Originally posted by CommC Joel,
I have sent you a save where unescorted IJN Nells attacked a TF protected by CAP. This is from the Coral Sea scenario #1. Human Allied, Computer IJN.
Here is the combat report:
Air attack on TF, near Cairns at 7,56
Japanese aircraft
G3M Nell x 9
Allied aircraft
Wirraway x 9
Japanese aircraft losses
G3M Nell x 2 destroyed
G3M Nell x 7 damaged
Allied Ships
CV Lexington, Torpedo hits 2, heavy damage
Attacking Level Bombers:
4 x G3M Nell at 200 feet
0 x G3M Nell at 200 feet
3 x G3M Nell at 200 feet
Hope this helps.
Thanks,
CommC
THIS DOES NOT HELP! It's not even an example of a poor target choice on the part of the AI. Losing a few Nells to put 2 torpedoes into a US carrier is a great swap for the Japanese.
If your Nells had ignored the CV, and instead attacked a transport and a couple of Sub Chasers covered by base CAP at Cairns, then you would have an example of a problem worth sending to 2by3.
Granted, this is not exactly what we are looking for, but it is an example of unescorted bombers attacking a naval target protected by CAP. Apparently, this did not happen in versions before 2.2.
Whether this is a bug or not is another discussion or which AI behavior is better, I don't know.
I think most would agree, if the AI chooses to risk bombers by attacking unescorted a capital naval target protected by some light CAP, thats OK. What we are looking for is unescorted bombers attacking a minor naval target protected by heavy CAP. As I said before, I haven't seen this yet, but I'll keep an eye out for it and snag the save if I do. Obviously, some people have reported seeing this.
Originally posted by CommC I think most would agree, if the AI chooses to risk bombers by attacking unescorted a capital naval target protected by some light CAP, thats OK. What we are looking for is unescorted bombers attacking a minor naval target protected by heavy CAP. As I said before, I haven't seen this yet, but I'll keep an eye out for it and snag the save if I do. Obviously, some people have reported seeing this.
The question the should be asked, is how did the IJN AI 'know' that the CAP was light? I mean, from the time the Lexington was spotted to the time the strike was launched from Rabaul, to the time the bombers actually reached the Lex, the situation reported by the recon plane could have changed dramtically.
So is the descision to launch dependent on the aggressiveness of the Squadron leader ? Attacking the Lex at such a distance would be quite a risk. Or is that the Nells launched regardless of wheather there was heavy CAP or not (not something they could know) and were simply lucky that upon arriving at the target CAP was light?
Or are targeting descisions made completely differently? Like if you put your Nells on Naval attack is it assumed that they are airborne on 'sweep' mode looking for ships to attack, and having spotted the Lex and noticed there was light CAP decided to attack? An on the spot descision as it were?
In time of war the first casualty is truth. - Boake Carter
Seems to be a fair amount of confusion over what constitutes
a STUPID attack as opposed to just a RISKY one. Some things
to consider...
Your reccon aircraft WILL get a fairly accurate picture of the CAP
over a TF. To do their jobs, they have to maintain contact with
the TF for a period to determine it's base course and speed (it
might be doing a zig-zag---no good reporting it's course as 180
if it turns to 140 10 minutes later.) During this time the scout
will normally get a reasonable notion of the CAP strength (if any)
from trying to dodge it---and as CAP is mounted on a rotating
basis it won't change greatly in most instances.
No matter what your preference, if your reccon spots enemy CARRIERS you are almost assuredly going to TRY to strike
them. The AI considers CV's to be TOP priority, and will try
more desperate measures to get a shot at them! This "design
choice" is NOT currently "on the table" and will remain for the
forseeable future. Only problem they may look at here is the
AI's seeming reluctance to match potential escorts up with
such strikes. The number of complaints seem to indicate that
there MIGHT be a problem in that regard.
STUPID comes into play when something like this happens. You
spot a TF (let's make it a CV TF) that might be coming towards
Moresby, so in your next turn you put all Moresby's bombers on
naval strike, with half the fighters on escort. Suprise..., you
guessed right and it actually comes toward you during the night!
You watch eagerly as your PBY's pick it up again, and your naval
strikes start launching. And then you get reports of this kind of
nonsense:
1) Your SBD's attacked the Carrier TF, but no fighters went
along, and they were totally shreaded by the CAP.
