PO Discussion

[Menschenfresser]

If anyone cares...other than Ralph.

PO movement has come a long way. I just finished a game of Bob Cross' Soviet Union 1941 as the Soviets. The PO did a good job overall when it came to spreading his units out and attacking with some degree of intelligence. In other words, he did hit some weak spots or by-passed roadblocks.

On the other hand, his biggest flaw presently, according to what I've seen, is his decision-making skills as to when and where to attack. No, it's simpler than that.

First, he seems to do a good job getting units up to the line and in proper proportions. But in both this scenario and in Trotsky's War, the PO lost the game by repeatedly making attacks that cost him horribly. Turn after turn he made attacks against the same stack where he'd lost 20-40% before. In both games, I've stopped playing because the PO has killed himself without any effort on my part.

I'm not sure what's going on here. He doesn't get a lot of rounds normally. Usually two. His attacks also result in evaporations for either him or me. Both of these aspects make me think he's using high loss tolerance settings. A human who has just attacked a stack of questionable strength and lost 20% to the enemy's 3% would try and think of something else to do. The PO will attack again until he's evaporated all his units and then he'll move up the parent HQ and start attacking with that.

Yes, I know the general response to posts like this: play PBEM, 'what do you expect?', etc.

I don't know. But I'd like to recommend something I've recommended before. If it isn't being done already, let the PO cheat. Not with a plus to combat, but let him be a chess program. Let him run proposed attacks. Give him the ability to run the results of the attack he's planning. Allow him to run it with various settings. Or with various units in the hexes surrounding. How many rounds does it burn? Losses taken and inflicted? I realize this would require logic for the PO to determine what is and isn't acceptable. But it would already be a bonus if the PO, analyzing the attack of, say, a tank regiment, would run 10 test attacks against both enemy hexes he faces. And, of course, attack the one with the better average result.

I'm not sure what's causing the PO's suicidal behavior. I realize these are scenarios where the PO is on the hot and heavy offensive. Probably set up for attack by the designer. But he seems to lack any sense of force preservation. Thoughts?

[jmlima]

ORIGINAL: Menschenfresser
...Yes, I know the general response to posts like this: play PBEM, 'what do you expect?', etc.
...

If that's the only answer this gets, they might as well advise you to go and play board wargames...

The PO does lack a fundamental understanding of 'big' issues. It lacks any strategic skills, and probably because of that fails completly to evaluate the issue of losses in an overall level. It all seems to be treated at a tactical level, which is kind of funny in a game branded as 'operational'... [8|]

My view, there is very little purpose in playing the PO in these big strategic games, or any scenario requiring sensible maneuvering like the desert, other than to see if the scenario is working in relation to events, etc.

[Menschenfresser]

ORIGINAL: jmlima
My view, there is very little purpose in playing the PO in these big strategic games, or any scenario requiring sensible maneuvering like the desert, other than to see if the scenario is working in relation to events, etc.

True. That's why I list Barbarossa and a hypothetical Russian invasion as examples. Both have tons of units for the PO and very little need for complex maneuvering. In fact, that's all I need it to do. Move forward, evenly, and attack, sensibly.

[jmlima]

ORIGINAL: Menschenfresser
... In fact, that's all I need it to do. Move forward, evenly, and attack, sensibly.

For that it needs the operational perception of threat to flanks, force preservation, schwerpunkt, timing... [;)]

None of that it's there.

[Curtis Lemay]

In "Soviet Union 1941" the German PO is set to Berzerk Strategic Bias and all formations are set to Ignore Losses. That was standard procedure for the aggressor PO in ACOW. Norm's PO was a really timid thing. Perhaps Ralph has skewed that by now, though, and scenarios will need some re-calibration. Regardless, this is a very hard scenario to calibrate the POs since the Soviet side can't be played by the PO without human help in disbanding the manpower levy units. That means no over-night PO vs. PO tests.

It does seem like, if the PO is only getting a couple of rounds per turn, then it is under a serious disadvantage. If that could be improved it would obviously help. But my long held position is that the PO is a black hole, down which an unlimited amount of coding effort could be pored without any discernible effect. Ralph feels otherwise, though, and I suppose it seems he has made some progress with it.

[Menschenfresser]

I don't think so. I'm not asking it to pick the best avenue of attack. Or even defend. Just not attack the same stack that gave it 40% losses and took almost none. Blindly shifting it's attack one hex to the south or north would be preferable.

Total pipe dream, I know.

[ralphtricky]

ORIGINAL: Menschenfresser

I don't think so. I'm not asking it to pick the best avenue of attack. Or even defend. Just not attack the same stack that gave it 40% losses and took almost none. Blindly shifting it's attack one hex to the south or north would be preferable.

Total pipe dream, I know.

Not really, one of my goals with Elmer is to have him not do stupid things. He may not do brilliant things, but I won't want him to do stupid things. People don't notice average play, they do notice obviously stupid play[X(].

