I.e., does the exception only apply on the first turn, or all though a game? And does it matter which side the AI is playing or what year, etc., etc.

Lastly, if this is a genuine bug, is BFC taking steps to ensure it won't appear in CMx2?

Michael

It is difficult to be certain what is happening on subsequent turns, because we don't know what orders the AI is giving its troops (including pause orders, which I doubt the AI uses but it's possible). What I have seen is ambiguous, however. On the one hand I have seen units that were not moving at the start of turn 2 or turn 3, for example, suddenly start moving immediately on that turn. That strongly suggests to me that they had no command delay (since even the use of pauses rarely causes a move to start precisely at the beginning of the following turn).

On the other hand, I also saw units that were stopped at the beginning of a turn start moving only in the middle of that turn, which suggests a command delay was involved. I should add that these two behaviors were observed side-by-side in the same scenario (same type of units: Russian green infantry). So, make of that what you can!

Anyway, I've tested this in a variety of scenarios and QBs, from both CMBB and CMAK, so it is fundamental to the game design -- not something that's scenario specific like nationality, unit experience level, year, or anything else of that nature.

Assuming what I am seeing is real and not some sort of optical illusion (and heaven only knows if Steve is going to tell us), my theory that this is designed to offset a lot of starting and stopping behavior in the AI's path-finding system is probably not correct. A better theory may be that the path-finding appears to involve a lot of zig-zagging around (meaning a lot of waypoints). Humans are better at figuring out how to get somewhere on the map with the fewest possible waypoints. Even then, I'd still be more inclined to believe that this is about speeding up game play than to give the AI some sort of advantage (and, anyway, the AI is so handicapped as it is, why not give it an exemption from command delay )

[flamingknives]

Has anyone checked if Human controlled troops are subject to command delays on the first turn?

Just a thought.

[Wicky]

edit - nuffin

[aka_tom_w]

....and, anyway, the AI is so handicapped as it is, why not give it an exemption from command delay

When I first read this thread that was my exact thought.

Its probably meant to be that way (no AI command delay) and its probably not a big deal if this is the FIRST time it has come up in the past 5-6-7 (?) years of EXTENSIVE game play and play testing!

-tom w

[SteveP]

Originally posted by aka_tom_w:

</font><blockquote>quote:</font><hr /> ....and, anyway, the AI is so handicapped as it is, why not give it an exemption from command delay

When I first read this thread that was my exact thought.

Its probably meant to be that way (no AI command delay) and its probably not a big deal if this is the FIRST time it has come up in the past 5-6-7 (?) years of EXTENSIVE game play and play testing!

-tom w </font>

[McIvan]

Originally posted by flamingknives:

Has anyone checked if Human controlled troops are subject to command delays on the first turn?

Just a thought.

I assume you're not serious but just in case you are, I can confirm having watched the first replay of a QB last night, that, yes, human controlled troops are subject to command delay.

Regards

Ivan McIntosh