The history of maritime war offers no typical cases of actions between groups of torpedo-boats and ships under way, from which to deduce tactical principles.
To this fact, perhaps, is to be attributed the great diversity of ideas that are generally held upon this important subject.
In the study of such questions as this, eminently practical methods should be employed, rather than simple theoretical analysis. But, on the other hand, when defensive action on the part of the ships is entirely lacking, only a relative value may be assigned to experiments made in time of peace, and even certain exercises of methodical simulated attacks, based only upon special suppositions, may perhaps easily violate tactical criteria, rather than furnish useful experimental data.
And indeed, in analyzing the exercises carried out in time of peace, we encounter, among other circumstances, one in particular; which is, that the torpedo-boats, sure of not running any risk, and of not being disturbed in their maneuvers, follow in the main an exactly prearranged line of conduct, and, in general, the ships do not even attempt to maneuver opportunely for preventing the attack.
Instead, it would evidently be necessary always to suppose that the adversary will not remain passive, but will make the maneuver that seems to him most suitable; this does not exclude the possibility that it might result in situations disadvantageous to him, and he who attacks must be prompt to discern them, and know how to profit by them.
*Revista Marittima Italiana, May, 1910.
The tactics, therefore, should not be rigid, composed of too precise schematic dispositions, but susceptible of adaptation to the special conditions of every single case and to diverse eventualities.
In fact, fortuitous circumstances may give occasion to special tactical initiatives; thus: fog, the moon, the state of the sea, the hydrographic conditions, the conditions of visibility, the positions of the destroyers, the zones illuminated by the search-lights, the contemporaneous discovery of more objectives in different directions, may determine different lines of conduct and have a decisive influence upon the result of the attack.
In order to profit by all these eventualities energetic initiative and quick perception are necessary, but one must not be too much hampered by precise dispositions. Thus, for example: to lose the opportunity of making a good launching solely to maintain oneself in the prescribed geometric formation would be a grave error.
Hence in the attack of a group of several torpedo-boats there is necessary the maximum co-operation among the various units, good mutual understanding, and maneuvering inspired by general principles rather than by prearranged fixed rules, because the line of conduct of each unit will depend upon the manner of directing the rays of the search-lights, upon the firing, upon the maneuvers of the ships, and upon many other unforeseen circumstances.
Therefore, it seems to me fitting to affirm very clearly that certain maneuvers often used during exercises are for the sole purpose of conventional training and have little tactical value.
In short, we must not confound the application of general principles, which may have varied forms, with submission to absolutely rigid rules.
It seems to me that this idea is not sufficiently recognized and emphasized, and that there exists rather a tendency to crystallize into special forms, without keeping well fixed and dominant in the mind those fundamental principles by which one should be inspired in the choice of a line of conduct.
These fundamental principles are certainly well known to all, but, notwithstanding this, perhaps because they were not sufficiently emphasized and very clearly formulated in synthetic manner when passing to examine the practical application, it not infrequently comes to pass that, on account of the diversity of appraisements and habits, the matter ends with the domination of certain systems that are not in complete harmony with the spirit of said principles.
To-day everything tends to establish the belief that with modern anti-torpedo-boat artillery, the torpedo-boats cannot rationally hope to reach the launching distance in the day time; therefore it seems to me that we may at once affirm that the use of the modern torpedo-boat must be limited to the night time, while by day we should rather employ submersibles, whose conditions of invisibility may be said to be ideal.
The anti-torpedo-boat armaments and the characteristics of the torpedo-boats being changed.1 it follows therefrom that it will be necessary to hold a different idea of their use from that which was held in the past, when, that is to say, the probabilities of hitting and destroying the torpedo-boats with guns were other than at present, the number of the torpedo-boats was also different, and the sacrifice represented by the loss of a certain number of them was much more limited. (In what I shall say I refer specially to sea-going torpedo-boats, armed with three launching tubes, two forward and one aft.)
It seems to me that for the tactical employment of torpedo-boats it would be necessary to be specially inspired by the idea of dispersion of the different vessels of the attacking group, and concentration of their effect.
It does not appear to me either impossible or too difficult to secure the result that the torpedo-boats of the attacking group (provided that the number of units is reasonably limited), conveniently dispersed in order to have the minimum probability of being hit and the maximum probability of not being all discovered at the same moment, may be able to hurl themselves against the target in such a way as to arrive at a launching position simultaneously or at very small intervals, and, therefore, with the maximum probability of useful effect.
Torpedo-boats are very fragile hulls without any defense whatever, which, in order effectively to use their proper weapon, must be able to arrive at a very short distance from the ships, without
1 The increase in tonnage corresponds to an increase of the general offensive capacity, that is to say, armament, speed, and sea-going qualities. To-day the number of torpedo-boats is smaller, but each one is armed with a greater number of launching tubes. For the present, then, we may expect that in one and the same attack each torpedo-boat may execute several launchings, and that the diverse units of the attacking group may be more widely dispersed. (Author's note.)
having first exercised offensive action, and while the powerful defense of the ships, composed of the guns, is still in full efficiency. It follows from this that the fundamental principle of torpedo-boat tactics must be that of surprise, and, therefore, action in the darkness of the night with several units contemporaneously, but in dispersed order, and at high speeds. And indeed, only the rapidity and impetuosity of the attack and the protection of the darkness can give good probability of arriving in safety at proper launching distance, before the ship is able to discover the attacking vessel and bring its means of defense efficaciously into action.
The simultaneous attack of a proper number of units spread out in the same sector of defense of the ship, cannot but increase the objectives for the firing, thereby rendering concentration impossible and so enormously diminishing the efficacy of the defense, and, vice versa, increasing the probability that some unit of the attacking group may bring the attack to a successful conclusion, and perhaps also without even being discovered, thus effecting a surprise.
When a torpedo-boat is discovered by the ship there is a tendency to concentrate attention upon it; and it is then that the undiscovered boats may profit by their favorable situation, withdrawing from those illuminated by the search-lights and hurling themselves toward the ship across the unlighted zones. However, the number of units that can attack simultaneously is always limited by the necessary dispersion and by the scant amplitude of the space that includes the positions and the profitable launching distances.
The importance of surprise as a factor of success can never be too strongly emphasized; in order successfully to carry it out there is necessary a proper dispersion, and also intelligent cooperation among the various units of the attacking group. For example, there must be no hesitation in sacrificing one's own unit, or in delaying one's own attack, when such a course may insure the success of the maneuvers of the others. Often, therefore, the first boats discovered will be sacrificed, and their action may consist simply in obliging the ship to definitely reveal her position by her own search-lights and gun-fire, or in continuing to draw the fire upon themselves, thus permitting the others to advance together without being discovered.
If, instead, the torpedo-boats of the same tactical group were used singly and successively, or compactly, there could be no considerable dispersion of the enemy's fire in one and the same sector of offense, and each one would be exposed to immensely greater risks.
Torpedo-boat tactics should be tactics not only of surprise but also of initiative. To obtain this end, if, on the one hand, it is neither profitable nor necessary to have too restrictive fixed forms, on the other hand it is indispensable to maneuver in order and according to definite and uniform general rules. Therefore the order must be understood to be only a means of arranging and conducting the various units prior to the attack in order afterwards to hurl them against the enemy from determined positions.
Summing up, it seems to me that the fundamental principles upon which all torpedo-boat tactics must be based may be said to be the following:
1. Contemporaneous attacks with a proper number of units in such fashion as to have sufficiently scattered order.
2. Attacks regulated in such a way that the various units of the tactical group may arrive at launching position simultaneously, or at very short intervals, without any one of the boats disturbing the maneuvers of any other.
