Type 93 Torpedo

Uncommon Valor: Campaign for the South Pacific covers the campaigns for New Guinea, New Britain, New Ireland and the Solomon chain.

Moderators: Joel Billings, Tankerace, siRkid

User avatar
Nikademus
Posts: 22517
Joined: Sat May 27, 2000 8:00 am
Location: Alien spacecraft

RE: Type 93 Torpedo

Post by Nikademus »

ORIGINAL: jwilkerson

One problem with the study BigB cites, is that it is still not exhaustive. We were even treated with an explaination about what was not included. But I guess fundamentally my question was different. My question was not about surface actions. My question was about damage caused by T93 torpedoes to vessels of all nations. (What damage was caused by T93 torpedoes in any and all circumstances to any and all vessels of any and all nations). That question overlaps the cited study, but it is not the primary question answered by that study.


The study cited was born out of multiple thread discussions about the potential impact of the Type93 back in the 90's on the warships1.com BBvsBB forum. The problem i had with it at the time it was being put together (by a fellow longtime poster/member of that forum) remains the same then as now.....that it was for all the sources cited a simple exercise in counting the hits and coming up with a % and then concluding that something theoretical could never or would probably never have happened....despite the fact that, as mentioned, the study was not exhaustive nor did it take into account the myriad difference in condtions for each battle.

Despite it's limitations...both admitted and postulated...the real irony in the days since has been how this study-post (preserved as an "article" for the warships1 board) has been taken out of context over the years and/or taken as gospel. It was always about postualating on whether or not Japan's pre-war decisive battle plan, specifically the part where the torpedo-men came into play, had any real chance of success.....but has since then been used to justify a wide variety of opinions mainly centering around smaller night operations....not the "Decisive Battle"
mdiehl
Posts: 3969
Joined: Sat Oct 21, 2000 8:00 am

RE: Type 93 Torpedo

Post by mdiehl »

The irony to me is how any quantitative information regarding the success rate of Japanese equipment is disregarded as "taken out of context" or otherwise accompanied by non-sequitur disparaging remarks (to wit "regarded as gospel") when it contradicts the Gospel of the Invincible Type 93a.

The plain fact is that Czarnecki's analysis was QUITE comprehensive, included all of the major engagements in which the torp was used, and did NOT suggest (as was imlpied before) that therefore " then concluding that something theoretical could never or would probably never have happened...." That was, in fact, not the conclusion. What it said was that the Type 93a does not seem to have been a winning component of the decisive battle doctrine given the examples of its actual success in combat.

To disagree with Czarnecki's conclusion, one would either have to reject the use of historical data for testing the quality of the decisive battle doctrine, or else one would have to conclude that the only data that should be included should be data that reflect well on said doctrine. Sure, if every battle were always just like the Japanese wished they could be, or every outcome predictably always just like Tassafaronga or Savo, then one can make the Type 93a into the uberweapon that mythology characterizes it to have been.

But the cold hard reality of facts and circumstances of use show that it wasn't. Once you become adequately informed about the stats (the mean, mode, and median hit rates at day and at night) you start to think about why in some circumstances torpedoes fired by surface ships worked very well, and why in most engagements they were completely ineffective.
Show me a fellow who rejects statistical analysis a priori and I'll show you a fellow who has no knowledge of statistics.

Didn't we have this conversation already?
User avatar
jwilkerson
Posts: 8251
Joined: Sun Sep 15, 2002 4:02 am
Location: Kansas
Contact:

RE: Type 93 Torpedo

Post by jwilkerson »

"Comprehensive" needs a context to be relevant. My context, as thread starter in this thread was "How much damage was caused by T93 to warships of any nation, by any means?" The context of the cited study is damage caused to enemy ships in surface battles. The study might be comprehensive within the question it was trying to answer, but not in terms of the question that formed the theme of the thread!

WITP Admiral's Edition - Project Lead
War In Spain - Project Lead
User avatar
Wirraway_Ace
Posts: 1509
Joined: Mon Oct 08, 2007 2:28 pm
Location: Austin / Brisbane

RE: Type 93 Torpedo

Post by Wirraway_Ace »

ORIGINAL: mdiehl

The plain fact is that Czarnecki's analysis was QUITE comprehensive,...

