Weak enemy tank AI

[Ron]

Originally posted by JasonC:

Starting a dozen tanks in LOS of a pair of shooters and charging to 50m away (1) isn't realistic at all and (2) won't reveal the problem. Cower occurs at the instant LOS is first achieved. You see it when "breaking from cover", when an enemy crests a hill, etc. If you start the whole thing already in LOS, you will not see it. It is also probably more common on "hunt" orders.

...I don't call losing 6 T-34s to kill 2 Tigers in a 9 on 2 "skill", I call it a "suicidal exchange". It may sometimes be warranted by the tactical situation, if nothing better is possible, as I carefully allowed.

Well either my writing skills are poor or your reading comprehension is. Your first paragraph doesn't remotely resemble what developed.

The T-34s weren't in the 'open' to begin with and didn't charge mindlessly towards the Tigers. The envelopment took several minutes using as I said what little cover there was. Only one T-34 was lost in the approach before launching the attack, the rest after it had begun.

As for some how relating the "losses" to skill... Sure after discovering the situation I could have simply bailed or retreated my force, but this is a game the last I heard and while the situation was highly unfavourable for the Russians, the most was made of it.

Ron

[JasonC]

You said 800m, which is a reasonably long way to close, very open, little cover, 9 T-34s start. You said only 4 made it close, which means 5 did not make it close. You said 1 of the close ones cowered which means 3 did not cower. Only 3 survived the engagement.

The obvious reading is the ones that lived long enough to get close and actually delivered their shots survived. Why obvious? Because they are known to have lived that long, and once they do deliver their shots the number of live Tigers ought to be dropping, and from 2 to 0 is not far to drop.

A last shooter, at the least, from among those that got close had to survive the engagement to win it. The most obvious way every one of those fits together is 5 burning wrecks along the approach route or performing distraction and 1 dead by the Tigers.

It can go as far as 3 and 3 without explicit contradiction of any of the statements made, though nothing explicitly said suggests as much. No reason is given why the others would not have closed.

It was not stated that only 1 T-34 remained alive near the Tigers, which would mean an outcome that could easily have gone the other way. Which would have been relevant and wasn't mentioned. So there was no reason to suppose half of the dead tanks were right next to the Tigers.

2 or more could be dead distractors around the same time as the close engagement, rather than killed earlier while closing. I could not read it as 1 at range and 5 dead close since that implies 6 or more got close, which directly contradicts the statement that only 4 did so.

With ranges all under 800m and it specifically said there was little cover (which rarely suffices to continually hide 9 fast moving tanks continually for a long period), there was no reason to suppose only 1 loss in the approach.

It takes 3-4 minutes to drive T-34s to close range from 800m away. 2 Tigers can fire several dozen of times in that period. Even at moving targets, shots so close do not usually all (or nearly all) miss.

You can put it down to my reading or to your writing, I really don't care...

[Lady Roxanne]

Jason,

In your "keyholed" Tiger example with the T-34 coming around the woods for a rear shot, I would have ordered the T-34 as follows:

1) Fast Move waypoint to a position behind the Tiger (obviously with LOS).

2) armor cover arc that will see the turret facing toward the Tiger when the waypoint is reached.

3)Button Up!! The combination of Fast Move and being buttoned often causes the TC to not see the scary Tiger the instant LOS is achieved, so no cowering.

Although I'm not sure (I'd have to test it), my guess is that the T-34 would have gone all the way to the waypoint set, with the turret facing toward the Tiger, and your Tiger would be history.

Once stopped at the waypoint with the turret facing the tiger, the T-34 TC WOULD see the tiger, even though buttoned, and the mission would be accomplished.

Edit: I tested this. About 5 seconds after the T-34 got to the waypoint I plotted, it spotted and targetted the tiger. It also started retreating. While reversing, the T-34 fired one shot. I would have preferred if it would have fired and THEN started reversing back out of LOS.

On the bright side, the T-34 did make it to the waypoint I plotted even though well past the point of LOS; and it DID fire a shot.

Kris

[ May 12, 2003, 01:18 AM: Message edited by: CrankyKris ]

[Lady Roxanne]

I did it!!

At 100 meters the T-34 killed the Tiger with a shot to the rear. The previous test was at 300 meters.

Other differences:

I did NOT button the T-34 TC. I did NOT use a cover arc. Instead, I targetted the Tiger specifically (sharpshooter spotting).

The T-34 rounded the bend, saw Tiger, buttoned up(2-man turret), targetted while completing fast movement to waypoint. After stopped, he fired. He then fired again!! The only movement after arriving at the waypoint was some rotation of the hull. No cowering!!

Interesting, was how quickly the regular Tiger TC spotted the T-34, even though buttoned with the turret facing away from the T-34. I guess a non-moving vehicle can spot a moving one easily, even while buttoned.

Kris

[Mattias]

It worked the same way for my Cranky..

If the Tiger was "pre-targeted" the T-34´s went through with the attack, if not they would cower.

(100x100 patch of woods, Tiger at base, the T34's running down the sides)

The Tiger was destroyed.

M.

[JoePrivate]

Originally posted by JasonC:

You said 800m, which is a reasonably long way to close, very open, little cover, 9 T-34s start. You said only 4 made it close, which means 5 did not make it close. You said 1 of the close ones cowered which means 3 did not cower. Only 3 survived the engagement.