3. Attacks regulated in such fashion that the various units may be found conveniently arranged in case of failure to effect a surprise or when the ship maneuvers to prevent the attack.
4. Attacks inspired by general principles, but not determined by rigid rules. I have already said that the orders are but a means for conducting the various units prior to the attack, and, therefore, they should present characteristics such as to respond, first, to the exigencies of the phase of search and approach preceding the attack, and that then they may permit of easily obtaining the conditions required by the fundamental tactical principles. Every order or formation should present the following requisites:
1. Elasticity, such as to permit sufficient liberty of movement to the various units, and a certain facility for changes of course.
2. Simplicity, so as not to require too much attention to maintaining the formation.
3. Safety, so as to reduce to a minimum the necessity for solicitude about collisions, in case of injuries or variations of speed, and the probability that any unit may disturb the maneuvers of another.
In considering the problem of the formations it is well also to note:
1. That too close formations of attack form a gross compact mass and gross columns of smoke, and facilitate the discovery of the torpedo-boats and the firing against them.
2. That in close formations of attack, or in those in which the various units are upon the same or nearly the same bearing with respect to the ship, it may be said that once the first torpedo-boat is discovered, the others will almost surely be discovered very soon after.
3. Too long formations interfere with mobility, not permitting the rear vessels promptly to see and know what is taking place at the head, the transmission of news and orders being difficult, and they bring to the maximum the intervals of time between the launchings of the different units.
4. In column of vessels, the intermediate units, that have not clear water either ahead or astern, may easily disturb the maneuvers of their immediate neighbors; therefore, in column, especially by night, there is always found a tendency to excessive elongation, and, when the speed is high, there may easily be dangerous closings.
5. The formations should be such as easily to permit the units to follow their respective group leaders, see and imitate their maneuvers without necessity for signals, and easily to permit the closing in of the various units for the transmission of information and orders, etc.
6. Formations should be such as to permit, first, the various units of the tactical group to discover the target simultaneously, and afterwards to dispose themselves easily, in a way to avoid the possibility of one boat masking it from another, to the end that, during the run of approach, all may keep it continually in view, follow its movements, and make a good estimation of the data for pointing.
7. Formations should be such as not easily to permit the smoke of any torpedo-boat to envelop the others—a condition that, at high speeds, is generally met with in a maximum degree, especially in columns of vessels.
8. Formations in which, by night, with lights obscured, each torpedo-boat may easily see another, that is, when the boats present their broadsides to each other, and not their breadths only, are advantageous for the necessities of nocturnal navigation with lights obscured.
9. Formations should be such as to render easy the subdivision into groups in case of meeting with destroyers, and an opportune maneuver in case of failure to effect a surprise and when the ship maneuvers in time to prevent the attack.
While not holding it to be improbable that torpedo-boats, especially if assisted by a supporting ship (nave appoggio), under the normal conditions of darkness of our nights, may be able to make out the larger masses of the ships at a distance sufficient to enable them to arrange themselves in the most convenient form for the attack while approaching to launching distance, it will have to be admitted that very bad cases could present themselves in which the torpedo-boats might find themselves, almost without warning, in contact with the ships in a way that would oblige them to attack in about the same order as that in which they arrive, without having time to arrange themselves very differently.
Therefore, the normal formation should not only be such as to be suitable to the period of search for the enemy, but should also lend itself sufficiently to the attack that in the worst cases at least some of the units may be able to attack quickly in the order in which they present themselves.
The problem of formations is intimately connected with the most convenient number of units in the squadron; a number which is limited by the sum total of all the above mentioned exigencies and especially by the matter of assuring the maximum flexibility, facility of maneuvering and transmission of orders; and also by the necessity for subdivision into two groups in order to be able opportunely to envelop the ship in given cases; and by the probability that some of the torpedo-boats may be destroyed by the guns of the ship. This number must not be greater than the limit which permits a proper dispersion and allows all the units of the tactical group to arrive almost together in good position for enabling them to launch almost simultaneously.
It is true that with the increase in number of the units of the tactical group the probabilities increase that at least one unit will arrive at launching distance without being disabled, but it is also true that exceeding a certain limit cannot but increase the difficulties of navigation, the likelihood of confusion and discovery, and above all, and in a maximum degree, the probabilities of the good effect of the guns of the ships, without returning an adequate increase of practical probability of success for the torpedo-boats.
Let us see what indications concerning this most convenient number can be obtained from consideration of the probabilities of the destruction of the torpedo-boats by the ship.
Let us suppose the time available for firing to be on the average2 one minute; the practical rapidity of nocturnal fire of modern anti-torpedo-boat calibers3 to be six shots per minute, which in practice, under the conditions of nocturnal firing against torpedo-boats, can give a probability of hitting of 0.1 for each shot (that is to say, the probable number of effective hits equals 10% with approximation of a fraction less than unity); and that a single good hit is sufficient to disable a torpedo-boat.
The shots that the ship can fire against each torpedo-boat, even under the most favorable hypothesis, are limited in number, and as a single hit suffices to disable the torpedo-boat, it will be necessary to determine the probability of placing at least one shot in the target.
In this special case, the probable number of effective hits considered is minimum, that is, equal to one; therefore, it cannot, with sufficient approximation, be held to be equal to the product of the probability assigned to each shot by the number of shots fired, since this product, being only a result approximating to within a fraction less than unity, would be affected by too great a relative error, and might lead to conclusions unworthy of attention.
Evidently, therefore, in our case, the principle of this product cannot be accepted, since, given the degree of approximation, the exact corresponding value could as well be a fraction less than unity as one. In short, the fact of having ten effective hits out of
one hundred shots, does not mean that we may have equal hopes of obtaining one out of ten.
It will then be necessary to have recourse to a rigorous determination, much more exact, by applying the laws relative to recurring events.
2 See Table 1; but it is considered that it will not always be possible to begin the firing quickly, at 1500 meters, and that the launchings can be executed at distances greater than 500 meters and with angles of impact less than 90°. (Author's note.)
3 Calibers capable of disabling a torpedo-boat with a single shot. (Author's note.)
If 0.1 is the probability of hitting, the contrary probability for each shot of missing the target will evidently be 1—0.1=0.9. If the number of shots be n, (0.9)n will express the total probability of not hitting with all those shots (that is, the event of not hitting is repeated n times), while P=.1— (0.9)n will represent the probability of placing in the target at least one of the series of n shots.
Wishing to determine the number of shots necessary in order to have almost the certainty4 of being able to place at least one in the target it will be necessary to place the value of P somewhere near unity; for example, let P=0.90, and then solve the said equation with respect to n. From this equation we have:
n=log(1-P)/log (0.9) = log (1-0.90)/log (0.9)=21.
Hence twenty-one shots must be fired in order to have a 90% Probability of placing at least one in the target. And as in the minute of profitable time, each gun can fire six shots, it follows that in order to have good probability of placing one effective hit in each torpedo-boat, it will be necessary to oppose each one with at least four guns, if the boats are in sufficiently scattered order.
But on ships at present in service, each quadrant can be covered by a maximum of ten anti-torpedo-boat guns;5 therefore, in this view of the case, on each side the number of torpedo-boats that attack simultaneously, provided they are sufficiently dispersed in a way to constitute different targets, should be greater than two.
If, for example, the defense were organized in a way that would enable it to direct the fire of two guns against each torpedo-boat attacking in the same sector, the probability of destroying each torpedo-boat with one hit would be: 1—(1—0.1)14=1-0.2=0.8; that is to say, the ship would have an 80% probability of destroying each torpedo-boat discovered, and vice versa, each torpedo-boat discovered would have a 20% probability of not being disabled before being able to launch a torpedo.