But does not attempt to include the damage/risk that the Type 93a represented to the ship carrying it.
ORIGINAL: mdiehl

Once you become adequately informed about the stats (the mean, mode, and median hit rates at day and at night) you start to think about why in some circumstances torpedoes fired by surface ships worked very well, and why in most engagements they were completely ineffective.

I am surprised that one might find the mode and median as valuable tools for describing the success of Type 93a attacks. A posteriori, I find mean based measures more effective for data distributed in this way....[;)]
User avatar
Wirraway_Ace
Posts: 1509
Joined: Mon Oct 08, 2007 2:28 pm
Location: Austin / Brisbane

RE: Type 93 Torpedo

Post by Wirraway_Ace »

ORIGINAL: jwilkerson

"Comprehensive" needs a context to be relevant. My context, as thread starter in this thread was "How much damage was caused by T93 to warships of any nation, by any means?" The context of the cited study is damage caused to enemy ships in surface battles. The study might be comprehensive within the question it was trying to answer, but not in terms of the question that formed the theme of the thread!

So we can be helpful by citing examples of when a Type 93a caused damage, other than by striking enemy warships, to help you derive differential rates for damage caused to a ship struck while carrying Type 93a torps vs those with other torps?
User avatar
Nikademus
Posts: 22517
Joined: Sat May 27, 2000 8:00 am
Location: Alien spacecraft

RE: Type 93 Torpedo

Post by Nikademus »


The irony to me is how any quantitative information regarding the success rate of Japanese equipment is disregarded as "taken out of context" or otherwise accompanied by non-sequitur disparaging remarks (to wit "regarded as gospel") when it contradicts the Gospel of the Invincible Type 93a.

The irony is that the article was neither disregarded in it's entirty, nor was it claimed that the Type93A was invincble.
The plain fact is that Czarnecki's analysis was QUITE comprehensive."

Incorrect. As mentioned, "comprehensive" requires a clear context to be relevent. The author admitted in the article, and during forum conversations during it's construction that it was not an exhaustive, complete study.
To disagree with Czarnecki's conclusion, one would either have to reject the use of historical data for testing the quality of the decisive battle doctrine, or else one would have to conclude that the only data that should be included should be data that reflect well on said doctrine.

Straw Man. The disagreeing with a conclusion, in part or in whole does not mean rejection of historical data since historical data requires interpretation and to be taken in proper context. That is why historians can (and often do) argue over the same set of data points. In this case it doesn't address at all JoeW's question as he has patiently pointed out. [:)]
But the cold hard reality of facts and circumstances of use show that it wasn't.

I don't recall these selective facts to include an actual playing out of the Japanese plan in wartime. That would be hard to do since it never happened as postulated. From a study of "Kaigun" it most likely would not have occured as planned and not by a long shot. The Japanese were entirely too optimistic in their appraisals and Joe's analysis lends support the conclusion already fielded by Evans and Peattie. To say it could not happen at all however under the circumstances postulated cannot be stated with authority.
mdiehl
Posts: 3969
Joined: Sat Oct 21, 2000 8:00 am

RE: Type 93 Torpedo

Post by mdiehl »

@jwilkerson - OK. Yes, vis the question that you asked, Czarnecki's analysis does not address exactly the question. I'm not sure how or whether one could come up with some reliable estimate of the likelihood of a Type 93a to cause damage to the ship carrying it or to a friendly.

@Wirraway Ace
am surprised that one might find the mode and median as valuable tools for describing the success of Type 93a attacks. A posteriori, I find mean based measures more effective for data distributed in this way....


Mean is a pretty good measure when the distribution is unimodal and otherwise reasonably well-behaved (not too skewed, not too much kurtosis). The IJN Type 93a % hit rate is neither. Thus, the mean hit rate overall is 6.7% regardless of context, but the standard deviation is quite high.

WTF why not here you go. Number of cases 21 (I've been adding to Czarnecki's tables), mean hit rate 6.7%, 95% confidence interval 3.07-10.35%, minimum 0%, maximum 25%, lower quartile 0% upper quartile 12.5%, std deviation 8.00.

Context matters. In daylight the mean hit rate is 0.6%. At night it's 7.7%. About 15% of Japanese torps that hit ships struck ships that were already dead in the water or substantially slowed because of prior battle damage, but this does not include for ex the scuttling by torpedo of abandoned enemy vessels.

