by Attack on Pearl Harbor: Strategy, Combat, Myths, Deceptions
The flow chart provides an altogether different impression of the attack. Rather than a smooth, “textbook” operation, there were significant problems with visual identification of targets, command and control, and mutual interference.
The groups targeting the carrier anchorages demonstrated the faulty principles upon which the attack was based. The idea of a prioritization scheme that the aviators were to follow based on ship identifications made “on the fly,” plus the authority given to individual aircraft to break off targets at their discretion, meant that the Soryu and Hiryu torpedo attack aircraft scattered, wasted most of their torpedoes on poor targets, and disrupted the Kaga and Akagi groups’ attack on Battleship Row. Of the 16 bombers assigned to the carrier anchorage, only five attacked battleships. Utah and Helena were both attacked under the supposition that they were operational battleships.
The differences in the flow charts show the gap between planning and reality. What could have been a smooth attack was marred by considerable confusion. The second chart is based on the testimony of the Japanese aviators who survived the war. It is likely, given the timing, that there were many more cases of mutual interference and aborted attacks than are recorded.
All of the B5N Kate losses were from Kaga’s air group, which attacked last—five of the last seven torpedo bombers were shot down, four before they could deliver their weapon. But for the premature warning caused by Fuchida’s failure to communicate the correct attack plan, the attackers might have gotten in and out without losses.
The black boxes in the chart show the immediate post-battle Japanese assessment of the attack. Soryu and Hiryu together reported 11 hits distributed over three battleships, and three hits on a heavy cruiser (14 hits out of 16 aircraft). Akagi reported 11 hits distributed over three battleships (11 hits out of 12) and Kaga eight hits on two battleships (eight hits out of 12). Thirty-three hits were claimed. Four aircraft were shot down before they could launch their torpedoes, and one torpedo was jettisoned due to a near-collision with another plane, leaving 35 torpedoes launched for a claimed 94% hits.
Overall Assessment of the Torpedo Bombers
The effort of the torpedo bombers must be seen from two viewpoints.
First, they accomplished Yamamoto’s objective to sink four battleships. Mission accomplished, well done; for such a daring attack there is no higher accolade.
Second, their efforts could have accomplished so much more. Eleven torpedoes accomplished their objective out of 40 brought to the battle. The rest were wasted in overkill, against improperly identified targets, or simply missed hitting a ship. A large degree of the blame can be attributed to the planners’ ideas of how the attack should be conducted. There were shortfalls in target distribution instructions, designated attack profiles, and command and control. Blame must also go to the strike commander for fouling up the ordered attack plan, which resulted in premature warning for the defenders.
The Japanese did not appear to use radio communications for tactical control, which effectively meant their commanders had no control after battle was joined. Fuchida himself was in a level bomber attacking from 3,000 meters altitude while the heavy-hitters, the torpedo bombers, upon whose shoulders the fate of the attack rested, went in at 20 meters. Fuchida could exert no control over the torpedo attack while he was busy dropping his bomb. He was in no position to exert control for the entire duration of the first wave’s attack. He transmitted the order, “Charge!” and command and control went by the board.
The torpedo bomber leaders had limited ability to control their own subordinates. Their aircraft were strung out in lines miles long, out of reach of anything other than the most basic “follow me” leadership. There is no evidence that any torpedo leader attempted to re-direct any portion of the attack against more appropriate targets other than by leading their own wingmen, that is, in those cases where the wingmen decided to follow.
Subordinates were given the authority to pick their own targets, a necessary principle when applying a prioritization scheme. However, this also allowed very junior airmen to make mistakes without a way for the more experienced leaders to exercise effective supervision. Aircraft missed turns, lost sight of their leaders, and had near-collisions and aborted runs as aircraft milled about looking for targets while trying to get in and out as rapidly as possible. Many of the torpedo bombers effectively made their attack decisions independent of their leadership.
The inflexibility of the planning and training efforts must also be faulted. The large number of B5N Kates allocated to high-altitude bombing with AP bombs was excessive once it was known that there were no torpedo nets protecting the battleships in the harbor. Trading in 20 AP bombs for 20 torpedoes would have been like trading three AP bomb hits for 12 torpedo hits. Granted, the torpedo hits would likely have been directed against cruisers and destroyers, but most would argue that sinking three or four cruisers and some destroyers in exchange for a small chance of crippling an additional battleship was a worthwhile trade.
The prioritization scheme for the attack was useless, and mostly ignored. Not one torpedo was intentionally launched at a cruiser.
Looking at the torpedo attack as a whole, the planners’ efforts merit no praise. The torpedo bomber aircrews, hampered by a poor approach plan and an unworkable prioritization scheme, had a lot to overcome. Poor delivery techniques and target selection limited the effective torpedoes to less than one third, an unimpressive result.
