Jump to content

Does AI cheat?


Recommended Posts

Just an observation...as I am creating a scenario, and obviously want to play both sides to make sure it is "fair" from both sides...or, at least 'possible to win' from either side.

I have a red RPG team, with AI orders of ambush @ 600m...when I am playing blue, this succesfully works, it is pretty much invisible, even for me, who designed it, and in tests thus far, 4 of 5 times, even with myself subconsciously "searching" for something as designer I know is there, but as a unit commander I would not..my first notice of it was when one of my vehicles either blew up(3 times) or was immobilized (1 time)..so, I can say it works well under AI control...now, when I play the red side in the test, the very same unit is spotted and destroyed at long range by the same vehicles it blew up without them seeing it, under AI command. So my question is, does AI cheat? and if not, how is this working...what under 'human' command is equivalent to AI's "ambush at xxx meters"...I had assumed it was 'hide' with a target arc, but that is what I was doing and it quite obviously is not the same thing.

Thanks,

Ron

Link to comment
Share on other sites

AI-controlled units do seem to have a superior spotting ability to human-controlled ones.

If you're solo playtesting a scenario to optimize it for 2 players, you may want to try WEGO Hotseat mode, which allows you to issue orders to each side in turn, then shows replays from each side's perspective.

I am not the world's biggest expert on Target Arc, so I'll let someone else advise you on that one.

EDIT: I just remembered, yes, when you order the AI to have a unit Ambush at xxx meters, it sets a circular Target Arc of that size. When you click on such an AI-controlled unit while playing in Scenario Author Test mode, a "yellow fog" appears around it. For a while I thought it was a bug.

Link to comment
Share on other sites

If this is true then it....well......SUCKS.

I have not heard anything firm on this.

I see it has been read a lot of times, but yes, nobody must be sure since not much in the way of replies...if it is true it does somewhat suck...but possible to work around...just would have to rethink my normal tactics...i know snipers i have controlled have stayed hidden quite well oftentimes, but yes, there does seem to be some sort of AI warning to it's units "Hey...look over there, that hidden guy has an RPG" lol

Link to comment
Share on other sites

I do NOT believe that the AI-controlled forces are getting "hints" on what to target that a player-controlled unit would NOT have. So I don't believe that the "AI is cheating".

Using a combination of "Hide" and "Target Arc" seems the most likely way to set an ambush with the current set of commands, but I believe that the "Hide" command may be causing problems. According to the manual the "Hide" command causes infantry forces to go prone and avoid getting spotted. This MAY cause spotting issues/reduced spotting effectiveness and your RPG team may not be spotting as effectively if they were NOT hiding.

Doing a 'Hot Seat' of the scenario with minimal player interaction with the 'ambushing forces' may give an idea as to what is happening.

Link to comment
Share on other sites

Not casting aspersions on the game engine here, as at this point I am delighted with the results I'm getting in Ramadi. But it has seemed to be consistently the case that AI RED units in buildings spot my nonmoving non-Ordered BLUE units in buildings first. Even when RED were the ones that moved last.

This seems to be generally my case as well..also not casting aspersions on the game, it is possible to work around it, if we know it is set this way.

Link to comment
Share on other sites

I do NOT believe that the AI-controlled forces are getting "hints" on what to target that a player-controlled unit would NOT have. So I don't believe that the "AI is cheating".

Using a combination of "Hide" and "Target Arc" seems the most likely way to set an ambush with the current set of commands, but I believe that the "Hide" command may be causing problems. According to the manual the "Hide" command causes infantry forces to go prone and avoid getting spotted. This MAY cause spotting issues/reduced spotting effectiveness and your RPG team may not be spotting as effectively if they were NOT hiding.

Doing a 'Hot Seat' of the scenario with minimal player interaction with the 'ambushing forces' may give an idea as to what is happening.

As the "player" on the RPG side, i left them hiding,and they were shot and killed at long range though,without moving. As the blue side player, they were able to move, apparently,spot me,and target blue vehicles, all without being seen,even with me as designer,knowing where to look. If I understand what you are saying here right, it should be reversed...that when I "hid"them..they cant see,but cant be seen,either.It seems a fair trade to give up spotting ability for hiding ability...but if the hiding,when I was ordering it, did not still hide them,allowing blue vehicles to fire on them at long range 600m+,then not sure how "hiding" works for human controlled, but it seemed to work well enough for AI controlled, so that they could also still spot me before I spotted them.