The mean number of Allied ships hit per engagement is 1.3 ships. Minimum 0, maximum 5 but these are based on my own incomplete data.

&c &c
Show me a fellow who rejects statistical analysis a priori and I'll show you a fellow who has no knowledge of statistics.

Didn't we have this conversation already?
mdiehl
Posts: 3969
Joined: Sat Oct 21, 2000 8:00 am

RE: Type 93 Torpedo

Post by mdiehl »

Incorrect. As mentioned, "comprehensive" requires a clear context to be relevent. The author admitted in the article, and during forum conversations during it's construction that it was not an exhaustive, complete study.


Nevertheless, the contexts are well known to both that author and to me making use of his tallies. His study actually IS quite comprehensive when you look at the actual number of surface-surface engagements fought. Indeed, his data are somewhat skewed in favor of the IJN, because they exclude some battles in which many Type 93as were launched but did not hit, on account of the set-ups not being setups imagined in the IJN DBD.
Straw Man. The disagreeing with a conclusion, in part or in whole does not mean rejection of historical data since historical data requires interpretation and to be taken in proper context.

It's not a straw man argument given that you have stated flat out that you reject the analysis because it does not envision a set up that presumes that the Torps work into the DBD in the way that the IJN wanted them to. In essence you have claimed nothing more than that 'Czarnecki's analysis is irrelevent because under the right set of circumstances the Type 93a could have worked out the way the IJN wanted in the DBD.' Well, no, actually, the point is that the data don't support that claim. The IJN DBD envisioned a series of engagements preceding to and leading up to a big battle. If we treat these data as indicative of the kinds of successes the IJN would have in a series of engagements (them being actual data from a series of engagements and all that), then the Type 93a isn't the war winner the Japanese hoped it would be. It's not even as good as UV and WitP make it out to be.
I don't recall these selective facts to include an actual playing out of the Japanese plan in wartime.

Yes I understand. The absence of data conclusively favorable to the Japanese implies the inapplicability of actual data from the actual war. Your rebuttal is bravo sierra.
To say it could not happen at all however under the circumstances postulated cannot be stated with authority.

I don't recall Czarnecki or anyone else saying that. As I and Czarnecki noted context matters. The only way you can plausibly cook up a scenario in which the Type 93a accomplishes its role in the DBD is to assume that most instances of real world Japanese torp use would resemble those instances in which the IJN realized their BEST performance, such as at Savo Island or at Tassafaronga. The data clearly indicate that those actions were statistically atypical. Therefore, it logically follows that the Tuype 93a living up to its role in the DBD is at best, statistically speaking, highly improbable. When you look at the contexts that would be required to override the stats (which involve commodore situational awareness, radar, orders of battle, and similar factors), the probability of the Type 93a fulfilling its desired role in the DBD approaches nil.
Show me a fellow who rejects statistical analysis a priori and I'll show you a fellow who has no knowledge of statistics.

Didn't we have this conversation already?
User avatar
Nikademus
Posts: 22517
Joined: Sat May 27, 2000 8:00 am
Location: Alien spacecraft

RE: Type 93 Torpedo

Post by Nikademus »

Nevertheless, the contexts are well known to both that author and to me making use of his tallies. His study actually IS quite comprehensive when you look at the actual number of surface-surface engagements fought. Indeed, his data are somewhat skewed in favor of the IJN, because they exclude some battles in which many Type 93as were launched but did not hit, on account of the set-ups not being setups imagined in the IJN DBD.

The contexts are not well known. He admitted that in his own study. It and your comments also have no context to what JoeW is asking.
It's not a straw man argument given that you have stated flat out that you reject the analysis because it does not envision a set up that presumes that the Torps work into the DBD in the way that the IJN wanted them to.

Show me where I stated that I flat out rejected the analaysis.
In essence you have claimed nothing more than that 'Czarnecki's analysis is irrelevent because under the right set of circumstances the Type 93a could have worked out the way the IJN wanted in the DBD.'

Incorrect. I pointed out that I had some issues with how he ultimately interpreted his data. We had a number of good discussions about it actually. You are also deliberately igorning that I have stated.....twice that ultimately Joe's article can be seen as supporting the conclusion of Evans/Peattie who stated in their book that it was highly improbable that the Japanese plan would work. I have no disagreement with that.
It's not even as good as UV and WitP make it out to be.