The lack of practical results based upon extensive experiments induces the to stop with the consideration of the indications sup-
4 Considering that for the ship there exists the contrary probability of not succeeding in discovering all the torpedo-boats in time, in compensation it should have at its disposal a number of anti-torpedo-boat guns sufficient to assure to it the best probabilities of disabling those discovered. (Author's note.)
5 We are considering calibers capable of disabling a torpedo-boat with a single effective hit. (Author's note.)
plied by the theory of probabilities, which can give a certain idea (doubtless preferable to none at all) concerning the forecasting of the probable useful effect of artillery in nocturnal firing against torpedo-boats, and limit the field of vague supposition.
I will suppose, for example, that the firing is begun by night against a torpedo-boat at a distance of 1500 meters with guns capable of disabling the torpedo-boat with a single effective hit, and whose rapidity of nocturnal firing is six shots a minute, which is one shot every to seconds. It is necessary to distinguish two cases, viz: of firing with fixed sight against a target not illuminated by a search-light, and of firing against a target well illuminated by the search-light and, therefore, permitting sight adjustment, although it may be difficult to accomplish it in practice. Let the relative velocity of the torpedo-boat's approach toward the ship be, for example, 150 meters in every 10 seconds; and let the target offered by the vital parts of the torpedo-boat, the amplitude of which varies according to the way in which it is presented, be on the average,6 10 meters in length and two in height.
I will suppose that an error in the sight produces a vertical displacement of the mean point of impact proportional to said error in the ratio of 0.90 meters for every 150 meters of sight error; that the firing is executed with the sight set at 1000 meters
in the case of the target not illuminated, and is also begun with the same setting of the sight when the target is illuminated; that a first adjustment of the sights to a distance measured by telemeter cannot be made until after 20 seconds and with an approximation of so meters; that the adjustment of the firing may be obtained with three salvos and that it may be maintained thereafter; that the effects of the practical errors of nocturnal gunlaying against targets so movable and so little visible may be represented by mean deviations proportional to the distances, which we will suppose the following: at a distance of 600 meters the vertical equals two meters (corresponding to two-tenths of a degree of error in elevation), and the horizontal equals to meters.
6 The mean inclination of the line of sight (see Table I) with respect to the longitudinal axis of the torpedo-boat is equal to 3e, the length of the vital part of the torpedo-boat is 20 meters, therefore the length of the vital target is 20 sin. 300 = 10 m. (Author's note.)
(corresponding to one degree of error in lateral train),7 and that they may respectively increase with the increase of the distance in the ratio of 1 meter and 5 meters for every 150 meters of further distance.
Let us see what may be the probable number of shots placed in the target in the two cases cited during the time employed by the torpedo-boat in reaching the distance of 600 meters from the ship (which signifies a launching distance of about 500 meters).
I will consider first the case of the target not illuminated. At a distance of 1500 meters we shall have the first salvo—of n shots if n be the number of pieces—with a sight error of 500 meters which will produce a depression of the point of impact with respect to the center of the target of 3 meters.
Indicating by P(x/d) the probability of hitting relative to a mean deviation d, and to a size 2x of the target, which is the probability of having a deviation less than x, we find that the probability of hitting for each of these shots will be:8
P=1/2[P(3+1/8)-P(3-1/8)] P(5/40)=0.310-0.157/2 x 0.076
=0.153 x 0.076/2=0.005814.
And, therefore, the probable number of effective hits for this first salvo of n shots will be (with an approximation of a fraction less than unity):
N=n X 0.005814.
For the second salvo made at a distance of 1350 meters, similarly we have:
P=1/2(P(2.10+1/7)-P(2.10-1/7)] P(5/35)=0.274-0.094/2 x 0.088
=0.180 x 0.088/2=0.00792. N=nx0.00792.
7 We must consider that, especially by right, the center of figure of the visible parts of the torpedo-boat, which is what attracts the attention of the gull-pointer, is always considerably more elevated and laterally displaced with respect to the center of the vital parts which constitute the profitable target; that the time conceded for pointing and firing with the established rapidity of lire is only 10 seconds; that the mobility of the target is considerable and its visibility minimum; that the inevitable variations of intensity of the light of the search-light produce very detrimental effects; that it is not easy to see a torpedo-boat through the port-hole of a redoubt. (Author's note.)
8 In these calculations I have used the tables and formulas in Alger's Exterior Ballistics. (Author's note.)
For the third salvo at the distance of 1200 meters we have:
P=1/2[P(1.20+1/6)-P(1.20-1/6)] P(5/30)=0.225-0.019/2 x 0.1
=0.206x0.1/2=0.0103. N=n x 0.0103.
For the fourth salvo at the distance of 1050 meters we have:
P=1/2[P(0.30+1/5)-P(0.30-1/5)] P(5/25)=0.163-0.088/2 x 0.127
=0.251x0.127/2=0.015938. N=n x 0.015938.
For the fifth salvo at the distance of goo meters we have:
P=1/2[P(0.60+1/4)-P(0.60-1/4)] P(5/20)=0.250-0.064/2 x 0.157
=0.314x0.157/2=0.024649. N=n x 0.024649.
For the sixth salvo at the distance of 750 meters we have:
P=1/2[P(1.5+1/3)-P(1.5-1/3)] P(5/15)=0.492-0.1/2 x 0.207
=0.392x0.207/2=0.04572. N=n x 0.04572.
For the seventh and last salvo at the distance of 600 meters we have:
P=1/2[P(2.4+1/2)-P(2.4-1/2)] P(5/10)=0.825-0.424/2 x 0.310
=0.401x0.310/2=0.06155. N=n x 0.06155.
Therefore, the total number of probable effective hits during all seven of the salvos will be:
n x [0.005814+ 0.00792+ 0.0103+ 0.015938+0.024649+0.04572 +0.06155]=n X 0.171891, that is, 17%.
When the target is illuminated by the search-light we have a first salvo with the sights set at 1000 meters, with a distance of 1500 meters under the same conditions as those of the first salvo in the case above analyzed, which, therefore, will give the following result:
N=n x 0.005814.
Similarly for the second salvo at a distance of 1350 meters, also with the sight set for moo meters, we shall have:
N = n x 0.00792.
At the third salvo at a distance of 1200 meters the conditions change since the sight will be set for the first measured distance with an approximation of 50 meters in this case we shall have:
P=1/2[P(0.30+1/6)-P(0.30-1/6)] P(5/30)=0.133-0.088/2 x 0.1
=0.221x0.1/2=0.011. N=n x 0.011.
At the fourth salvo at the distance of 1050 meters the error in the setting of the sight is reduced to 25 meters, and we have:
P=1/2[P(0.15+1/5)-P(0.15-1/5)] P(5/25)=0.145-0.206/2 x 0.127
=0.251x0.127/2=0.015938. N=n x 0.015938.
At the fifth salvo at a distance of 900 meters the firing is adjusted, and we have:
P=P91/2) P(5/10)=0.310 X0.310=0.0961. N=n X 0.0961.
Then in this second case the total number of effective hits during all seven of the salvos is:
n x [0.005814 + 0.00792 + 0.011+ 0.015938+ 0.024649+ 0.042849
+ 0.0961] =n x 0.204270; that is, 20%.