He is supposed to attack with at least 4:1 odds, or 3:1 after round 7 or 2:1 after round 5. I don't believe he takes artillery/air support into account, which may be a factor. I'll have to try the Soviet Union scenario and identify a couple of places where he does stupid things and see why. I think I've got enough logging to be able to tell, if not, I'll add more. I'll see what I can do for 3.5 after I put in the major features.

I don't think there's any reason he can't become competent at tactical play, strategic play is going to be beyond him for a long time.

[FJeff]

As a relative noob, which WWII scenarios do provide a decent Elmer?

[larryfulkerson]

ORIGINAL: FJeff
As a relative noob, which WWII scenarios do provide a decent Elmer?

I'm playtesting Autumn Fog right now and I'm really impressed with Elmer. There are minor problems of course, it IS Elmer after all, but those problems are those he has with every scenario, more or less. But for the most part it's an accurate contest.

[Menschenfresser]

Not many...but I recommend trying:

Ardennes44 as the Germans
2 Weeks in Normandy as the Germans
Gothic Line 44 as the Germans

Anything where the PO has straightforward goals and lots of units to achieve it. Be prepared to fight with one hand behind your back. Don't get clever or gamey.

[Curtis Lemay]

ORIGINAL: FJeff

As a relative noob, which WWII scenarios do provide a decent Elmer?

I can recommend a few of mine:

The German PO in "Germany 1945"
The German PO in "France 1944"
Either side in "Cambrai 1917"
The Japanese PO in "Okinawa 1945"

[berto]

ORIGINAL: Curtis Lemay
The Japanese PO in "Okinawa 1945"

Has this been recalibrated for 3.4? (Or will the scenario continue to work well, unmodified, despite the changes in 3.4?)

[jomni]

Can someone teach me how to use PO and make playing simpler?  I have the impression that I can delegate objectives to some formations but I'm kinda lost.
Or PO is just an AI script for the whole side and I shouldn't be playing with it.

[Obsolete]

When it comes to adding CHEATS, whatever you do please don`t go the Firaxis cop-out route.  Such as removing FoW for the AI and so on...

[ralphtricky]

ORIGINAL: Obsolete
When it comes to adding CHEATS, whatever you do please don`t go the Firaxis cop-out route.  Such as removing FoW for the AI and so on...

He does know the actual equipment so he knows if he's got reasonable odds for an attack even for an unknown unit. I want to eventually change to using probing attacks to make them known, or remembering the last known values or something, but that's going to take a while to implement, I want to make him stronger first, and add in a strategic layer.

[Menschenfresser]

ORIGINAL: ralphtrick
He does know the actual equipment so he knows if he's got reasonable odds for an attack even for an unknown unit. I want to eventually change to using probing attacks to make them known, or remembering the last known values or something, but that's going to take a while to implement, I want to make him stronger first, and add in a strategic layer.

From my point of view, I'd rather think of Elmer on the level of someone who's played a scenario enough times to know it well and maybe even viewed it in the editor. In other words, his equivalent human player has a good knowledge of the equipment in any given enemy unit based solely on previous play-throughs. That said, I personally see no reason to restrict his a priori knowledge of enemy equipment.

[ralphtricky]

ORIGINAL: jomni

Can someone teach me how to use PO and make playing simpler?  I have the impression that I can delegate objectives to some formations but I'm kinda lost.
Or PO is just an AI script for the whole side and I shouldn't be playing with it.

Tha beta download includes an Elmer tutorial in the documentation folder. The AI uses a combination of scripted paths and settings.

Please take a look and post any questions.

Ralph

[jomni]

ORIGINAL: ralphtrick

ORIGINAL: jomni

Can someone teach me how to use PO and make playing simpler?  I have the impression that I can delegate objectives to some formations but I'm kinda lost.
Or PO is just an AI script for the whole side and I shouldn't be playing with it.

Tha beta download includes an Elmer tutorial in the documentation folder. The AI uses a combination of scripted paths and settings.

Please take a look and post any questions.

Ralph

Great. I'll go dig that up.

[Telumar]

It's well known that some soviet units went into fighting right out of the trains during Barbarossa / Typhoon. But Elmer seems to have made quite a habit out of this. I am playing D21 (the version from 3.4 Beta) and he repeatedly has railed units directly into hexes adjacent to a german unit. Sometimes there is a soviet unit in a such a rail hex which could provide some cover, but sometimes Elmer doesn't even seem to care about this.

I can provide a .sal if required.

[ralphtricky]

ORIGINAL: Telumar

It's well known that some soviet units went into fighting right out of the trains during Barbarossa / Typhoon. But Elmer seems to have made quite a habit out of this. I am playing D21 (the version from 3.4 Beta) and he repeatedly has railed units directly into hexes adjacent to a german unit. Sometimes there is a soviet unit in a such a rail hex which could provide some cover, but sometimes Elmer doesn't even seem to care about this.

I can provide a .sal if required.

I'll put it down to look at in 3.5. Right now, he goes to the last friendly owned objective, I think.