Incoming!

[Nemo121]

Couple of points in response to the actual topic:
 
1. In WW2 naval combat with torpedoes I think 10 kilometres is pretty long range.
 
2. In statistical analysis I think you get onto very shaky ground when you say ( effectively ) "if we ignore instances where the Allies were surprised or where there were startlingly high hit percentages". Sorry but an analysis should include as many data points as possible and not just those ones which support your view.
 
3. You might argue that "being surprised" was rare and so there's rationale for removing torpedo attack results while under such conditions. I would respond that since it was quite possible that the attack on the night preceding the "decisive battle" would have elements of surprise I could use that same argument to say "We MUST assume there would be surprise so I'm only going to accept the 20 to 40% hit rates. At that rate each IJN DD would put 16 torpedoes into the water, get about 5 hits and account for 2 USN BBs per DD committed". This is, obviously, indefensible statistically as you are selecting your sample in order to create the results you want.
 
 
IJN torpedoes weren't war-winners by any stretch of the imagination and they were neither as effective or as terrible as the two extremes in this debate would seek to misrepresent them to have been BUT I think the MEAN performance of any weapons system across ALL of its combat outings is a pretty reasonable guide as to its performance in a hypothetical battle which may have featured surprise or radar detection or any other myriad confounders such that we can't just say "to model the hypothetical battle I'll assume there was no surprise".
 
 
Outruling instances from your sample because they featured high percentage hits and then turning around to say that the hit rate was pathetically low is really not defensible IMO in this instance.
 
 
 
herwin,
Aye, if I understand you correctly under surprise conditions one could expect a three to four-fold increase in hits above the 6% figure I used. As such would you not agree that the 6% figure ( the mean of all conditions ) is a reasonable figure to use to figure out what one could expect from a series of concentrated DD night attacks against a battleline ( with several hundred torpedoes being fired at the battleline over the course of these attacks) ?
 
 
 
mdiehl,
The whole point behind the IJN doctrine was that the decisive battle would not be occuring at the end of a long attritional war in which the USN got to "catch on". Their idea was to ambush the USN fleet seeking to relieve the Phillipines etc when it sailed 6 months to a year after the war started.
 
Personally I think their doctrine relied a bit too much on the enemy being dumb BUT if the USN had done what the Japanese expected I really don't have any problem imagining that 30 IJN DD ( 16 torpedoes each ) + 6 or so CLs ( another 16 torps each ) + the 2 Kitakamis ( about 80 torps each ) could have fired the guts of 750 torpedoes at ranges of 10km or so from the US battleline with a reasonable prospect of surprise on at least some of those occasions such that the mean hit % would have been about 6% - resulting in about 45 Long Lance hits on BBs and CAs. That's a lot of the battleline sunk there.
 
I think the real issue is that their concept of the strategic imperatives and strength of airpower was so off as to render the whole decisive battle concept erroneous. BUT if it had happened as they had thought it would I just don't see how you can argue with a mean hit % of 6% taking into account both battles in which they had the advantage of surprise and those in which they didn't.

[Charles2222]

ORIGINAL: CV Zuikaku

ORIGINAL: castor troy

ORIGINAL: AW1Steve






Oh thank you! Not only do you throw fuel on the fire , you take sides! Why don't you help try and cool this debate off? That would be the RESPONSIBLE thing to do![:(]

Steve, please tell me that you don´t find Zuikaku´s post insulting or a try to piss Mr. M off. Either I don´t know English at all (perhaps I should start discussing in German but I guess Mr. M would find speaking German crappy - sorry, can´t resist [;)]) or why isn´t this a permitted question as it definetely begs some truth in it? [&:]

Well, I'm not a native english speaker, and my english is very crappy, but I hope it was not insulting... sorry if I was...

You weren't.

[Nemo121]

FWIW I thought you (Zuikaku ) were asking a question and not being insulting.

[herwin]

ORIGINAL: Dili

If you fire your torpedos at 15km obviously you will get less hits. If the doutrine says to fire at 2km there will be more, all other things equal.

Torpedoes were an area fire weapon: "It's not the bomb with my name on it that I'm afraid of; it's the one addressed 'To Whom It May Concern'". As long as the target battle line extended outside the spread of torpedoes being launched, range was only important in giving the targets more time to react. If they knew they were going to be blessed with a visitation, and they turned in time, there wasn't much more they could do.

[Mike Scholl]

ORIGINAL: Nemo121
Personally I think their doctrine relied a bit too much on the enemy being dumb BUT if the USN had done what the Japanese expected I really don't have any problem imagining that 30 IJN DD ( 16 torpedoes each ) + 6 or so CLs ( another 16 torps each ) + the 2 Kitakamis ( about 80 torps each ) could have fired the guts of 750 torpedoes at ranges of 10km or so from the US battleline with a reasonable prospect of surprise on at least some of those occasions such that the mean hit % would have been about 6% - resulting in about 45 Long Lance hits on BBs and CAs. That's a lot of the battleline sunk there.