In practice, when the target is not illuminated, unless the night is exceptionally clear, the errors committed are much greater than those supposed, and, therefore, the percentage of effective hits never amounts to seventeen, but suffers a notable reduction. Similarly, as the immediate and continuous illumination of the discovered torpedo-boat is not always accomplished, and adjusting the firing and following it in the pointing is not easy, nor is it always possible to illuminate all the torpedo-boats that attack contemporaneously if they are sufficiently dispersed, and to have the intensity of the light constant, the calculated percentage of 20% may be held to be a mean approximating to the truth even admitting that when the target is illuminated the practical errors of pointing may produce errors smaller than those supposed.
Therefore, it appears to me that we may argue that, the number of hits being always small, limiting the percentage of effective hits to 10%, in the general average of the cases, cannot be held to be too pessimistic, since it is also necessary to consider that the above mentioned calculations have been made completely ignoring many other disturbing factors of moral and material nature the influence of which may be enormous.
From what has been said it appears that the principle of the dispersion of the various simultaneously attacking torpedo-boats is also imposed in order to avoid having several of them illuminated at the same time by the same search-light, which would bring about a condition immensely advantageous for the firing.
In general, ships at present in service have not more than four search-lights on each side, so that the number of torpedo-boats that, being sufficiently dispersed, can be illuminated at one and the same time, will be, at the maximum, four on each side. Excessively increasing the number of units in the attacking group means the abandonment of that proper dispersion which can give good probabilities of success in the surprise, and the assemblage of too many units crowded together in excessively compact order in the limited space included by the distances and the proper positions for launching;9 which is not well since, even without preoccupying oneself with the amount of the losses, one must take into consideration the moral effect and the disorganization produced by the sinking of units very close to each other.
Under the present conditions of launching distances, angles of impact, etc., it is held in general that the most convenient number of units of the squadron is four, or at most six; and many think that the tactical unit should be the half squadron, that is to say, the section; and this answers to the idea of action in subdivided groups, which is imposed in case the ship maneuvers to avoid the
9 The angular amplitude of this space diminishes with the increase of the ship-torpedo speed ratio. (Author's note.)
attack. (It is said that the Japanese attacked in groups of two units; Labres upholds groups of three units each.)
Some nations, as France, Germany and the United States, add to the squadron a unit of greater tonnage, that is to say destroyers, which, having 'greater scouting efficiency, can conduct the two sections of torpedo-boats and at the same time contribute either to the attack on the ship as a torpedo-boat or to the defense against destroyers.
The best conditions for the maneuver of attack of the torpedo-boats must be determined by necessities of a different nature and by their opportune combination, that is, with respect to the gun in a way to have the minimum probability of being discovered and hit, and with respect to the launching, in a way to have the maximum probability of hitting the ship. As regards the manner of confronting the fire of the ship it is necessary to consider:
1. That ships, in nights of normal darkness, cannot hope to succeed in sighting torpedo-boats at a distance greater than a mile.
2. Experience has demonstrated that by night, firing with rapid-fire guns, at a taget 6 meters long and 2.20 meters high, there is rarely obtained a percentage of effective hits greater than 3%. Lieutenant Dewar, speaking of experiments carried on in England, cites a result of only 6 shots placed in the target out of 128 fired with guns of 76 mm., with a firing distance a little greater than the normal launching distance.10
3. That the time in which a torpedo-boat remains exposed to fire before arriving at launching distance is very short, that is, about one minute.
Supposing the firing to be begun on an average at a distance of 1500 meters, the ship's speed to be 14 knots, the speed of the torpedo 34 knots, and the launching distance 500 meters, the times during which the torpedo-boat remains exposed to fire and the variation of distance and bearing for each interval of 10 seconds, for the two extreme cases, of attack on opposite courses at a lateral interval of 500 meters and attack with a normal approach, and an angle of impact of 90° in both cases, are shown in the following table:
10 It would be greatly interesting to make lengthy experiments in night firing with the ship in motion against a target towed at high speed and which is presented by surprise. (Author's note.)
TABLE 1.
Times of firing. | Speed of torpedo-boat =20 kts. Speed of ship =14 kts. Speed of torpedo=34 kts. Launching distance=500 meters. Angle of impact=90°. | |||
| Attack when on opposite parallel courses. | Attack when on a normal course. | ||
| Distance torpedo-boat to ship. | Bearing from ship’s head. | Distance torpedo-boat to ship. | Bearing from ship’s head. |
Seconds. | Meters. |
| Meters. |
|
0 | 1500 | 19° 30’ | 1500 | 60° |
10 | 1342 | 22° | 1387 | 60° 10’ |
20 | 1185 | 25° | 1265 | 60° 30’ |
30 | 1032 | 29° | 1142 | 60° 40’ |
40 | 890 | 35° | 1025 | 61° 30’ |
50 | 752 | 41° 30’ | 900 | 62° 30’ |
60 | 635 | 51° 30’ | 780 | 63° 30’ |
70 | 547 | 66° | 660 | 65° |
71.5 | 540 (Launch) | 68° | … | … |
80 | … | … | 540 (Launch) | 68° |
NOTE.—In case of attack on a normal course no account is taken of the change of course that it may be necessary to make in order to execute the launching if there are no bow launching tubes or torpedoes with angle gyroscopes. (Siluri angolati.)
It results from this table that in case of attack on a normal course the torpedo-boat remains continually, during the 80 seconds, in the sector of 600 forward of the ship's beam, that is to say, in the sector of maximum offense; while in the case of attack on a parallel course it enters it only after the first 30 seconds, remaining therein 41 seconds. Moreover, it is shown that the depth of the target exposed by the torpedo-boat to the fire of the ship, determined by the inclination of the line of fire to the longitudinal axis of the torpedo-boat, an inclination that is equal to the bearing given in the table in the case of the parallel course, and to its complement in the other case considered, is greater on the parallel course during the first 30 seconds, while afterwards it becomes continually and rapidly greater in the attack on the normal course.
4. That the probabilities of hitting with the guns increase with the increase in compactness of the formations of attack; since, when the torpedo-boats are in very close order, they almost constitute a single very extended target, easily visible, and, vice versa, they diminish with the increase of the dispersion of the torpedo-boats in the same sector and in a direction normal to the line of fire.
5. That the same probabilities increase with the increase in depth of the formation of attack in the direction of the bearing from the ship, so that they are at maximum when the various units keep themselves on or near the same bearing with respect to the ship, thus remaining exposed to an enfilading fire.
6. That the same probabilities increase if the torpedo-boats make considerable changes of course; since, in that case, the relative rapidity of movement greatly diminishes. In fact, even torpedo-boats that have the best maneuvering qualities, in order to complete the first 90° of change of course, occupy not less than 35 seconds at full power, and during that time their change of position is minimum.
I have forborne to cite all the other well known conditions that determine the probability of hitting a torpedo-boat with the gun; such as, the time during which it remains exposed to fire, sectors of maximum offense, distance and rapidity of change of bearing.
In a typical attack of torpedo-boats we may distinguish three successive phases:
1. The getting into contact by making out the ships; 2, the maneuver for taking up good position for launching; 3, the execution of the launching.
It is not necessary to believe that, in our waters and with modern torpedo-boats, in which the lookout is sufficiently elevated, on a night of normal obscurity and when in a good situation with respect to the land, the eventuality of sighting the larger masses of the ships, especially if they are grouped together, before the ships can discover the torpedo-boats, must be held to be a fortunate circumstance and very improbable. Experience has demonstrated that while ships, under the most favorable conditions, cannot hope to make out torpedo-boats at a distance greater than one mile, even when the destroyers have signalled their approach, on the contrary, torpedo-boats stand a good chance of making out ships between two and three thousand meters away.