Link to comment
Share on other sites

If the unit that is killing your RPG team is equipped with 'thermal imagers', then it is possible that such units will spot ANY unit that is within LOS. That's the effect of thermal imagers in the game as far as I'm aware. Possibly a bit too 'uber', but I don't know the true mechanics of spotting with thermal imagers.

So 'Blue' forces will have a bit of an advantage with spotting since more of them are equipped with thermal imagers. However I'm not sure why these units, when under player control, do not spot units within LOS as quickly as they seem to do under AI-control.

Perhaps there is something wrong with spotting routines for AI-controlled units. If there is, then I wouldn't classify it as an intentional "cheat".

To present this as a bug there would probably have to be quite a bit more testing to verify that there is a consistent behavior that should not be there.

Link to comment
Share on other sites

If the unit that is killing your RPG team is equipped with 'thermal imagers', then it is possible that such units will spot ANY unit that is within LOS. That's the effect of thermal imagers in the game as far as I'm aware. Possibly a bit too 'uber', but I don't know the true mechanics of spotting with thermal imagers.

So 'Blue' forces will have a bit of an advantage with spotting since more of them are equipped with thermal imagers. However I'm not sure why these units, when under player control, do not spot units within LOS as quickly as they seem to do under AI-control.

Perhaps there is something wrong with spotting routines for AI-controlled units. If there is, then I wouldn't classify it as an intentional "cheat".

To present this as a bug there would probably have to be quite a bit more testing to verify that there is a consistent behavior that should not be there.

The RPG teams were Syrian uncons, no thermals.

I am not really sure I would call it a bug, not even really sure I "mind it"being there, just curious if it is there, as it would change my approach somewhat to scenarios as designer, and to playing tactics as well.

Link to comment
Share on other sites

Using a combination of "Hide" and "Target Arc" seems the most likely way to set an ambush with the current set of commands, but I believe that the "Hide" command may be causing problems. According to the manual the "Hide" command causes infantry forces to go prone and avoid getting spotted. This MAY cause spotting issues/reduced spotting effectiveness and your RPG team may not be spotting as effectively if they were NOT hiding.

Wow, this post could have come right out of the old CMBO/CMBB/CMAK threads (and it made its appearance in all of them).

If a unit is hiding, (i.e. face down in the dirt so as not to be spotted) then it is to be expected that they are not going to spot well, yes?

Or to quote Sgt. Major Plumbly from "We Were Soldiers Once, and Young. to a young reporter in LZ X-RAY, "Son, You can't take pictures with your face in the dirt."

If you are looking then you can be seen. If you're not looking, then you will be surprised.

Link to comment
Share on other sites

Wow, this post could have come right out of the old CMBO/CMBB/CMAK threads (and it made its appearance in all of them).

If a unit is hiding, (i.e. face down in the dirt so as not to be spotted) then it is to be expected that they are not going to spot well, yes?

Or to quote Sgt. Major Plumbly from "We Were Soldiers Once, and Young. to a young reporter in LZ X-RAY, "Son, You can't take pictures with your face in the dirt."

If you are looking then you can be seen. If you're not looking, then you will be surprised.

Well, I think that you are missing the point...with human control, hiding,not looking, they are shot from long range...with AI control, not hiding, able to move and target, they are still not seen until they actually fire. Therefore my original question was how can they remain invisible under AI control while able to move and fire, and under human control, even hiding, they are seen and shot before an enemy closes to range of their "ambush"order....even though they CAN see the enemy while they are hiding,under my control, but are ordered not to fire until 600m...the same orders for AI mean that as the human I cannot see them until then. It is not a matter of "being surprised" by the enemy, it is a matter of letting the enemy come into the kill zone...which the enemy wont seem to do until it destroys them.

Link to comment
Share on other sites

The RPG teams were Syrian uncons, no thermals.