A point of context Joe did not attempt to make. Nor can you since you've never played the games in question.

quote:

Yes I understand. The absence of data conclusively favorable to the Japanese implies the inapplicability of actual data from the actual war. Your rebuttal is bravo sierra.

No, the absence of complete data and analysis accompanied by the absence of an actual battle under which the conditions the DB were to be fought means that one can conclude, based on Evans/Peattie that such a result is improbable, but not impossible.

I don't recall Czarnecki or anyone else saying that. As I and Czarnecki noted context matters. The only way you can plausibly cook up a scenario in which the Type 93a accomplishes its role in the DBD is to assume that most instances of real world Japanese torp use would resemble those instances in which the IJN realized their BEST performance, such as at Savo Island or at Tassafaronga.

Except that those battles do not resemble the DB and as Joe admitted, he was unable to completely evaluate the manner in which each battle was fought, including the the number of weapons fired by which ship, range and speed settings, spread type and target. Ultimately he counted launches and hits. Ultimately, as mentioned it can be seen as a supportive argument to Evans/Peattie's comments regarding the DB.
User avatar
Miller
Posts: 2227
Joined: Tue Sep 14, 2004 10:14 am
Location: Ashington, England.

RE: Type 93 Torpedo

Post by Miller »

I have come late to this one, but I can say as a Jap player with many games under my belt that the LL is way to accurate in UV. It is less so in WITP, but still hits more often than in real life......
User avatar
Nikademus
Posts: 22517
Joined: Sat May 27, 2000 8:00 am
Location: Alien spacecraft

RE: Type 93 Torpedo

Post by Nikademus »

I'm admitedly a long way out from my UV days but I don't recall seeing an overflux of torpedo hits by LL's as either player. In WitP, i've yet to ever duplicate a Tassafaronga, either in static tests or AI/PBEM. Came close a couple times. And no....this doesn't put me in Ike's camp over on the "Needs a patch" thread. [;)]
User avatar
decaro
Posts: 4004
Joined: Wed Aug 31, 2005 12:05 pm
Location: Stratford, Connecticut
Contact:

RE: Type 93 Torpedo

Post by decaro »

In the "Needs a Patch" thread, Tocaff and I ran several surface battles from the Eastern Solomons scenario by PBEM to prove the that LL was very effective, though not everyone on the thread was convinced.

We concluded that the torpedo wasn't weak, but Allied surface/gunnery radar was too strong in shutting down IJN LL attacks.

We made a suggestion in the CF thread to increase the effectiveness of radar over time, modified by the experience of the ship's crew; that way LL attacks would be more frequent early on in UV w/o having to tinker w/the torpedo.

Like I told Ike, CF is UV's next (only?) patch.
Stratford, Connecticut, U.S.A.[center]Image[/center]
[center]"The Angel of Okinawa"[/center]
Home of the Chance-Vought Corsair, F4U
The best fighter-bomber of World War II
User avatar
Nomad
Posts: 7273
Joined: Wed Sep 05, 2001 8:00 am
Location: West Yellowstone, Montana

RE: Type 93 Torpedo

Post by Nomad »

I would also like to see USN Radar not start showing up until October, 1942. As things are in UV, ships can get radar in June, 1942.
User avatar
tocaff
Posts: 4765
Joined: Wed Oct 11, 2006 9:30 pm
Location: USA now in Brasil

RE: Type 93 Torpedo

Post by tocaff »

Nomad
You should put the radar dates into the CF wish list thread.
Todd

I never thought that doing an AAR would be so time consuming and difficult.
www.matrixgames.com/forums/tm.asp?m=2080768
mdiehl
Posts: 3969
Joined: Sat Oct 21, 2000 8:00 am

RE: Type 93 Torpedo

Post by mdiehl »

The contexts are not well known. He admitted that in his own study. It and your comments also have no context to what JoeW is asking.

The contexts are well known to anyone who'se read about the battles. What more information do you think you must have that you currently don't have?
Show me where I stated that I flat out rejected the analaysis.


---->
the real irony in the days since has been how this study-post (preserved as an "article" for the warships1 board) has been taken out of context over the years and/or taken as gospel.

Which of your gospels did Big B offend?