The torpedo-boats that are successful in evading the destroyers and passing through their line, having seen the search-lights or the vague and confused black masses of the ships, will approach still nearer in order the better to make them out and to ascertain the direction in which they are moving, and then perhaps withdraw again, if necessary, and move ahead in order to take up favorable positions for launching.
Let us then consider the conditions of launching. In this connection it is necessary to note:
1. The launching tubes, in one and the same attack, can execute only one launch each against the enemy, being afterwards exposed to almost certain destruction; therefore, in their use great care must be taken to put them in conditions for having good probabilities of hitting.
2. Notwithstanding the increase in speed of modern torpedoes, launching from great distances against moving ships cannot be held to be of use, since it is difficult to estimate the data for pointing, that is to say, the speed and course of the enemy, and with the speeds of present torpedoes the influence of errors in the speed and course of the enemy cannot yet be said to be negligible.
3. Generally the speed of the enemy may be more accurately estimated than the course. In this connection it must be considered that, for an observer sufficiently distant, the profiles of two ships that are following directions almost symmetrical with
{figure}
FIG. 1.
respect to the normal to the mean line of sight, are the same, as indicated in the figure. Expert observers can estimate the speed of a ship with an approximation of about three miles, and the course with an approximation of about 20°.
4. The probability of hitting with the torpedo increases with the diminution of the launching distance, which, however, should never be less than 100 meters to the end that the torpedo may have time to regulate its course. The poor results obtained with the torpedo by the Japanese in the last war are to be attributed principally to the fact that they desired always to use the torpedo at the extreme limit of its range and, therefore, often finished by launching at too great distances.
5. The opportune launching distance diminishes with increase of speed of the target; therefore, as the speed of the ship increases, it will be well to diminish the launching distance.
6. Keeping the ratio between the speed of the target and that of the torpedo not greater than one-half, as is to-day practically the case, the probability of hitting a target of the dimensions of an ordinary ship, and which, therefore, presents an exposed surface which varies with the inclination of the trajectory of the torpedo to the course of the ship, that is to say, the angle of impact, is maximum for an angle of impact of 90°, and remains very good as far as an angle of impact of 30°,11 while it rapidly diminishes as soon as the said angle becomes greater than 90°, as is indicated in Table 2.
TABLE 2.
NUMBERS PROPORTIONAL TO THE PROBABILITY OF HITTING WITH THE TORPEDO.
Angle of impact. | Ratio between the speed of the target and that of the torpedo. | ||
Degrees. | ½ | 1/3 | ¾ |
0 | 3.88 | 4.61 | 5.40 |
10 | 3.58 | 4.27 | 5.03 |
20 | 6.42 | 7.83 | 9.29 |
30 | 8.52 | 10.75 | 13.00 |
40 | 9.89 | 13.18 | 16.22 |
50 | 10.68 | 15.10 | 19.14 |
60 | 11.09 | 16.87 | 22.03 |
70 | 11.34 | 18.65 | 25.16 |
80 | 12.91 | 21.05 | 29.42 |
90 | 12.98 | 25.14 | 36.26 |
100 | 7.90 | 16.48 | 24.58 |
110 | 5.17 | 11.51 | 17.69 |
120 | 3.55 | 8.37 | 13.14 |
130 | 2.50 | 6.14 | 9.87 |
140 | 1.78 | 4.52 | 7.38 |
150 | 1.24 | 3.22 | 5.34 |
160 | 0.81 | 2.12 | 3.56 |
170 | 0.41 | 1.09 | 1.84 |
180 | 0.44 | 1.16 | 1.96 |
NOTE.-The above table was calculated by P. R. Alger on the basis of actual conditions, that is to say, a target not represented by a geometrical point, but having dimensions and consequently an exposed surface that varies with the angle of impact.
7. The same probability of hitting with the torpedo (supposing the speed of the torpedo to be not less than 34 miles) is sufficiently good up to distances not greater than 700 meters (41 seconds of run with a speed of 34 miles), while it rapidly diminishes at greater distances.
8. The launching distance is also limited by the action of avoiding the possibility that the ship, during the run of the torpedo,
11 Supposing the angle of impact counted from 0° to 180° to be reckoned from the bow. (Author's note.)
may have time to execute a change of course sufficiently wide to enable it to avoid the torpedo, or to render the angle of impact so small as to exclude the possibility of the functioning of that weapon. Under the actual conditions of speed of torpedoes and with the present maneuvering qualities of ships (about 80 seconds of time being necessary in order to complete the first 90° of change of course) it is not advisable, even in the most favorable case, to launch at distances greater than 700 meters (41 seconds of run).
TABLE 3.
MAXIMUM DEVIATIONS WITH RESPECT TO A TARGET REPRESENTED BY A GEOMETRICAL POINT.
Calculated on the basis of the following data:
Speed ratio torpedo to target = 2.
Error in estimation of target's speed = 1/10.
Error in estimation of target's course = 1/5 (that is, about one point, 11 1/4°).
In the table it is supposed that the two errors produce deviations in the same direction, the worst case.
Bearing at the instant of launching, counted from the head of the target. | Distances in meters. | ||
500 | 750 | 1000 | |
Degrees. | Meters. | Meters. | Meters. |
0 | 33 | 50 | 67 |
10 | 36 | 54 | 72 |
20 | 39 | 58 | 77 |
30 | 41 | 62 | 82 |
40 | 43 | 65 | 86 |
50 | 45 | 67 | 89 |
60 | 45 | 67 | 90 |
70 | 44 | 66 | 87 |
80 | 38 | 57 | 76 |
90 | 33 | 50 | 67 |
100 | 49 | 73 | 98 |
110 | 65 | 97 | 129 |
120 | 80 | 120 | 159 |
130 | 92 | 138 | 184 |
140 | 102 | 153 | 204 |
150 | 108 | 162 | 216 |
160 | 109 | 164 | 219 |
170 | 107 | 161 | 213 |
180 | 100 | 150 | 200 |
Note—This table was calculated by P. R. Alger considering the deviation to be the distance between the target and the torpedo at the moment when the line joining the target and the torpedo is normal to the trajectory of the torpedo.
9. An angle of impact of 30° is sufficient to insure the functioning of modern torpedoes. And I insist upon the fact since it seems to me that there continually exists a tendency to exaggerate the importance of an excellent angle of impact to the detriment of so many other indispensable conditions for enabling one to have good probability of arriving in position and in condition to execute the launching of the torpedo.
10. Launchings executed at excessive distances, beyond the range of the torpedo, may also be effective, thanks to the speed of the ship, when, the angle of impact being less than 90°, the ship in its movement advances toward the torpedo and has a component of speed which is to be added to that of the torpedo; while the error in distance may more easily prevent the torpedo from arriving at the target, if the angle of impact is greater than 90°.
11. While for a torpedo-boat situated in a sector sufficiently forward of the ship's beam the run of the torpedo is always less than the distance from ship to torpedo-boat at the moment of launching, for a torpedo-boat situated in a sector abaft the beam the opposite is the case.
12. The pointing in launchings executed on the change of course is less difficult than is supposed, since one is prepared for the angular movement, which is uniform, while in pointing with a constant course one may, especially with flat hulls, have occasional easily sensible yawings.
13. The estimation of the speed of the target and the launching distance is much more easy when the course of approach of the torpedo-boat is normal to the course of the target than when it is the contrary.
14. The launchings of the various units should always be made from positions such as to exclude the possibility of the torpedoes on their proper courses being run into by the torpedo-boats.
