Anemic Pearl Harbor Strikes

[mdiehl]

Please correct me if I'm wrong, but in WitP a ship "sunk" is not refloatable, same as GGPW, right? So, "damaged, on fire" is the proper result for three of the USN BBs historically sunk at Pearl Harbor (but later refloated), and "Sunk" is the proper WitP result for the one USN BB at PH that was written off completely (USS Arizona). Or arguably two "sunk" (Arizona and Oklahoma) since the latter was refloated but not returned to service.

So the historically correct result in WiTP should be one or two USN BBs sunk, 2 DDs sunk, 1 minelayer sunk, a whole bunch of a.c. destroyed, and a whole bunch of ships damaged to varying degrees, some damaged to within inches of their electronic database "lives."

[pauk]

LOL, anemic attacks at PH happens in Allied edition too? [:D]

[Alfred]

ORIGINAL: Keldun

...I know we can play a dec 8 game but it would be nice to be able to reproduce the battle that started the war, wasn't it the purpose of the option of 7th dec surprise ?

No, that isn't the intent of 7th Dec Surprise.

The intent is to reproduce the unpreparedness of Allied forces anywhere in the PTO to meet the initial Japanese attacks. Even if you combine it with "Historical Start" you still don't get a faithful reproduction of the Pearl Harbor strike because of die rolls.

No, the only way you are guaranteed to replicate the Pearl Harbor strike is to play the 8th December scenario.

Alfred

[Lecivius]

I'm going to have to go on random rolls.  I refer you to my dated post http://www.matrixgames.com/forums/tm.asp?m=2850354 where the entire battle line in PH went under.  I had the latest bete patch as well.  It hurts, but the sheer randomness of the game is what makes it interesting.

[LoBaron]

ORIGINAL: Keldun

ORIGINAL: LoBaron
Before I forget: There were "only" 2 BBs lost at PH. So your best result is in
accordance to reality - at least considering what the term "ships sunk" on game
terms. [;)]

Actually there were 4 bb sunk in the pearl harbour attack but they managed to refloat 2 of them later.

I retried many times as well to reproduce similar result (with the 7th dec surprise option)but never managed even though after like 20 tries, i did get 3 bb sunk but the number of aircraft or smaller ship sunk wasn't anywhere close.
I know we can play a dec 8 game but it would be nice to be able to reproduce the battle that started the war, wasn't it the purpose of the option of 7th dec surprise ?

Thats why I wrote '"ships sunk" on game' terms. In game sunk ships are sunk. A BB with 80 major flotation could well be interpreted as capsized or
or run aground, and later refloated. [;)]

As for your question: In fact the dec 8th scenario was the purpose of reproducing the battle that started the war.
The dec 7th surprise just recreates the historical prerequisite.

Edit: Alfred I should have simply copy pasted your response... [8D]

[Keldun]

That certainly is overboard but there should be a good middle ground between that and very few aircraft and 0-1 bb sunk +small ships most of the time. Your result extreme but it is like someone getting 60 points from 10 dices with a single roll (probably super rare [;)]).

[Keldun]

ORIGINAL: Alfred

ORIGINAL: Keldun

...I know we can play a dec 8 game but it would be nice to be able to reproduce the battle that started the war, wasn't it the purpose of the option of 7th dec surprise ?

No, that isn't the intent of 7th Dec Surprise.

The intent is to reproduce the unpreparedness of Allied forces anywhere in the PTO to meet the initial Japanese attacks. Even if you combine it with "Historical Start" you still don't get a faithful reproduction of the Pearl Harbor strike because of die rolls.

No, the only way you are guaranteed to replicate the Pearl Harbor strike is to play the 8th December scenario.

Alfred

How about boosting a bit more damage done with the 7th dec surprise option to non flying airplanes (in ph, clarkfield,wake and maybe others?) and ships docked in ph on the 7th december so that we can get a result a bit closer to what it was at the start of the war a bit more often?
Or is there any reason why we shouldn't try to get a result close more often?
I' m not asking for the exact same faithful result, I' m just asking to be able to reproduce something close more often.
At the moment I see lecivius having an extreme result and the topic starter and myself seem to have always gotten much worse result after many tries.