Their whole concept of the "attrition battle" to proceed the "Decisive Battle" was pretty rigid, and depended a lot on the US Navy advancing in much the same manner as the British Grand Fleet at Jutland. Had they been presented with such a target, then it's highly likely that their light forces, operating at night with massed torpedo batteries of unprescedented (and unimagined) range, would have had a serious chance of inflicting a first class disaster.

The real failings of this doctrine were that the Japanese became so "invested" in it that they refused to allow for the enemy deviating from his part in "the plan". IRL, the big strengths of the "Long Lance" proved to be it's high speed, large punch, and lack of a visable wake. So many of the "design specs" for the weapon proved quite usefull, even if the monsterous range wasn't.

[herwin]

ORIGINAL: Nemo121

herwin,
Aye, if I understand you correctly under surprise conditions one could expect a three to four-fold increase in hits above the 6% figure I used. As such would you not agree that the 6% figure ( the mean of all conditions ) is a reasonable figure to use to figure out what one could expect from a series of concentrated DD night attacks against a battleline ( with several hundred torpedoes being fired at the battleline over the course of these attacks) ?

OPERATIONS RESEARCH

Destroyers were about half as vulnerable as major surface units, and it was also a question of what the ship spacing was in the line of battle and how many lines were present.

To replicate the Solomons actions and cover other possible night actions, I'd define three lines: screen, major warships, and non-combatants. The screen would consist of destroyer-class gunships and smaller cruisers (up to about 6500 tons or so). The major warship line would consist of larger cruisers, battlecruisers, and battleships. Non-combatants would consist of carriers and auxiliaries. Major warships would be spaced at 1000 meters. Non-combatants and carriers would be spaced at 1000 meters. Destroyers would be spaced to cover the other two lines--call that the target array length--with a minimum spacing of 500 meters, so that excess destroyers miss the party. Torpedoes would be distributed over the target array. Allied torpedoes would attack one line--the first one with significant numbers of ships. Japanese torpedoes would attack two lines, similarly beginning with the first one with significant numbers of ships. Determine the target array length, and the percentage of the target array length (defined above) for a given line that consists of ships--if surprised, use the ship lengths; if not, use the ship beams--that's the probability that a torpedo will hit a ship. Start with the torpedoes, and allot them randomly to ships in the first line. See which actually hit, and save the remainder for the next line--if there is another line. Repeat for the second line. Report.

Ignore the Mogami performance at Balipapan, where she sank five Japanese transports.

[mdiehl]

In statistical analysis I think you get onto very shaky ground when you say ( effectively ) "if we ignore instances where the Allies were surprised or where there were startlingly high hit percentages". Sorry but an analysis should include as many data points as possible and not just those ones which support your view.

Statistical analysis has to account for central tendencies and outliers. That is why any statistician worth two bits doesn't only look at the mean, but also the median and the mode. Sophisticated fellers will even break it down into percentiles. In any small sample, one expects statistical outliers, and these have to be on the one hand accounted for and on the other closely examined, because outliers often indicate the influence of some unaccounted for factor that is absent from the more normative results.

You might argue that "being surprised" was rare and so there's rationale for removing torpedo attack results while under such conditions.

Why would one argue that? A better approach is to ask oneself which conditions led to the outlying results and to assume that outliers would be more reasonable when atypical conditions occur.

IJN torpedoes weren't war-winners by any stretch of the imagination and they were neither as effective or as terrible as the two extremes in this debate would seek to misrepresent them to have been BUT I think the MEAN performance of any weapons system across ALL of its combat outings is a pretty reasonable guide as to its performance in a hypothetical battle which may have featured surprise or radar detection or any other myriad confounders such that we can't just say "to model the hypothetical battle I'll assume there was no surprise".

The mean of a distribution is only a good measure of central tendency when the distribution is normal. If the distribution is not normal, you can't use the mean and standard deviation to represent the distribution because you couldn't recreate the distribution using those measures alone. Every competent study of central tendency begins with a histogram; no competent study of central tendency presumes at the outset that the mean *must* be a good measure of central tendency.

The whole point behind the IJN doctrine was that the decisive battle would not be occuring at the end of a long attritional war in which the USN got to "catch on". Their idea was to ambush the USN fleet seeking to relieve the Phillipines etc when it sailed 6 months to a year after the war started.

The battle was expected to be attritional and involve a series of engagements beginning in the central pacific and ending close to Japan. One supposes that if Japan could have gotten ONE fight with the USN involving all the ships of both navies in March 1942, the plan might have born fruit. If they'd waited until March 1943, they'd lose, because by then US radar would have provided a tactically decisive advantage.