15. The probabilities of hitting increase when the different units of the attacking group maneuver so as to execute simultaneous or nearly simultaneous launchings, which present possibilities of the compensation of errors and give the same advantages recognized in salvos of artillery. If p is the probability of hitting with each single launching, P= 1—(1—p)n. will be the probability of hitting with at least one torpedo when the number of the launchings is n, and p being always nearer to zero than to unity, evidently P will always be considerably greater than p.
16. The principles of probability applied to torpedo launchings can give us a certain approximate idea of the relative value of launchings executed under different conditions.
Let us suppose that the mean error in the estimation of the speed of the target is, at the maximum, ?V/V=1/3; that the error in the estimation of the course is, at the maximum, ?R=1/2 (or about 30°, unity representing a radian); that the speed ratio of ship and torpedo is 1/2; and that the distance from ship to torpedo-boat at the instant of launching is 600 meters.
I will first consider an intermediate launching position situated in the zone of the more advantageous positions, corresponding to the bearing of torpedo-boat from ship of 40°, that is, to an angle of impact of about 60°.
From the accompanying Table 4 we take the deviations due to the two errors in the estimation of the speed and of the course, which, with the established data, give the following:
D' =600 x 1/2 x 0.2558=51.16, that is to say about 50 meters.
D" =600 x 1/2 x 0.3042=91.26, that is to say about 90 meters.
Therefore, the mean value of the total mean deviation resulting from these two independent deviations will be:
2√502 +902=102 meters, that is to say about 100 meters.
Passing to the consideration of the case of launching from the less advantageous position in the zone further astern, corresponding to angles of impact greater than 90°, taking for example the intermediate position corresponding to ship-torpedo-boat bearing of 90° (symmetrical with that corresponding to 60° already considered), that is, to the angle of impact of 120°, we have for the value of the mean deviation about 130 meters.
Having considered the cases of mean angles of impact of 60° and 120°, if the ship is too meters long the normal surface of target exposed will evidently be:
100 x sin 60°= 100 x 0.866=86.6; about 80 meters.
With mean deviations of too and 130 meters, the probabilities P and P' of hitting a target 80 meters long, that is to say, of having a deviation less than 400, are (using Alger's tables):
P(40/100)=0.250, or 25%
P(40/130)=0.18, or 18%
But I have already said that if, for example, against every torpedo-boat the ship can direct the fire of two guns, under the conditions established on pages 13 and 14, every torpedo-boat discovered by the ship at a distance of 1500 meters will have a 20% probability of arriving at launching position without being disabled. Therefore, the total probability of realizing this condition and at the same time the other one of hitting with the torpedo (I am considering that the torpedo-boat makes but one launching), will be for every single torpedo-boat discovered, the following, in the two cases considered:
0.20 X 0.25=0.05 Or 5%. 0.20 X 0.18=0.036, or 3%.
If the torpedo-boats concurring in the attack were three in number, sufficiently dispersed to constitute different targets, and should all three succeed in launching their torpedoes, and the three launchings were executed under the same conditions, that is to say simultaneously, neglecting the small differences due to slightly different angles of impact, and admitting the same conditions established above, the probability of hitting the target with at least one of the three torpedoes would be, for the two cases considered:
1 —(1 —0.25)2 = 1 — 0.421875 = 0.578125, or 57%.
1 —(1 —0.18)3 = 1 - 0.6 = 0.4 or 40%.
Hence the enormous advantage of launching simultaneously clearly appears. But for each of the three torpedo-boats that attack together the probability of not being hit by the guns is 0.20, and, therefore,
1—(1-0.20)3=1-0.5=0.5,
expresses the probability that at least one of the three units discovered by the ship will arrive at launching position without being disabled.
And finally. the total probability that besides this condition there may at the same time be also realized another, that the fortunate torpedo-boat that succeeds in launching its torpedo may hit the ship, will be the following for the two cases considered:
0.5 x 0.25=0.125 or 12%
0.5 x 0.18=0.090 or 9%.
TABLE 4.
Bearing of torpedo-boat from ship’s head at the instant launching | Angle of the trajectory of the torpedo with said bearing. | D1/S ?v/v | D2/S ?R |
Degrees. | Degrees |
|
|
0 | 0° | .0 | .3333 |
10 | 4° 59’ | .0586 | .3321 |
20 | 9° 51’ | .1193 | .3278 |
30 | 14° 29’ | .1843 | .3193 |
40 | 18° 45’ | .2558 | .3042 |
50 | 22° 31’ | .3329 | .2799 |
60 | 25° 40’ | .4174 | .2410 |
70 | 28° 02’ | .5053 | .1839 |
80 | 29° 30’ | .5649 | .0996 |
90 | 30° 00’ | .6667 | .0 |
100 | 29° 30’ | .7221 | .1273 |
110 | 28° 02’ | .7482 | .2723 |
120 | 25° 40’ | .7379 | .4260 |
130 | 22° 31’ | .6881 | .5776 |
140 | 18° 45’ | .6031 | .7175 |
150 | 14° 29’ | .4828 | .8362 |
160 | 9° 51’ | .3370 | .9258 |
170 | 4° 59’ | .1700 | .9820 |
180 | 0° | .0 | 1.000 |
Ship’s speed / Torpedo’s speed = ½
D1= Deviation due to error in estimation of the speed.
D2= Deviation due to error in estimation of the course.
S= Distance from ship to torpedo-boat at the instant of launching.
?v/v=error in estimated speed / ship’s speed.
?R =Error in estimated course expressed in circular measure.
NOTE.-This table was calculated by Alger.
So that, under the above mentioned conditions-which are anything but optimistic-in the simultaneous attack of three torpedo-boats that are discovered by the ship, all at the same time, at a distance of 1500 meters, and are sufficiently dispersed to constitute
different targets, launching at a distance of about 600 meters with a mean angle of impact of 60°, we may expect a 12% probability that at least one will succeed in launching a torpedo and hit the ship with it, while if the mean angle of impact were 120° there would be only about 9% of probability.
Admitting that we may have a 60% probability that at least one of the torpedo-boats of the attacking group will succeed in reaching the launching position undiscovered, and thus surprise the ship, let us determine what should be the minimum number of groups of three units each, that must attack the ship from advantageous positions, supposing that all the attacks are made under the same conditions, in order to have, even with the torpedo-boats discovered and exposed to fire, a 40% probability of success against the ship.
If the probability of success for the attack of a single discovered group is 12%, evidently we shall have:
n= log(1—0.4)/log (1-0.12) = log 0.6/log 0.88 = 1.7782/1.9445 = 0.2218/0.0555 = 4
that is to say, against each ship there should be hurled at least four groups of three units each.
At this point the question spontaneously presents itself whether, having four sections of three units each, so as to be able to hurl to the attack two of them on each side of the ship, it would be well to make simultaneous attacks with the two sections on the same side so arranged that one of them may be in a position further astern than the other with respect to the ship, as indicated in Fig. 7, or successive attacks, both of them from advantageous positions, as shown in Fig. 6.
The probability of success for an attack of a group of three units, executed from an advantageous position (mean angle of impact, 60°) is 12%, and the relative probability for a similar attack from the position further astern (angle of impact=120°) is 9%, therefore, the total probability for these two simultaneous attacks will be 12+9 = 21%.