[CaptBeefheart]

GoTraurig: Apologies if you mentioned this above, but did you actually look at the Allied side to determine the real losses? Fog of war can give you bogus results.

Cheers,
CC

[GoTraurig]

Commander Cody: Yes, I was doing my testings via hotseat and examined the results from both sides.

LoBaron: I didn't intend for this thread to be a complaint that my results were not historic. A historical result on any of the strikes or major battles is not necessarily expected or required.

Although, after rereading my initial post, I said the historical result was the standard. Using the "historical" result (2 BBs sunk) as the standard, i.e. the average result for the Japanese side in a PBEM, was my understanding of what to expect at the time. In retrospect, to be clear, I should have stated that almost all my strikes were weaker than what I perceive the average from looking through AARs and reading player discussions.

Everyone: My main purposes for this thread were to achieve some combination of: 1. To verify that I was setting up the strike properly. 2. To ensure I wasn't experiencing a bug. 3. To seek advice on minimizing the potential for a really bad result. 4. To learn what results should be expected. 5. To determine why my strikes were so much weaker than those I usually see in AARs and other discussions.

So far I have at least partially achieved 1, 3, 4, and 5, with my thanks and appreciation to you all. Currently my only related problem is ensuring more than half my Kates fly - I understand it could be just bad die rolls, but they were awfully consistent for a while.

I've also pretty much accepted that the solution, and all the advice I'll probably receive about this, is to ensure that the Kates have the proper orders and just hope they decide to fly.

Lecivius: That sure was an interesting strike. Thanks for bringing an outlier, on the other side of the results, for me to view - too bad you had to experience it.

[LoBaron]

GoTraurig, I did not see your post as a complaint.

My motivation to post the 'The art of "averaging out"' thread is much
older already, because misinterpreting historical results is a more or less common
occurance on the forum.

If there was any additional reason at all, it was Kelduns post suggesting to "boost a bit more damage"
to "so that we can get a result a bit closer to what it was at the start of the war" concerning PH.

This statement suggests a classic misconception of what should be a "common result", because the expected
average is based on viewing a single historical event as representative, simply because it happened that way.

[PaxMondo]

ORIGINAL: LoBaron
This statement suggests a classic misconception of what should be a "common result", because the expected
average is based on viewing a single historical event as representative, simply because it happened that way.

+1

History also has had very lucky events as well as unlucky.

[Sardaukar]

I'd say that historically Japanese were bit lucky to get result they did (number of sunk BBs), especially with USS Arizona's magazine explosion.

[mdiehl]

This statement suggests a classic misconception of what should be a "common result", because the expected
average is based on viewing a single historical event as representative, simply because it happened that way.

Your statement also suggests a classic misconception, that the presence of only one data point means that that data point isn't more or less a resonable expectable outcome.

Face it, alot of the things ConSims try to model are "one offs." But it makes more sense to make your "average simulated result" approximate the only data point that you actually have, than it does to make your average result something ELSE.

In game terms, "sunk" means "unrecoverable." Therefore, a historical result in game terms is 2BB sunk, 1 CM sunk, and 2 DDs sunk. Since a ship in PH had to be pretty well obliterated to be simply unrecoverable (like Arizona), I think based on what we know both of the Pearl Harbor ship facilities, the harbor depth, the Japanese equipment, and the general robustness of BBs, that leaning towards the "historical result" as a design goal for the game (or any consim covering the same subject) is a rather good way to do things.

[Nikademus]

ORIGINAL: GoTraurig
discussions.

Everyone: My main purposes for this thread were to achieve some combination of: 1. To verify that I was setting up the strike properly. 2. To ensure I wasn't experiencing a bug. 3. To seek advice on minimizing the potential for a really bad result. 4. To learn what results should be expected. 5. To determine why my strikes were so much weaker than those I usually see in AARs and other discussions.