BUT if it had happened as they had thought it would I just don't see how you can argue with a mean hit % of 6% taking into account both battles in which they had the advantage of surprise and those in which they didn't.

The mean proportion of hits per battle is, with respect to WW2 IJN surface ship torpedo shots, a statistically poor measure of central tendency. It does not accurately describe the historical distribution. The reason why it does not do so is because torpedo success or failures were highly contingent on their use in favorable circumstances that, in the majority of IJN surface ship torpedo shots, were not extant at the time of the battle. That is why most of the time Japanese torpedos all missed.

[AirGriff]

Against all statistical odds, I have emptied the refridgerator of cold libations and just awoke from a daze. Having finished reading this thread from where I left off yesterday, I believe it is highly probable, let's say 3-1 odds, that I will go ask a neighber for something stronger to drink.

[JWE]

Statistics is a wonderful analysis tool so long as the input parameters are reasonably uniform. In warfare, each engagement has vastly different inputs, so the analytical result is not so much predictive as it is responsive (c.f. Deming). Mr Shewhart’s Bowl is a bit thin with only one ticket at the bottom.

Tools needs must be appropriate to the desired analysis. Statistics are not appropriate to prompt operational inspection, only longer term analysis.

For the philosophy, and math behind the statement, had a professional colloquy with (then Maj., now Maj. Gen.). Dan Bolger on this exact subject in the Nov? 1987 issue of The Artillery Journal (now Fires). Can discuss elsewhere if you are interested.

[Panther Bait]

I agree with JWE, the number of engagements combined with the complexity of a typical surface engagement, provides a pretty small sample set for analysis.  It makes the outliers hard to correlate with the rest of the data.  How extreme are they really?  One in a hundred?  One in a thousand?  Hard to say when you have 27 events to include in your analysis.
 
Mike
 
P.S. I will agree though with those that say radar makes torpedo attacks much less effective.  The change in course/aspect ratio typically necessary to unmask the torpedo batteries is fairly easy to recognize and react to, particularly at long range.  I seem to remember reading several times in Morrison when US forces detected likely torpedo launches on radar and just had to change course long enough to clear the torpedo water before returning to their original course.  It might give the firer time to break an engagement off, if desired, but it didn't generate many hits.

[Distiller]

ORIGINAL: Nemo121

Personally I think their doctrine relied a bit too much on the enemy being dumb BUT if the USN had done what the Japanese expected I really don't have any problem imagining that 30 IJN DD ( 16 torpedoes each ) + 6 or so CLs ( another 16 torps each ) + the 2 Kitakamis ( about 80 torps each ) could have fired the guts of 750 torpedoes at ranges of 10km or so from the US battleline with a reasonable prospect of surprise on at least some of those occasions such that the mean hit % would have been about 6% - resulting in about 45 Long Lance hits on BBs and CAs. That's a lot of the battleline sunk there.

i share the view, more or less, even though i would have gamed a lot in the IJN admirality's position.

the real and unimaginable blunder of IJ was to plan for literally decades for the deceisive battle only to trash all hopes to actually pull it of by a sneak attack (PH of course), which can only induce caution and a rethinking of doctrine among the USN while understanding said doctrine was the only real chance at suckering a superior force into the planned ambush. iow, they never believed in the deceisve battle and therefore, the range of the Type 93, the extra knots of speed for their BBs achieved through expensive refits, their sub doctrine, pretty much everything they did was a waste of time, to an extent.

what i would have tried:

If it's deceisive, commit all your available forces, everything, if you lose, you're dead anyway, so KB and all batdivs sortie, along with all DDs and CLs you can bring. subs will sortie or be vectored to the expected location of battle (or bases for refueling to gain options, especially high speed transit).

upon locating the enemy fleet, go to great lengths to make it a night engagement. sprinting is one option but it's treacherous as it reveals a desire for battle, better to make it look like a surprise skirmish against a scout force. you can sprint as long as you remain undetected, though. before the meeting, clandestinely detach all DDs with type 93 and all CLs you can spare, and of course bring Oi and Kitakami. approach the enemy line in two divisions, pretend not to notice their presence and let all ships with reloads launch their first salvo with max range setting when in range and once reasonable firing solution has been achieved. next, hope their hydrophones are drowned out by the fleet's own noise and the T93 isn't too loud at slower settings for max range. (iirc 35kts, 20nm)

continue to close until you're either too close for the TF leaders' comfort or you're engaged and launch all the rest, preferably at the high speed setting, if within range at that setting. (10nm at 50kts)

finally feign terror and scatter in seemingly disorganized fashion, a lot depends on conveying that image, even more on concealing your true numbers - it should be noted that close formation keeping and at least some understanding of radar might be required in order to do that. collisions would suck, but it's showtime, use it or lose it. if you score a lot of hits, a carrier strike at dawn will be just what the doctor prescribed, followed by a battle line engagement, maybe at afternoon. a lot would depend on the location of this carnage, if a decent airbase is near, hundred Betties and Nells could provide more torps and may even return and rearm once or twice if the carriers and therefore CAP had been neutralized before.