If, on the other hand, the two attacks were successive and both from the advantageous position, the ship could increase the number of guns directed against each torpedo-boat, after having ascertained that the attack is made in a single sector; for example, after 20 seconds, that is at the fourth salvo, there could be concentrated in the forward sector a part of the guns destined for the defense of the after sector. In this case, therefore, if at the fourth salvo the fire of three pieces could be directed against each torpedo-boat instead of two as at first supposed, the probability of success for each of the two successive attacks would be:
1 — (1 — 0.1 )18 = 1 — (0.9)18 = 1 — 0.149 = 0.851
1 — (1-0.149)2 = 1 - 0.616 = 0.384
P = 0.384 x 0.25 = 0.096.
And as a total for the two successive attacks we should have:
0.096 x 2=0.192 or 19%.
This shows that there is about the same probability of success in the case of the simultaneous attack of the two groups as in the case of successive attacks. In practice the advantage of the successive attacks from favorable positions for launching increases, since, evidently, the guns destined for the defense of the after sectors of the ship cannot direct their fire on the torpedo-boats that present themselves in the forward sectors after only 20 seconds as has been supposed.
17. In attacks on opposite parallel courses the launching tubes can be trained ahead up to the maximum angle, equal to the minimum angle of impact admissible, while in attacks on normal courses it will be well to keep the forward tubes at the maximum angle of chase thereby to render minimum the change of course necessary for launching.
18. If it is true that launching at short distance increases the probability of hitting, it is also true that in waiting to arrive at short distance one risks being sunk and not launching at all. So that it seems to me that for torpedo-boats that have several launching tubes, and that are advancing under fire, it may be better to execute the first launch at a distance not the minimum, which gives sufficient probability of hitting, and the others at shorter distances.
The most convenient launching distance for a torpedo-boat discovered, and, therefore, exposed to fire, should be established on the basis of the probability of not being disabled by gun-fire and of the probability of hitting with the torpedo, if these probabilities could be known with sufficient approximation.
Evidently the first probability diminishes with the diminution of the distance while the second increases, and, therefore, the most convenient launching distance will be that corresponding to a maximum of the total probability that expresses the contemporaneous realization of these two happenings.
In order to have some idea—if not an exact one, certainly less vague than none—of the relative value of the various launching distances, we turn to the first example cited on page 1111 from which it results that n x 0.17 is the number of hits effective against torpedo-boats during the time employed in arriving at the distance of 600 meters from the ship, if the number of guns directed against it is n.
Similarly from the same example we find that the number of effective hits in the time necessary for arriving at the distances of 750, 900, 1050 and 1200 meters are, respectively:
n x 0.11, n x 0.06, n x 0.03, n x 0.02.
If the number n of the guns directed against each torpedo-boat were equal to two, we find that, in the times employed in reaching the distances considered, the probabilities of placing at least one shot in the torpedo-boat are as follows:
At the distance of 600 meters P’ = 1 — (1— 0.17)14 = 1 — 0.073
“ “ “ “ 750 “ P” = 1 — (1— 0.11)13 = 1 — 0.247
“ “ “ “ 900 “ P”’ = 1 — (1— 0.06)10 = 1 — 0.538
“ “ “ “ 1050 “ P”” = 1 — (1— 0.03)8 = 1 — 0.783
“ “ “ “ 1200 “ P””’ = 1 — (1— 0.02)6 = 1 — 0.885
and vice versa, the probabilities of reaching the said distances without being disabled will be for the torpedo-boat:
For the distance of 600 meters = 0.073
“ “ “ “ 750 “ = 0.247
“ “ “ “ 900 “ = 0.538
“ “ “ “ 1050 “ = 0.783
“ “ “ “ 1200 “ = 0.885
Supposing that, with a mean angle of impact of 60° (ship's speed = 1/2 torpedo speed; the angle of impact of 60° corresponds to the ship-torpedo-boat bearing of 40°), the mean practical deviations relative to the launching are, respectively, 100, 200, 300, 400 and 500 meters for the distances considered, the probabilities of hitting a target whose exposed surface normal to the trajectory of the torpedo is 80 meters (angle of impact 60°) will be represented in the following manner:
For the distance of 600 meters P(40/100) = 0.250
“ “ “ “ 750 “ P(40/200) = 0.127
“ “ “ “ 900 “ P(40/300) = 0.082
“ “ “ “ 1050 “ P(40/400) = 0.064
“ “ “ “ 1200 “ P(40/500) = 0.04
Therefore, the total probabilities relative to the contemporaneous realization of the two happenings of not being hit by the guns and of hitting with the torpedo, will be:
For the distance of 600 meters = 0.073 x 0.025 = 0.01825
“ “ “ “ 750 “ = 0.247 x 0.127 = 0.03136
“ “ “ “ 900 “ = 0.538 x 0.082 = 0.04411
“ “ “ “ 1050 “ = 0.783 x 0.064 = 0.05011
“ “ “ “ 1200 “ = 0.885 x 0.04 = 0.03540
These data give us to understand that the most advantageous ship to torpedo-boat distance for launching with a mean angle of impact of 60°, considered in regard to the profitable effects of the gun and of the torpedo, is about 1050 meters, corresponding to a launching distance of
1050 x sin 40°/sin 120° = 1050 x 0.74 =777 or about 750 meters.
19. It seems logical to me that the torpedo-boats which have several launching tubes should increase the probability of hitting the target by launching from all the tubes rather than economize torpedoes with the grave risk of carrying them to the bottom along with themselves.
20. In order to diminish the probability of being discovered it will be necessary to guard against the possible emission of flame and too much smoke from the funnels, and, therefore, it cannot be said to be well to force the speed too much before being discovered.
After what I have said it seems evident to me that, without going further, we may establish the following:
1. That on the basis of the considerations above set forth, the most convenient positions for starting for an attack are, from all points of view, those included in the two quadrants forward of the ship's beam, and corresponding to an angle of impact not greater than 90°; a fact well known to all. It follows from this that it will always be desirable to gain a position sufficiently forward of the beam before starting on the attack.
2. That, without doubt, it is to be held convenient to attack successively with several squadrons and contemporaneously from both sides of the ship, or, according to the possibilities, successively at intervals from one and the same side, etc., but it will never be advisable to abandon the proper dispersion of the units of the same squadron or section. Attacks conducted from both sides of the ship simultaneously greatly disconcert the ship, and, given the natural tendency to follow the first torpedo-boats discovered, attacks at successive intervals present the probability of having a certain number of torpedo-boats which may arrive undiscovered at position for launching while attention is entirely directed to the preceding ones already discovered.
It is clear that the position for launching against a ship underway, with an angle of impact not greater than 900, nearest the beam that a torpedo-boat can occupy in the forward quadrants, is that which corresponds to a torpedo's track normal to the course of the ship; and, therefore, the angle, reckoned from the ship's head,
{figure}
FIG. 2.
which includes all the positions advantageous for launching, depends upon the ship-torpedo speed ratio. Supposing this ratio to be not greater than one-half—as is the case in practice the zone of the positions most favorable for launching is included in the angle of about 60°, whose cotangent is equal to one-half.
Always supposing the ship-torpedo speed ratio to be not greater than one half, the condition that the mean run of the torpedo must be 600 meters, together with the other condition that the angle of impact cannot be greater than 135°, determine the extreme position that a torpedo-boat may occupy in the after quadrant of a ship in order to execute the launching. (See Fig. 2.)
Solving the triangle formed by the track of the torpedo having an angle of impact of 135° and the contemporary track of the ship (speed ratio=1/2) we find that the aftermost position of the attacking torpedo-boat can be, at the maximum, that which corresponds to a bearing of 107° from the ship's head, which is not more than about 20° abaft the beam.