So far I have at least partially achieved 1, 3, 4, and 5, with my thanks and appreciation to you all. Currently my only related problem is ensuring more than half my Kates fly - I understand it could be just bad die rolls, but they were awfully consistent for a while.

I've also pretty much accepted that the solution, and all the advice I'll probably receive about this, is to ensure that the Kates have the proper orders and just hope they decide to fly.

There's not alot you can do to fuddle the results as randoms will have their way. You could forgo attacking the airfields and set all your bombers to Port to maximize the ship attacks but even then your results might not be better than if you left them alone. By default the only thing i change for Turn 1 assuming no 1st historical turn setting is undo the Straffing orders for the fighters. I'd rather not waste them. However I believe destroying some of the airpower has as much use as the ships themselves even though the results tend to be on average only 1/2 of what was achieved in real life. In my last two results I scored Zero battleships sunk and in the latter i got two. Think the best result i ever got was four.

[LoBaron]

ORIGINAL: mdiehl

This statement suggests a classic misconception of what should be a "common result", because the expected
average is based on viewing a single historical event as representative, simply because it happened that way.

Your statement also suggests a classic misconception, that the presence of only one data point means that that data point isn't more or less a resonable expectable outcome.

Hm, well youre one of the "If someone says you have to be careful to identify something as 'A', he states it must be 'B'" guys, are you?

Face it, alot of the things ConSims try to model are "one offs." But it makes more sense to make your "average simulated result" approximate the only data point that you actually have, than it does to make your average result something ELSE.

What makes you think that on scenarios like PH theres only one data point available?

That would only be true if you see the attack as a holistic event generating only one data point, that it is not possible to extract data from a multitude events, down to a scale of penetration probability
of a 800kg bomb hit on a specific armour type, and recombine the data to work with that information.
Which is obviousely wrong.

In game terms, "sunk" means "unrecoverable." Therefore, a historical result in game terms is 2BB sunk, 1 CM sunk, and 2 DDs sunk. Since a ship in PH had to be pretty well obliterated to be simply unrecoverable (like Arizona), I think based on what we know both of the Pearl Harbor ship facilities, the harbor depth, the Japanese equipment, and the general robustness of BBs, that leaning towards the "historical result" as a design goal for the game (or any consim covering the same subject) is a rather good way to do things.

Thank you, I never knew... [8D]

[mdiehl]

Hm, well youre one of the "If someone says you have to be careful to identify something as 'A', he states it must be 'B'" guys, are you?

Not that I am aware of. But I'm not entirely sure what it is you just wrote.

What makes you think that on scenarios like PH theres only one data point available?

I think you missed something. I said there's one HISTORICAL data point. There was only one Pearl Harbor attack. Where "data" with respect to the outcome refers to "number of ships sunk vs damaged by class" there's one data point. That was not to say you can't analyze other things. Only that if one judges the overall result of the PH attack in a Consim, the one historical data point is better than operating in an information vacuum.

That would only be true if you see the attack as a holistic event generating only one data point,

I agree. But that was essentially the nature of this thread wasn't it? People posting the results of their PH attacks and claiming that they weren't sinking enough BBs. They weren't talking about not penetrating enough armor or not generating a big enough blast radius on average with their 500 Kg bombs, they were talking about the results overall.

that it is not possible to extract data from a multitude events, down to a scale of penetration probability
of a 800kg bomb hit on a specific armour type, and recombine the data to work with that information.

Might be doable if anyone had really good data on that and had modeled that extensively prior to designing the combat algorithm. To my knowledge there are very few people who CAN do such calculations, fewer who have, and most of said info is buried in a whole lot of (possibly outdated) Rock Island and US Naval Arsenal ballistic data. Were they used to design the combat routines in WitP.

Which is obviousely wrong.

Not necessarily. If you're just churning butter into cheese without really doing a whole lot of detailed research on penetration tests, angles of impact, and so forth, and using the penetration tests and the angles of impact as the basis for your Consim combat alogorithm, then "winging it" with such concerns has little chance of producing objectively reasonable results. You might as well go to a holistic assessment, because at least THAT would be indexed to something real.