---->
No, the absence of complete data and analysis accompanied by the absence of an actual battle under which the conditions the DB were to be fought means that one can conclude, based on Evans/Peattie that such a result is improbable, but not impossible.

Dwelling on a straw man argument about the absence of complete data and the other straw man argument about the absence of an "actual battle" strikes me as a dismissal of the analysis. You've clearly stated that you don't think the real world battles represent operational realities as they might be encountered in, well, the real world. QED.

Since you keep saying that, however:
No, the absence of complete data

1. The data are thorough and very representative of typical circumstances in which battles of the kind in the Decisive Battle doctrine were fought. One can add more data but they do not seem to improve Japanese successes. I've certainly added data to the results cataloged before but the outcome hasn't gotten better. If you can find a battle or a sequence of battles that substantially falsify the results of Joe's analysis that he hasn't included, pipe up.
the absence of an actual battle under which the conditions the DB were to be fought


2. All of the battles included by Czarnecki were battles in which the conditions represented conditions that one could expect under the decisive battle doctrine. The suggestion to the contrary makes no sense at all, since the DB doctrine envisioned a campaign consisting of a serious of attritioning battles in which presumed USN numerical superioirty would be whittled down to parity or inferiority prior to one last great decisive battle. The battles included by Czarnecki do a very good job representing what that serious of attritioning battles might look like, and they conclusively demonstrate that Japan was not going to get the desired attrition rate prior to that last great battle.

You *seem to* be relying on a tautology here -- dismissing the relevance of these historical engagements because they're not outcomes favorable to the Japanese as required in the decisive battle doctrine. In effect, the theory is sound because the data are rejected. That is logically an unsound practice. Given the extant data from real world combat, the preponderance of the evidence indicate that the DB doctrine wasn't going to work.

Finally, if you did not mean to reject the analysis as germane to proof testing WitP or UV, why would you accuse Big B of taking it "out of context" or as "gospel?" The data are germane to a discussion of Type 93a torp effectiveness unless you imagine the contexts in which the Type 93a was used were "taken out of context" by the IJN. (Which gets us back to that tautology you keep deploying and then tripping over). Big B posted the Czarnecki analysis without comment. Which gospel did he violate? Clearly, the existence of Czarnecki's analysis violates some sort of gospel to which you are clinging.
Incorrect. I pointed out that I had some issues with how he ultimately interpreted his data.

Same difference from where I sit.
You are also deliberately igorning that I have stated.....twice that ultimately Joe's article can be seen as supporting the conclusion of Evans/Peattie who stated in their book that it was highly improbable that the Japanese plan would work. I have no disagreement with that.


Then why are you here writing off the relevance of the analysis? If you think the basic conclusion is correct what after all is the point of suggesting that the use of the data compiled by Czarnecki seems inappropriate for groundtruthing WitP or UV? Or were you just impressing us all with your lofty dismissal of the analysis (while agreeing in principal with the end result) just as a way of objecting to Big B mentioning them here at all?
A point of context Joe did not attempt to make.

No one other than yourself imagined that he attempted to apply those data to WitP or UV.
Nor can you since you've never played the games in question.

Assumes facts that are in fact not correct. I have played UV and rather extensively. Moreover, there are more than enough AARs available from WitP to make a conclusion. But you already knew both, and that makes you at best a kind of cartoon distortionist playing at discourse.
Except that those battles do not resemble the DB and as Joe admitted, he was unable to completely evaluate the manner in which each battle was fought, including the the number of weapons fired by which ship, range and speed settings, spread type and target. Ultimately he counted launches and hits. Ultimately, as mentioned it can be seen as a supportive argument to Evans/Peattie's comments regarding the DB.

1. The DB doctrine did not require specific statement of torpedo settings vis speed, depth, torpedo spread, or any of these things. Therefore, their exclusion from Joe's analysis does not obviate the applicability of his analysis to an examination of the likely outcome of the DB.

2. For it to obviate applicability to the evaluation of the DB doctrine on account of some imagined scenario in which substantially different settings would be use, one would have to presume that the torpedo speed, depth and spread settings in all of these battles consistently deviated from Japanese doctrine and training. There is no evidence to warrant such presumption.
Show me a fellow who rejects statistical analysis a priori and I'll show you a fellow who has no knowledge of statistics.