I am not really sure I would call it a bug, not even really sure I "mind it"being there, just curious if it is there, as it would change my approach somewhat to scenarios as designer, and to playing tactics as well.

Well I would call it a bug. A huge bug and I very much do mind it being there. The problem, like somebody said, is proving it. I think our best bet would be the guy who has tested enough to notice it. Maybe send all your thoughts directly to Steve?

Link to comment
Share on other sites

I'm surprised nobody has mentioned CIVILIAN DENSITY regarding this problem. Simply put, AFAIK, the higher the civilian density, the more difficult it is for BLUE units to spot RED UNCONS, regardless of all other factors. Conversely, civilian density does not change the ability of RED units to spot BLUE. I find both of these concepts make sense, since it is easier to spot a fully clad US soldier than it is to spot a guy who looks like any other guy but has a rifle under his cloak. Until he raises the rifle and starts shooting, of course. And civilian density applies the same way in urban and rural terrain.

Don't know whether I am explaining this entirely correctly or whether it means anything in terms of the problems people are complaining about here, I just thought it should be mentioned. In this regard, it is not the AI that has a spotting bonus, but the RED (Syrian) side. Depending on civilian density, that is. And this is not a bug but a realistic feature in my opinion.

Link to comment
Share on other sites

I'm surprised nobody has mentioned CIVILIAN DENSITY regarding this problem. Simply put, AFAIK, the higher the civilian density, the more difficult it is for BLUE units to spot RED UNCONS, regardless of all other factors. Conversely, civilian density does not change the ability of RED units to spot BLUE. I find both of these concepts make sense, since it is easier to spot a fully clad US soldier than it is to spot a guy who looks like any other guy but has a rifle under his cloak. Until he raises the rifle and starts shooting, of course. And civilian density applies the same way in urban and rural terrain.

Don't know whether I am explaining this entirely correctly or whether it means anything in terms of the problems people are complaining about here, I just thought it should be mentioned. In this regard, it is not the AI that has a spotting bonus, but the RED (Syrian) side. Depending on civilian density, that is. And this is not a bug but a realistic feature in my opinion.

I agree with you...however, once again I must explain that this is not what I am talking about...if this were so, then in the tests where I was the red player, I would have the same advantage..however, in those tests, MY red uncon is shot immediately...all other things being equal, the same population density as well...nothing in the scenario changes, except the side I as designer am playing...when the AI is controlling the uncon, it remains invisible until it strikes...pretty much WAD in my opinion,the result I am aiming for..but when I control it, it is shot immediately before it has a chance to carry out it's orders.

Link to comment
Share on other sites

I see, that is strange then. Aside from the obvious variations in outcomes from randomness in the spotting routines, I don't have any clue where this might be coming from. But it would certainly take a large sample of games with a fixed set of conditions to prove an AI advantage.

How many tests have you done, just out of curiosity?

Link to comment
Share on other sites

I was not even trying to test this, was just trying to make sure the scenario I am making can be won from red side, and from blue, so that I can adjust each one's VPs accordingly to give each a shot. I had done 5 from each when I noticed this trend and posted the topic...have done 5 more from each now..so a total of 10 from red,10 from blue, and overwhelmingly this has happened..never once from red have I gotten a shot off, and only twice from blue now have I managed to see them before THEY got their first shot off.

Link to comment
Share on other sites

The suggestion is that the AI gets a spotting advantage whatever unit we are talking about? So AI controlled RPG team spots and targets much quicker than the same team controlled by a human; and AI controlled Abrahms spots and targets much quicker than the same tank controlled by a human.

I tested this out in 'ATGM Ambush' on elite and without changing the set-up positions, and found there was no difference in the time it took for the Red ATGM to take out the leading Blue Bradley whether Red was controlled by me or the AI. Also, when controlled by me the Bradley's only spot the ATGM after it has fired. When controlled by the AI the Bradley's do not fire on the ATGM, even after it has fired, so I assume they have not spotted it.

So if there is an issue/bug it could well just be situation-specific rather than applying to the AI as a whole. Lots of testing would need to be done to prove this one way or the other.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Unfortunately, your content contains terms that we do not allow. Please edit your content to remove the highlighted words below.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...