Matrix Games Forums

After over a year and finishing a GC through 8/45 I had thought I'd seen everything the game could serve up. This was a new one.

Playing Allies against Japan AI, Historical, Scenario 2.

Bataan on 3/15/1942. Been fully retreated to in good order by Allies. Forts have been knocked down to zero, has been out of supply for over two weeks. No recon assets remain. No naval units of any kind. The AI has been doing mixed bombardment, regular, and shock attacks with a large force, and losing thousands of men. Allies gradually wearing down, etc., but disorg. and fatigue levels are generally good.

Then, this attack, with results I've never seen.

--------------------------------------------------------------------------------

Ground combat at Bataan (78,77)

Japanese Shock attack

Attacking force 2124 troops, 320 guns, 207 vehicles, Assault Value = 1383

Defending force 44436 troops, 785 guns, 730 vehicles, Assault Value = 1053

Assault collapses, survivors seek cover
Japanese ground losses:
159 casualties reported
Squads: 22 destroyed, 0 disabled
Non Combat: 33 destroyed, 16 disabled
Engineers: 9 destroyed, 0 disabled
Guns lost 1 (1 destroyed, 0 disabled)
Vehicles lost 44 (42 destroyed, 2 disabled)
Units destroyed 1


Allied ground losses:
123 casualties reported
Squads: 3 destroyed, 6 disabled
Non Combat: 0 destroyed, 8 disabled
Engineers: 0 destroyed, 0 disabled


Assaulting units:
65th Brigade
48th Division
10th Division
12th Division
19th Division
16th Division
7th Tank Regiment
19th Ind. Engineer Regiment
2nd Mortar Battalion
20th Ind. Mtn Gun Battalion
8th Medium Field Artillery Regiment
3rd Ind. Engineer Regiment
1st Medium Field Artillery Regiment
9th Ind.Hvy.Art. Battalion
20th Army
21st Ind. Engineer Regiment
Tonei Hvy Gun Regiment
22nd Medium Field Artillery Regiment
14th Army
1st Mortar Regiment
1st Hvy.Artillery Regiment
7th Ind.Hvy.Art. Battalion
15th Ind.Art.Mortar Battalion
5th RF Gun Battalion
10th Ind. Mountain Gun Regiment
3rd Engineer Construction Battalion

Defending units:
31st Infantry Regiment
4th Marine Regiment
2nd PA Constabulary Regiment
21st PA Infantry Division
91st PA Infantry Division
3rd/12th PA Inf Battalion
86th PS Coastal Artillery Battalion
1st PA Constabulary Regiment
51st PA Infantry Division
14th PS Engineer Regiment
11th PA Infantry Division
192nd Tank Battalion
26th PS Cavalry Regiment
4th PA Constabulary Regiment
194th Tank Battalion
71st PA Infantry Division
57th PS Infantry Regimental Combat Team
Manila Bay Defenses
41st PA Infantry Division
45th PS Infantry Regimental Combat Team
2nd PA Constblry HW Regiment
88th PS Field Artillery Regiment
803rd Engineer Aviation Battalion
Bataan USN Base Force
Cavite USN Base Force
II Philippine Corps
200th & 515th Coast AA Regiment
202nd PA Construction Battalion
USAFFE
Provisional GMC Gp
1st USMC AA Battalion
301st Construction Battalion
Far East USAAF
201st PA Construction Battalion
Asiatic Fleet
I Philippine Corps
1st PI Base Force
Manila USAAF Base Force
PAF Aviation
Clark Field USAAF Base Force
301st PA Field Artillery Regiment


--------------------------------------------------------------------------------

Clearly, the list of attackers didn't all attack. Perhaps the AI set all but one on Reserve. Perhaps it doesn't even use those settings. I don't know. But it threw a small unit to its death. I've never seen the AI make a move this bone-headed, and I've never, ever seen that text line "Assault collapses, survivors seek cover." The AI is a lot of things, but bad at calculating risks isn't one of them. The scripts have limits, but they don't make math errors, or get lazy, or have emotional responses to previous failure.

I'm sure there are randoms involved, but still, very interesting. The AI can still surprise me.

Interesting. Note how few casualties the attacker has. Indicating that the attack was broken off almost right away. I suspect very low morale or high fatique to be the issue. I actually would like to see that happen more often when you do an attack and the odds turn out to be too high.

It's not that uncommon a result. I've seen it about 6-8 times in 2 GC's. And I have a city in China where the AI regularly attacks me at 1:30-60 odds with a portion of its army - just about every turn.

My latest game is about the same as Moose's. It's Mid April and the PI units are about out of supply but still forcing the IJA into foolosh attacks at Clark where they lose thousands of men at very poor odds. I fear that I'll be toast by the end of the moth once supply hits zero for just about everybody. I'm still moving a small trickle of supplies in by sub, but much more is needed than I can provide.