2) Your B-17's ignored the CV TF and decided to bomb a
couple of transports in Rabaul Harbor---covered by 50 Zeros.
3) your 2 squadrons of B-25's decide the really vital target
is a DD with some MSW's off Lae. One of your escort units
agrees, and they merrily beat up this unfortunate force with
minor losses. NOTE that this is NOT A MISTAKE. The target
they chose was 90 degrees away from the bearing to the
CV TF---they didn't "miss" it, they weren't even looking for
it.
4) Your A-20's actually go to the carriers and bring some of
the escorts with them. They manage to get a hit on a CA
while taking heavy casualties.
5) ..But your Hudson's head to the Shortlands to make an
extended range strike on a tanker and some escorts CAP'd
by some Rufe's. They are driven off with some 40% loses
and a single hit to show for it.
6) Your Beauforts make a legitimate mistake and attack some
Transports in a TF following the CV's. They do some good and
take some loses.
This is the kind of thing that is worrysome. The AI often can't
decide what is a legitimate and dangerous target NOW! And
it scatters it's efforts all over the map on things that could be
safely ignored or put off until tomorrow.
It's other big candidate for the STUPID label is that it seems to
feel it HAS to make a "Naval Strike" if any TF is spotted within
range. The following would be an example:
As the Japanese Commander, you feel that the US carriers
might be about to launch an attack on Gili-Gili or the Shortlands.
So you manuever your CV TF southeast from Rabaul to a
position that you can strike this force from if it shows up. You
have your Val's resting and doing ASW, your Zeros are on CAP
with some allocated to escort "just in case". You don't expect
to sight anything yet, as you are moving to gain a position of
advantage---but just in case you see something worthwhile or
have miss-timed the US effort, you have your Kate's on "naval strike". At first light, a float plane from Buna sights an Allied
CA force in Moresby. With New Guinea between, it's no threat
to you, and can't be one today. But because Moresby is just
within the extended range of your Kate's, guess what happens.
If you guessed 50 dead Kate's ground to bits making a suicidal
and unescorted strike against THE ONLY PLACE IN THE WHOLE
THEATRE YOU KNOW IS CONSISTANTLY COVERED BY SWARMS OF
CAP..., "you win a cookie."
That's the other big problem. The AI doesn't seem to be able to
descriminate between actual threats/opportunities and doing
something suicidally stupid. It's supposed to do this, and it
isn't working. That's the kind of report they are looking for---
ones that will help them see if there's a bug that is allowing
either "attacks that make NO SENSE" or ones that splatter the
effort all over the Pacific when there is only one REAL THREAT to
be dealt with. If you have examples like this---please send them.
This is effectively a repeat of the test game I ran a few days ago. I am sending the save to Joel but I thought I would put the AAR on the public forum for people to see.
Situation.
Allied base at Port Moresby. 1 squadron of F5 a/c flying recon over Lae every turn (1-2 a/c at 20,000 ft). 2 squadrons of PBYs on 100% naval search at 6000 ft. 1 squadron of B25s (exp 70) on naval attack. No escort provided.
Japanese base at Lae with 4 A6m2 squadrons (on 50% CAP) at 9000 ft. 1 Japanese Transport TF in Lae harbour (2 DD, 2 PG, 4AP), 7 hex range from Port Moresby. 1 Japanese Transport TF at sea (2DD, 4AP), 7 hex range from Port Moresby.
Targets are virtually identical, range is the same. Recon missions over Lae were flown the day before and also on the same day as the Naval Strike (and was intercepted by a large CAP each time). Both targets were spotted several times by allied naval search during the AAR turn.
The AI routine sent 2 large strikes against the Lae TF and one small strike against the TF at sea.