Didn't we have this conversation already?
User avatar
jwilkerson
Posts: 8251
Joined: Sun Sep 15, 2002 4:02 am
Location: Kansas
Contact:

RE: Type 93 Torpedo

Post by jwilkerson »

ORIGINAL: Wirraway_Ace

ORIGINAL: jwilkerson

"Comprehensive" needs a context to be relevant. My context, as thread starter in this thread was "How much damage was caused by T93 to warships of any nation, by any means?" The context of the cited study is damage caused to enemy ships in surface battles. The study might be comprehensive within the question it was trying to answer, but not in terms of the question that formed the theme of the thread!

So we can be helpful by citing examples of when a Type 93a caused damage, other than by striking enemy warships, to help you derive differential rates for damage caused to a ship struck while carrying Type 93a torps vs those with other torps?

Exactly, for example in the Suzuya example in the first post in this thread. The Suzuya is not actually "hit" by enemy ordnance. A "near miss" causes the loaded T93 in one bank to explode. The fires from these explosions eventually cause additional T93 to explode and the ship ships. This ship could easily qualify has having been sunk by her own torpedoes.

Any Japanese ship hit by Japanese T93 would qualify as well as cases where T93 aboard Japanese ships exploded and caused crippling damage.
WITP Admiral's Edition - Project Lead
War In Spain - Project Lead
mdiehl
Posts: 3969
Joined: Sat Oct 21, 2000 8:00 am

RE: Type 93 Torpedo

Post by mdiehl »

This ship could easily qualify has having been sunk by her own torpedoes.

No it couldn't, at least no more more than Lexington could easily qualify as having been sunk by her own fuel tanks, Bismarck sunk by her own rudder, or Yorktown sunk by her own malfunctioning boilers.
Show me a fellow who rejects statistical analysis a priori and I'll show you a fellow who has no knowledge of statistics.

Didn't we have this conversation already?
User avatar
jwilkerson
Posts: 8251
Joined: Sun Sep 15, 2002 4:02 am
Location: Kansas
Contact:

RE: Type 93 Torpedo

Post by jwilkerson »

ORIGINAL: mdiehl
This ship could easily qualify has having been sunk by her own torpedoes.

No it couldn't, at least no more more than Lexington could easily qualify as having been sunk by her own fuel tanks, Bismarck sunk by her own rudder, or Yorktown sunk by her own malfunctioning boilers.

I don't understand the relationship you propose?
WITP Admiral's Edition - Project Lead
War In Spain - Project Lead
mdiehl
Posts: 3969
Joined: Sat Oct 21, 2000 8:00 am

RE: Type 93 Torpedo

Post by mdiehl »

Ships were full of things that could be damaged to the point of inoperability, that might lead to running aground or capsizing, such as the loss of a steering mechanism. Likewise they were full of things that could explode, like fuel in a.c. carriers, munitions in AKs, etc. Ammunition in a burning magazine could cook off (and in too many instances to name) did so, sometimes with catastrophic results.

Nevertheless, it would never have occurred to me to say that "The USS Arizona was sunk by her own shells" or "The USS Lexington was sunk by her own gas tank" or "The HMS Hood was sunk by her own gunpowder" or "The USS Yorktown was sunk by her own inoperable engines." All of these were contributing factors to the demise of the aforementioned ships, but the plain fact is that NONE of them would have been sunk without substantial damage inflicted by the enemy.

"Sunk by her own torpedoes is a phrase" whose only application has value in a discussion of circular torpedo runs.
Show me a fellow who rejects statistical analysis a priori and I'll show you a fellow who has no knowledge of statistics.

Didn't we have this conversation already?
User avatar
jwilkerson
Posts: 8251
Joined: Sun Sep 15, 2002 4:02 am
Location: Kansas
Contact:

RE: Type 93 Torpedo

Post by jwilkerson »

Hum,

I guess I have a world view that allows a more "multi-determined" view of the causes of some events. Ship sinkings is one of them.

Another example from the initial post in this thread is Mikuma and Mogami and Midway. Mogami was more heavily damaged by the collision and hence had already discharged her torpedos. This fact probably contributed to her survival. Mikuma, less heavily damaged had not discharged her torpedos this fact probably contributed to her loss.

WITP Admiral's Edition - Project Lead
War In Spain - Project Lead
Post Reply

Return to “Uncommon Valor - Campaign for the South Pacific”