My hope is that the IJA units will be so depleted that they will not be available elsewhere for many months. It seems like the thrusts into the DEI have been weaker than usual, so maybe I'm seeing that already. I don't know. But, I'm past the landing bonus, so if the IJA/N troops aren't already on a land mass, they won't be able to grab any large land mass quickly. They're on Java, but have only taken unoccupied bases. Those where I've concentrated by defenses are holding out, at least for now. I hold Koepang and Lautem in force, so there's no way they'll get those.

I've also found that, in China, I can use the tactic Moose describes very effectively. I mass a few corps, maybe six or seven fairly strong ones, push a couple of IJA units out of a hex, then wait for them to move back in and attack. I then attack again the next after and they really take casualties. The AI just keeps on doing the same foolish thing over and over. I've also learned to be patient and surround IJN units in China over a few months before attacking with a relative overwhelming force. After a few turns, the IJN units are gone and I move on to the next location. It takes several years, but I can ususally retake most of SE China that way. It doesn't really do much more than keeping Japan from expanding its holdings and giving me a few extra VP that I don't really need. But, at least I can avoind any big push towards the interior.

ORIGINAL: crsutton

Interesting. Note how few casualties the attacker has. Indicating that the attack was broken off almost right away. I suspect very low morale or high fatique to be the issue. I actually would like to see that happen more often when you do an attack and the odds turn out to be too high.

Yes, low casualties, but over 60 lost squads. Something doesn't add up there.

Edit: unless they only had a couple of men each. In which case, the AI shouldn't have attacked with them.

ORIGINAL: wpurdom

It's not that uncommon a result. I've seen it about 6-8 times in 2 GC's. And I have a city in China where the AI regularly attacks me at 1:30-60 odds with a portion of its army - just about every turn.

FWIW , this attack come off at 1:2 odds in the Allies' favor. There had been several days of 1:1 odds with regular attacks. As before, Allies at zero forts and zero supplies. In the regular attacks the Allies have been losing about 50% of the casualties of the Japanese. Not irregular, given the supply status/on-the-defense combo.

From an AV POV, the AI doesn't look TOO insane, except for the Shock attack piece. But why list the whole stack, and have only one, small unit actually go in to utter destruction?

The AI Bombarded the next turn, then Regular attacked the turn after, with 45,000 men. Bataan continues to hold.

Edit: my memory is faulty. There ARE no odds reported on this attack. or +/- results. I looked at all of the other ground actions that day, and only the Bombardment attacks do not have odds listed. All of the Deliberate attacks elsewhere do. More strange. Was this really a Shock attack?

I remember adding that text to the code.  I believe it happens if an attack has worse than 1:99 odds.  (I can't remember for sure at the moment.) 

It does sound like the AI is behaving stupidly.  I don't know much about how the AI works, so I won't comment on that.

Bill

ORIGINAL: wdolson

I remember adding that text to the code.  I believe it happens if an attack has worse than 1:99 odds.  (I can't remember for sure at the moment.) 

It does sound like the AI is behaving stupidly.  I don't know much about how the AI works, so I won't comment on that.

Bill

Excellent input! Thanks.

Playing detective (always dangerous with code one can't see) I wonder if the lack of an odds print-out, despite the report of a Shock attack, means that the attacking Japanese never got past the prep bombardment/counter-fire phase before the 1:99 odds occurred and fired the text. The lack of odds (never seen with a Bombarment attack type), combined with a reported Shock attack, is interesting.

I had never noticed until today that bombardment attacks don't spit out odds. Some of the forum membership may want to offer me a bridge property or two for my purchase consideration.

The way ground combat works is there is a phase where the units shoot at one another, then the odds are figured.  Most of the casualties happen during the shooting at one another phase, but the odds determine retreat, base capture, etc.  I just remembered that the assault collapses was added for the situation when an attacker is wiped out by the shooting at one another phase.  Odds can't be calculated because there is nothing left with which to calculate the odds.

Bill

I like the little "gems" the devs flavored the reports with. This one caught my eye while browsing through the ops report a couple of days ago from my PBEM.

xAK Toho Maru hits reef at 132 , 115

Turns out FOW got the wrong ship but an AMc nearby had sustain a bit of float damage while it plied the shallows.

It's pretty neat and unexpected IMHO (although I'd probably been more perturbed and less fascinated had it been one of my carriers running aground!). It is just another example why AE has dominated my gaming time more than any other! [:D]

ORIGINAL: wdolson

The way ground combat works is there is a phase where the units shoot at one another, then the odds are figured.  Most of the casualties happen during the shooting at one another phase, but the odds determine retreat, base capture, etc.  I just remembered that the assault collapses was added for the situation when an attacker is wiped out by the shooting at one another phase.  Odds can't be calculated because there is nothing left with which to calculate the odds.

Bill

That makes more sense than my theory.

Still wonder why just that one unit was thrown to the wolves.