The obvious reading is the ones that lived long enough to get close and actually delivered their shots survived. Why obvious? Because they are known to have lived that long, and once they do deliver their shots the number of live Tigers ought to be dropping, and from 2 to 0 is not far to drop.

A last shooter, at the least, from among those that got close had to survive the engagement to win it. The most obvious way every one of those fits together is 5 burning wrecks along the approach route or performing distraction and 1 dead by the Tigers.

It can go as far as 3 and 3 without explicit contradiction of any of the statements made, though nothing explicitly said suggests as much. No reason is given why the others would not have closed.

It was not stated that only 1 T-34 remained alive near the Tigers, which would mean an outcome that could easily have gone the other way. Which would have been relevant and wasn't mentioned. So there was no reason to suppose half of the dead tanks were right next to the Tigers.

2 or more could be dead distractors around the same time as the close engagement, rather than killed earlier while closing. I could not read it as 1 at range and 5 dead close since that implies 6 or more got close, which directly contradicts the statement that only 4 did so.

With ranges all under 800m and it specifically said there was little cover (which rarely suffices to continually hide 9 fast moving tanks continually for a long period), there was no reason to suppose only 1 loss in the approach.

It takes 3-4 minutes to drive T-34s to close range from 800m away. 2 Tigers can fire several dozen of times in that period. Even at moving targets, shots so close do not usually all (or nearly all) miss.

You can put it down to my reading or to your writing, I really don't care...

I was the opponent in the game versus Ron and a very nailbiting one at that I am not sure why the assumptions as I think the main point trying to be made was the T-34s actually fought instead of retreating.

While it is neither here nor there, some clarifications: the "approach" of his T-34s took 2/3 thirds of the game while the final flanking manuevers and attack only took 4-6 minutes tops, my Tiger's did get a T-34 early on but the rest of the KO's happened while attacking, all under 200m, a hidden ATG accounted for one T-34 as well. The Tigers took a fearful pounding - both immobilized and at least a dozen penetrations before the crews abandoned. I admit I was lucky to have them survive as long as they did and I doubt if I could have done any better were the roles reversed.

One thing to keep in mind with this "cowering", if you can get an overmatched tank into position where the game "thinks" it has a chance then it won't retreat. In the case of T-34s versus Tigers that means directly perpendicular to the Tiger's side preferably, or the rear. Any sort of angle and the T-34s will retreat, so I guess there is some sort of "skill" in that.

Good game Ron, hope for a rematch sometime

[aka_tom_w]

"One thing to keep in mind with this "cowering", if you can get an overmatched tank into position where the game "thinks" it has a chance then it won't retreat. In the case of T-34s versus Tigers that means directly perpendicular to the Tiger's side preferably, or the rear. Any sort of angle and the T-34s will retreat, so I guess there is some sort of "skill" in that. "

is the game engine really that sophisticated? :confused:

I don't have enough experience with the "cowar" behaviour in tanks that are under matched to know if it really does make a difference if the undermatched tank approaches from the rear or perpendicular to the flank :confused:

Can anyone else confirm if the game engine is sophisticated enough to rationalize or figure out that approaching an over matched tank from the flank or rear aspect will in fact suspend the "cowar" behaviour in the undermatched tank?

Jason C ? where do you stand on this issue?

this is interesting and in fact amazing if the game engine and the TAC AI are sophisticated enough to determine that the approach from the flank or rear will mean the TAC AI and will behave differently and not cowar?

How cool is that?

BUT is it really TRUE? :eek:

-tom w

[ May 12, 2003, 09:56 PM: Message edited by: aka_tom_w ]

[mchlstrt]

I guess we're trying to figure out how sophisticated it is. It would be helpful to post which patch people are playing with, if they know. BFC did something in this area for 1.0.3, I believe, so without the patch #s in their posts people could be talking apples & oranges. Hope so.

strt

[aka_tom_w]

Originally posted by mchlstrt:

I guess we're trying to figure out how sophisticated it is. It would be helpful to post which patch people are playing with, if they know. BFC did something in this area for 1.0.3, I believe, so without the patch #s in their posts people could be talking apples & oranges. Hope so.

strt

OK

Anyone else care to comment on just how sophisticated the engine is with regard to this "tank cowar behaviour" issue?

some interesting comments here

-tom w

[ May 13, 2003, 11:52 AM: Message edited by: aka_tom_w ]

[MikeyD]

The game doesn't have to be too sophisticated to decide whether to retreat or not.

A forward facing target falls within LOS. The computer reads your tank has a 2% chance of penetration at that range and angle while your opponent has a 100% chance of killing you. A no-brainer on what to do, the computer says.

Change the angle of the target to the vehicle so you've got an 80% chance of penetration vs 100% for the enemy. It's a fair fight now!

I swear I can almost hear my Russian tankers giggling with glee when they stumble upon a Stug arse-end to them. That same Stug with its gun (and 80mm armor) facing them is a whole 'nother kettle of fish.

Whatever problems may have been with 'cower', in my opinion the Beta v1.03 does a very much better job of handling it.

[ May 13, 2003, 01:04 PM: Message edited by: MikeyD ]

[aka_tom_w]

Hi MikeyD

Sounds like you are pretty sure that the cowar behaviour does not happen when an undermatched tank approaches from the flank or rear.

If that is the way it actually works in the game I think we can all agree that is the way it should be.

-tom w