AFTER ACTION REPORTS FOR 08/07/42
Weather: Partly Cloudy
Air attack on TF, near Lae at 9,33
Japanese aircraft
A6M2 Zero x 45
Allied aircraft
B-25D Mitchell x 9
no losses
Allied aircraft losses
B-25D Mitchell x 3 damaged
Japanese Ships
AP Asakasan Maru, Bomb hits 1
PG Nikkai Maru
AP Mogamigawa Maru, Bomb hits 1
Attacking Level Bombers:
4 x B-25D Mitchell at 6000 feet
1 x B-25D Mitchell at 6000 feet
4 x B-25D Mitchell at 6000 feet
--------------------------------------------------------------------------------
Air attack on TF at 14,34
Allied aircraft
B-25D Mitchell x 3
Allied aircraft losses
B-25D Mitchell x 1 damaged
Japanese Ships
DD Shiratsuyu
Attacking Level Bombers:
3 x B-25D Mitchell at 6000 feet
--------------------------------------------------------------------------------
Air attack on TF, near Lae at 9,33
Japanese aircraft
A6M2 Zero x 44
Allied aircraft
B-25D Mitchell x 12
no losses
Allied aircraft losses
B-25D Mitchell x 3 destroyed
B-25D Mitchell x 11 damaged
Japanese Ships
DD Asanagi
AP Asakasan Maru, Bomb hits 1, on fire, heavy damage
Attacking Level Bombers:
2 x B-25D Mitchell at 6000 feet
2 x B-25D Mitchell at 6000 feet
4 x B-25D Mitchell at 6000 feet
--------------------------------------------------------------------------------
Actual allied a/c lost : 4 B25s dest, 10 damaged (from 16 a/c). Morale is now 46. You could imagine what would have happened to Hudsons.
As allied medium bombers do, they made the TF in harbour pay.
This is exactly the kind of "brainless stupidity" that needs to
be ripped from the AI's heart. Two identical targets, at an
identical range, with only one (the one at sea moving) a threat
to do anything NOW! Had the total effort of the B-25's been
directed at the sensible target, it would have been pretty much
"trashed" with relatively light losses. Leaving the B-25's still
operational the next day in case the force at Lae threatened
to actually DO something.
Instead the moronic AI sends most of the effort against a well
protected force THAT ISN'T DOING ANYTHING! And now if there
is a threat tomorrow, the player has nothing of consequence
left to meet it. And they wonder that the player's get upset
and want some stronger control of their missions.
Stronger control of "naval strikes" IS NOT going to happen, for
a number of design choice reasons. But 2by3 is upset by the
way the AI is performing as a surrigate commander. They want
to make the AI stronger and more descriminating in it's target
choices for both sides. Which is why efforts such as your's are
important. They provide the raw data to see why the AI is
functioning like a brain dead hamster on drugs instead of
making the rational choices it's supposed to be making...
I've got a game against IJN as computer, scen 17, where in Mar 1943 it decides to send raids of unescorted Bettys and Nells down from Truk to attack my 5 CVs in 3 task forces sitting at whatever the base just south of Buka is. Of course this group of CVs had over 100 wildcats on CAP and shredded the 6ish odd strikes that came it's way. I don't think any one raid had more than 30 bombers in it. Think the IJN had lost over 100 bombers in the end - they didn't even turn away even when taking horrible casualties (aren't they supposed to do that now). I don't have an exact save but could go back and recreate from a save a few turns ago if this is of value to Matrix/2x3. I even think there was a nice juicy unescorted cargo TF sitting at Buka at the time. On a final note the very next day the AI sent whatever penny packets of Bettys and Nells down again (6-10 planes each) - you would think the AI would have got the joke by then.....
Please let me know if I should recreate or if this is not what is sought.
Originally posted by fcooke File should be attached - not quite as lopsided as original but pretty bad. If there's anything else you need, please give a shout.
jrcar, did the TF at 38,48 include any carriers? (I assume answer is yes, given the strong Wildcat CAP).
Then, that second strike with just 7 Zeros and 10 Bettys would be completely justified in my own opinion, and is just the kind of 'sacrifice attack' that a Japanese war-time commander would order in the same circumstances.
A single torpedo hit on an enemy carrier would have greater influence -by far- in the war situation that the loss of some 17 aircraft. The risk is justified; who knows if the CAP will be found napping, or if a single Samurai-minded pilot will be able to crash himself on a flat-top?
Just remember the last strike by the survivors of HIRYU air group at Midway; by any logic, that pitiful handful of planes should have been butchered by the CAP, however it was enough to finish-off YORKTOWN.
Of course, I agree with you that such an attack is not always justified by the target's size or potential threat; but against carriers or battleships? Yes, by any means.