Lack of SEAD in the “No Surprise” Scenario
The volume of AA fire undoubtedly contributed to the below-expectations performance of the torpedo attack. The total of 19 confirmed torpedo hits23 was well under the 82.5% hits recorded in the last practice session (a projected score of 33 hits out of 40 torpedoes), and also under the planners’ earlier expectation of 27 hits. While fewer hits ought to be expected under combat conditions, the hit percentage compares unfavorably with the performance of the British at Taranto, where 9 of 12 torpedoes (75%) scored hits in a night attack under more trying conditions.24
The problem of American AA fire could have been mitigated by better strike planning. The only difference between the “surprise” and “no surprise” plans was in the order in which the strike aircraft would approach their targets. There were no changes in the assigned targets. In the “no surprise” plan, the assumption was that the bombers “would create so much confusion and draw such heavy fire upward that the torpedo bombers could sneak in virtually unseen, securing a high percentage of direct hits and suffering little damage in exchange.”25 But all of these dive-bombers were to attack airfields, not ships or AA positions. Guns on the ships that would be firing on the torpedo bombers could generally not bear in the direction of the dive-bombers attacking Ford Island, and were out of range of the ones attacking Hickam Field. The “no surprise” plan did not, could not, divert any fire away from the torpedo bombers.
Genda and Fuchida’s move to change the order of the attack if surprise was not achieved was less than useless—it instead provided warning so that the defenders could prepare to receive the torpedo bombers’ attack. There were no provisions to directly suppress or divert those AA batteries most likely to interfere with the torpedo bombers’ runs, that is, the harbor-side batteries of the battleships and the guns of the ships moored at the Navy Yard.
The first wave included 54 dive-bombers assigned to hit airfields. Some of these could have been reasonably employed in clearing the way for the torpedo bombers, particularly considering that in the “no surprise” situation, many of the defending fighters could already be airborne and past the ministrations of dive bombers. The first bombs, instead of hitting a hanger and reconnaissance aircraft at the south tip of Ford Island, ought to have been used to suppress AA fire. There were also 45 A6M Zeros in the first wave, all of which were assigned to attack airfields. Remarkably, the nine fighters—only nine!—accompanying the torpedo bombers departed ten miles short of the harbor to strafe the Marine Corps Air Station at Ewa. Conside
ring the value and vulnerability of the torpedo bombers, these fighters ought to have escorted the torpedo bombers all the way to the target, and maintained top cover for the duration of their attack. Some of them could have provided SEAD support—a single strafing run against Bagley could have gone far towards saving several torpedo bombers, which might have resulted in another two or three hits. In other words, a single strafing run might have been worth half a battleship.
The idea that the torpedo bombers were to attack “at almost the same instant” applied only to the initiation of the attacks which, remarkably enough, they nearly achieved—torpedoes were in the water against targets at the west and east sides of Ford Island within minutes. However, there was enough time, between the following aircraft, that ships were able to sequentially engage them as they arrived (according to US witness reports) in clusters of three to five aircraft with pauses between clusters. Within each cluster witnesses reported the aircraft attacked with between six to twenty-five seconds separation between aircraft. The attack stretched out over at least eleven minutes and possibly longer, hardly a “simultaneous” attack.
Suppression of a target ship’s AA fire was a concept established in the 1930’s, and employed by the Japanese in their “massed attack” doctrine.26 The tactic consisted of fighters strafing the AA batteries of the target ships, or bombers delivering contact-fused bombs, just prior to or at the same time as the torpedo planes’ attack. The intent was to kill the gun crews or have them hiding under cover while the slow, vulnerable torpedo planes made their runs. Such measures could have been highly effective against the American battleships, which had many of their heavy AA guns (5”/25s and 3”/50s) in exposed mountings without splinter vprotection.27 This tactic was in the Japanese kit bag. It should have been employed.
Assessment: Level Bombers of the First Wave
Of the 50 800kg bombs assigned against inboard battleships, eight scored direct hits on battleships—six on inboard and two on an outboard ship.
The distribution of the 800kg bomb hits was effective. All inboard battleships were hit: two hits on Tennessee, two on Maryland, and two on Arizona, including the hit that caused the spectacular detonation of the ship’s forward magazine. Two bombs hit an outboard battleship, West Virginia. There were two additional “collateral damage” hits on the repair ship Vestal, moored outboard of Arizona. Vestal suffered severe flooding and had to be beached.
Fuchida claimed that he instructed the level bombers to concentrate on one battleship to ensure its destruction. This statement does not match with what happened in the attack. The table shows Fuchida’s possible briefed instructions to concentrate on a single battleship, the best possible distribution of attacks, the distribution actually achieved during the battle, and hits achieved.
Group Attacks on Battleship Row and Hits
Of the ten salvoes, six salvoes scored hits for 60%, two less than the 80% expected. Usually, as a rule of thumb, combat results are a third less than training results, which appear to be a good approximation in this case. Four of the hitting salvoes also put bombs into outboard ships, boosting the overall hit percentage to 10 hits out of the 50 bombs carried by the level bombers.