the most essential component (surprise) could only be achieved by restraint during the opening battles in the DEI, this means no Betties against Force Z and no LLs against ABDA under any circumstance - quite a downside. the worst objection would have to be that such a plan leaves initiative to the enemy. no sortie of the fleet, no deceisive battle and you're stuck with your self imposed rules of engagement.


the historic plan included sacrificing the Kongos to break through the screen - i wonder if that would have even been necessary given the LL's range, it wouldn't have been conducive to surprise, that much is certain. i wonder whether gambling entire nations is acceptable conduct, but suppose great risk is a cheap substitute for great power.

[Terminus]

And the IJN would have been able to utilize the full range of the Long Lance in a night battle, and the USN would have placidly steamed into all those fine torpedoes, suspecting nothing... Suuuuuure...[8|]

[mdiehl]

Tread carefully there Terminus, you're starting to sound like.... me. [;)]

[pasternakski]

ORIGINAL: Distiller
feign terror and scatter in seemingly disorganized fashion

This post is ironclad evidence you would have excelled at that.

[mdiehl]

BIG Liopleurodon.

[Distiller]

ORIGINAL: Terminus

And the IJN would have been able to utilize the full range of the Long Lance in a night battle, and the USN would have placidly steamed into all those fine torpedoes, suspecting nothing... Suuuuuure...[8|]

heck, i never said it was a good plan (just one which revolves around that particular torpedo) it just happens to be the kind of engagement it was apparently designed for. what is the long range for if not salvo launch at formations at absurd ranges? if you can't make it happen, you lose and realisitcally chances are slim. day launch is out of the question, everyone would know, if you can't avoid an ambush yourself you lose, if you miss, same, etc.

we already knew that.

[pasternakski]

ORIGINAL: Distiller
it just happens to be the kind of engagement it was apparently designed for. what is the long range for if not salvo launch at formations at absurd ranges?

This is exactly the conceptual trap that often made the Japanese slaves to their own technology. Just because the danged thing will zip along for miles and miles doesn't mean that you need to abandon study of tactical doctrine that will result in your wonderful weapon being applied to maximum effect.

What the Japanese failed to do was concentrate on command, control, communications and task force maneuver discipline. They often showed themselves to be just as disorganized and incompetent in night actions as their American adversaries. In the end, they were just as hide-bound to obsolete ideas as anyone else.

Thank he-who-has-no-name for the mistakes the Japanese made that cost them the war, or we might have wound up with control of aircraft engine production in WitP and midget submarines and hordes of sea trucks in AE...

[Terminus]

At 20 miles range/35 knots speed (your figures, not mine), the torpedoes would take around half an hour to reach their designated targets. Do you perhaps want to consider if said targets might be in an ENTIRELY DIFFERENT ZIP CODE by then?????

[Mynok]

Heck, that's enough time to invent the Enterprise and raise Scotty from a child so you can get beamed up.

[mikemike]

ORIGINAL: Terminus

At 20 miles range/35 knots speed (your figures, not mine), the torpedoes would take around half an hour to reach their designated targets. Do you perhaps want to consider if said targets might be in an ENTIRELY DIFFERENT ZIP CODE by then?????

If you go by the well-known quip that states that military organizations always plan, equip and organize to win the previous war, it is reasonable to assume that the IJN started off with a concept to win Jutland (although, after all the naval treaties, nobody would have expected to meet quite that target-rich an environment again). The point is that at Jutland, the battle squadrons held fairly close formations, the distance between ships of a squadron being 400 metres or less, and that course changes were avoided as much as possible, except when indispensable tactically, or if a torpedo attack was recognized. Particularly at night, the risk of collisions was too high to start zig-zagging without a very strong reason. During combat, the doctrine called for holding course and speed as constantly as possible to maximize rate and accuracy of gunfire. Given this expected enemy behaviour, it does not seem so fantastic that the IJN could be confident of predicting enemy position half an hour or more into the future with enough precision to make a mass torpedo attack using torpedoes with unprecedented range and trackless running (the enemy therefore not interpreting the behaviour of the Japanese light forces as a torpedo attack because they would have seemed to be wildly out of range) a viable proposition. At Jutland, shooting several hundred Type 93s at the Grand Fleet's battle line in such a manner would have resulted in a blood bath. Under the altered conditions prevailing in WWII, the IJN would have been better served by having torpedoes that ran at 60 knots over a range of 12k to 16k metres than by the Type 93s with their range of 49k metres at 35 knots.