[Shark7]

ORIGINAL: Nikademus

ORIGINAL: GoTraurig
discussions.

Everyone: My main purposes for this thread were to achieve some combination of: 1. To verify that I was setting up the strike properly. 2. To ensure I wasn't experiencing a bug. 3. To seek advice on minimizing the potential for a really bad result. 4. To learn what results should be expected. 5. To determine why my strikes were so much weaker than those I usually see in AARs and other discussions.

So far I have at least partially achieved 1, 3, 4, and 5, with my thanks and appreciation to you all. Currently my only related problem is ensuring more than half my Kates fly - I understand it could be just bad die rolls, but they were awfully consistent for a while.

I've also pretty much accepted that the solution, and all the advice I'll probably receive about this, is to ensure that the Kates have the proper orders and just hope they decide to fly.

There's not alot you can do to fuddle the results as randoms will have their way. You could forgo attacking the airfields and set all your bombers to Port to maximize the ship attacks but even then your results might not be better than if you left them alone. By default the only thing i change for Turn 1 assuming no 1st historical turn setting is undo the Straffing orders for the fighters. I'd rather not waste them. However I believe destroying some of the airpower has as much use as the ships themselves even though the results tend to be on average only 1/2 of what was achieved in real life. In my last two results I scored Zero battleships sunk and in the latter i got two. Think the best result i ever got was four.

And if you don't make the airfield attack on turn 1, a successful turn 2 attack becomes a far less achievable goal. I generally leave the fighters alone for turn 1, but change all Kates to port and put the Vals on airfield. It does make it easier to stick around 2 more turns when I do that. I'm sure every one does things just a bit differently though.

[Keldun]

ORIGINAL: LoBaron

GoTraurig, I did not see your post as a complaint.

My motivation to post the 'The art of "averaging out"' thread is much
older already, because misinterpreting historical results is a more or less common
occurance on the forum.

If there was any additional reason at all, it was Kelduns post suggesting to "boost a bit more damage"
to "so that we can get a result a bit closer to what it was at the start of the war" concerning PH.

This statement suggests a classic misconception of what should be a "common result", because the expected
average is based on viewing a single historical event as representative, simply because it happened that way.

What's wrong with reproducing something close to what happened at ph if we don't assign any commands?
Why should this game make it as hard as possible to reproduce a result close to what happened in a historical battle ? ( I m talking about the planes too, never seen anything close to the 180+ destroyed during that battle.)

[Alfred]

For a start, the algorithms for 7th December 1946 have to be the exactly the same for each of the potentially 1300+ turns.

You want the same result, as it has been said ad nauseum, play the 8th December 1941 start. Die rolls will always introduce randomness.

Over the years there have been plenty of posted results which demonstrate that given an appropriately sized sample, there is nothing wrong the Pearl Harbor strike.

Alfred

[Keldun]

The first turn is different from the other 1k+ turns with the 7th dec option.
I' m at about 40 tries now and haven't once gotten anything close. How many more sample do you want? I can bet you without much doubt that if i tried 20 more times i wouldn't get a result close either.

And about die rolls, of course it is random but at the moment a lower result than what happened historically is almost guaranteed, not much randomness in the chance of getting a lower result then.

But if you add a bigger constant in damage to the 7th december option, the number of destroyed machines would be higher.
Those following numbers are completely fictional but lets say that a bb ship needs 90 damage to sink if we add a constant of +10 damage to the current +5 damage only once, as soon as that ship got hit the chance that the ship will sink will still be higher on average and we may have a slightly higher chance to get a result close.
That would only affect the first turn and not the rest of the game, I' m not sure why people are so against the idea since its an option that can be disabled.

And about the 8th december scen, I'd rather like to see the war in action in game from the start rather than know that it happened.
Or I could just make a scenario starting near the end of the war just before japan surrendered and imagine all the historical battles if I only cared about the result.