This was a phenomenally good performance for level bombers. The pattern mean point of impact (MPI) CEPs were likely on the order of 100 to 150 feet, compared to US heavy bombers which typically had a MPI CEP on the order of 2,000 feet. Japanese level bombers at the start of the war were very effective—they devastated Clark Field in a fatal blow to US air power in the Philippines, burned Cavite Naval Shipyard to the ground, and sank the aircraft transport Langley while she was underway during the Java campaign. In contrast, USAAC aircraft employing the celebrated Norden bombsight during the Battle of Midway dropped 291 bombs from altitudes between 4,000 to 25,000 feet for zero hits.28
The Japanese level bombers’ efforts were sabotaged by faulty weapons. Six of the ten bombs failed, seven if a bomb that passed entirely through Vestal before exploding is included. A dud rate of 5% is the norm, but 60% is outrageously high. In spite of having two independent fuzes, two of the bombs completely failed to explode, and four others had low order detonations that delivered only a small fraction of their destructive potential. The modifications that removed weight to convert the shells into bombs weakened them so that some shattered on impact.
The two bombs that penetrated the 4-inch Class “B” armor covering the turret roofs on Tennessee and West Virginia hit a joint between two armor plates and squeezed between the overlapping edges, a process which shattered the weakened bomb casing and cracked open the explosive cavity, scattering the explosive. The resulting low order detonations did only minor damage, as the turrets were empty of powder. Both turrets were repaired and their turret top armor replaced and thickened.29
Overall, it was a totally unsatisfactory performance by the Japanese weapons establishment.
The simulation was re-run to examine the potential of the attack under various conditions. The hit rate per salvo was changed to 60%. A 60% dud rate was used on bombs that hit engineering spaces. A 50% dud rate was used for magazine hits, assuming that some low order detonations might be enough to ignite powder that was broken open by the bomb’s impact. Since the magazines were shielded by gun turrets empty of powder, the probability of hitting a magazine was cut in half.
(12) AP Bombing Under Actual and Expected Conditions
The shaded bars show the original expected performance of the attack. The black bars show the results under the attack conditions. As would be expected, the expectations for blown up and crippled ships are reduced significantly. Over the 1,000 trials, 650 battleships were destroyed by magazine explosions.
(13) Distribution of Number of Inboard Battleships Sunk or Crippled
The second chart shows the numbers of battleships crippled or sunk under three sets of conditions. The “expected conditions” used the original hit rate assumptions as shown earlier. “Duds and Empty Turrets” uses the historical 60% hit rate and includes the historical dud rate, shielding of the magazines by the empty turrets, and distributes the attacks evenly over the targets. “Actual Attack Distribution” adds the actual attack distribution to the “Duds and Empty Turrets” conditions. Each bar shows the number of times out of 1,000 runs that zero, one, two, or three battleships were crippled or sunk.
The leftmost results (“Expected Conditions”) were shown earlier and are included for comparison. In about 85% of the trials either two or three battleships were crippled or sunk.
In the center set of bars (“Duds and Empty Turrets”) the most likely result is that one battleship would be crippled or destroyed, with only a 27% chance that two or three would be crippled or destroyed, down from 85%. The chance that no battleships would be crippled or sunk has risen from nearly zero to 30%, a significant risk that the entire level bombing effort could have come to naught.
The last set of bars (“Actual Attack Distribution”) shows the additional effect of the maldistribution of attacks as they occurred historically, where five groups targeted Tennessee, three groups targeted Arizona, and one group apiece attacked California and Maryland. The probability that none of the inboard battleships would be crippled or sunk increases to 35%. The expected value of the attack (the weighted average number of inboard battleships destroyed) drops from 1.02 battleships to .845 battleships, indicating that the actual attack distribution reduced the potential performance of the level bombers’ effort by about 18%.
The “surprise” plan called for the first wave D3A Val dive-bombers to wait until after the level bombers delivered their attacks. Instead they hit Ford Island early, generating clouds of smoke that interfered with the level bombers’ sight picture. Some of the level bombers aborted several passes and ended up dropping on whatever battleship was visible at the time. Thus, the 18% reduction in attack potential can partially be attributed to Fuchida’s communications error with the flare guns.
The most dramatic difference is in comparing the chanc
es for a clean sweep (destroying or crippling all three inboard battleships) under expected conditions versus the actual conditions. The chances drop from over 40% to less than 2%. The chances of getting two or three battleships drops from 85% of the trials to under 20%. Under historical conditions, the most likely result is that one battleship would be crippled or destroyed. So, the faulty bombs reduced the level bombers’ damage potential by one to two battleships.
The level bombers destroyed one battleship, meeting Japanese hopes.
The hit on the Arizona has often been represented as a “one in a million” or “lucky” hit. The simulation shows that the battleships were more vulnerable to the high level bombers than previously believed, and the loss of at least one battleship to AP bombs would be expected.
Overall, the performance of the level bombers was very good under trying circumstances. Compared to the performance of high-level bombers throughout the war, their effort was outstanding.
Assessment: Dive-Bombers of the Second Wave
Seventy-eight D3A Val dive-bombers arrived over the harbor to hit ships.30 With the 250kg GP bomb expected to be ineffective against battleships, and no carriers present, under the original prioritization scheme the dive-bombers’ assignment was to sink cruisers. Either just before or just after launch their assignment was changed. They were told to finish off ships damaged in the first wave attacks.