Hence we may distinguish two zones profitable for launching: the most advantageous limited by the said angle of 600, and the other corresponding to angles of impact greater than 90°, which extends as far as 20° abaft the ship's beam.12
This being laid down let us pass to examine the most important and culminating moment of the attack, which is that of launching the torpedo, thence to deduce from the sum total of the conditions set forth what may be the best disposition of the different units of the squadron for the execution of the launching, which is the final object toward which all must strive and to the good result of which all must sacrifice.
I have already said that the ideal conditions towards which it is necessary to tend are dispersion of the different units and concentration of their offense; only in this way can we have good hopes of success in the surprise and of enormously increasing the total probabilities of hitting with simultaneous launchings, probabilities that for each single launching considered alone are minimum.
It follows from this that the various units of the tactical group should arrive contemporaneously in position for launching.
But the space that includes the positions profitable for launching is determined not only by the sector of 60° forward of, and 20° abaft the beam of the ship, but also by the limitations of the angle of impact which I will set down at 45° on one side and 135° on the other, and by the mean distance convenient for launching which I establish at 600 meters.
The different units of the same tactical group should, therefore, under the best conditions, arrive contemporaneously on the arcs AB and AD (Fig. 2). Considering, instead of the arcs, the chords of the arcs, which differ very little from them, it is evident that the lines of maximum length along which the torpedo-boats may be dispersed in order to occupy contemporaneous positions for launching are AD=AB=2BC=2 x 600 x sin 22°=600 x 2 x 0.37 =444.
Hence the maximum length of the arrangement in scattered order which permits concentration of launchings is about 400 meters on each of the aforesaid lines. And as, for the principle of dispersion, it cannot be said to be convenient to close too much the distances between the various units that join in the simultaneous attack, it seems to me that the number of the said units can never be greater than three for each of the two lines. It is concluded then that we may affirm that the squadron can, at the maximum, be composed of six units, and that the tactical unit must be the section; in this connection I mention again that it
{figure}
FIG. 3.
appears that the Japanese at times attacked successively by sections composed of two units each (see "The Truth about Port Arthur," page 99, line 17), and that Labres establishes as the tactical unit the group formed of three torpedo-boats.
The lines AB and AD of the formation not only respond to the principle of the concentration of the launchings, but for the most part are also excellent lines of dispersion for confronting the fire of the guns, being nearly normal to the lines of fire.
The torpedo-boats dispersed along the most advantageous line AB can have various courses whose extreme limits are, on the one side, courses opposite to that of the ship, and on the other, courses normal thereto.
12 If the ship-torpedo speed ratio could be kept less than 1/2 the angular extension of the zones profitable for launching would be greater. (Author's note.)
The first case, with opposite parallel courses, is doubtless the most advantageous because it permits the maximum relative swiftness of approach and keeps the torpedo-boats, as is seen from Table 1, for about half the time in the ship's sector of minimum offense; therefore, wherever possible, it should always be preferred,13 also because, in night attacks, whatever maneuver the ship may make, the two sections, being separated, can always oppose it with the opportune counter maneuver. In this case it must be held to be best to envelop the ship, attacking with the two separated sections from both sides at once (Fig. 3).
This form of attack, by sections on both sides; is specially advantageous when the ship, having made out the torpedo-boats when sufficiently far away, may have time to maneuver. The two most convenient maneuvers for the ship are the following: either to head directly toward one of the torpedo-boats (Fig. 4), probably the first one discovered, and turn toward it in a way to keep it always directly ahead (the torpedo-boat remains continually in the ship's sector of minimum offense), or to turn with the helm hard over (Fig. 5) so as to try to leave the torpedo-boats astern,
13 Considering a ship that has an extreme forward sector of 300 covered by three anti-torpedo boat guns, and the sector of 60° forward of the beam covered by 3 + 3 = 6 guns (rapidity of night firing six shots a minute) and admitting that in the attack on opposite parallel courses the torpedo-boats (see Table 1) may be during the first four salvos in the aforesaid sector of 30°, and in the attack on a normal course they may always be in the said sector of 60°, we see that if the number of the boats attacking simultaneously is three, each one might be exposed to the fire of a single piece in the first phase of the attack on the parallel course, and to that of two pieces in the second phase; while in the attack on a normal course each will always remain exposed to the fire of two pieces.
Therefore the total number of shots fired at each torpedo-boat in the available minute of time will be:
For the attack on the parallel courses = 12 (time of firing 71 seconds).
For the attack on the normal courses = 18 (time of firing 8o seconds).
Admitting that in the firing one may expect 10 per cent of effective hits, the probability of disabling each torpedo-boat with one hit is:
For the parallel courses, P = 1 — (1 — 0.1)12 = 1 — 0.28 = 0.72
For the normal courses, P = 1 — (1 — 0.1)18 = 1 — 0.15 = 0.85
Hence the probability of arriving at launching position with an angle of impact of 90° and without being disabled is represented by 0.72 for the parallel courses and by 0.85 for the normal courses; so it appears that the difference is not so great as at first might have been supposed. (Author's note.)
that is to say, in a disadvantageous position for launching. Considering that the ship occupies 80 seconds in completing the first 90° of the turn, and that the distance of good and clear discovery cannot, for the ship, exceed 1500 meters, Fig. 5 indicates how, with an opportune maneuver, and with sufficient intervals between the different torpedo-boats, a part of the units of each group could always be in good position for launching. This, however, is one of the critical cases that requires the greatest amount of
{figure}
FIG. 4.
{figure}
FIG. 5.
co-operation among the various units. Thus, for example, if the ship should keep her head toward the first torpedo-boat discovered, the effort of that boat should be to keep the attention of the ship directed toward itself, so as to allow its followers to launch, and afterward it might itself attempt to launch when the attention of the ship is perhaps attracted toward the others.
This torpedo-boat, if not destroyed, will finally find itself ahead of the ship and at a somewhat short distance, but it would never do to attempt to turn and run, since to turn through 90° would employ not less than 35 seconds, a time more than sufficient to allow the ship to run it down, while by steering to cut across the ship's circle of evolution, it will pass also at a short distance, but with the maximum relative speed, and there will remain for it the hope of being able perhaps to launch a torpedo, although under bad conditions; which will always be better than having the certainty of running the same chance of destruction without attempting any launching at all.
Passing to consider the other extreme case, that is, when the attack is made on a course normal to that of the ship, we note that,
{figure}
FIG. 6.
{figure}
FIG. 7.
evidently, the attacks of the two sections that compose the squadron, if executed on the same side and in the same sector of the ship, must necessarily be successive, as is shown in Fig. 6; or it could be developed in different sectors as shown in Fig. 7.
From what I have said it seems to me that we may deduce the principle that, in order to obtain the best conditions, it would be necessary to attack by sections, and to assume formations developed ahead of the ship and along a line nearly normal to the line joining the ship and the group leader; understanding by formation, not a rigid formation, but a simple line of dispersion.
In the form of attack determined by such standards, the various units of the section would be almost in line abreast when the courses of approach are normal to that of the ship, while in the case of opposite parallel courses they would be almost in column of vessels, a little inclined toward the ship, and each torpedo-boat would keep the one immediately ahead a little on the bow.
In this manner, having the launching tubes trained at different angles, a series of successive simultaneous launchings could be executed, with angles of impact slightly different for the different units; the first series, for example, when the group leader is about
{figure}
FIG. 8.
1000 meters distant from the ship (launching distance=700 meters) and the others at shorter distances. And if the boats were provided with torpedoes having angle gyroscopes, each torpedo-boat, in the attack on a course normal to that of the ship, could execute three launchings at once from its three tubes, with which (see Fig. 8) it could constitute for the ship a danger zone about 100 meters in extent, and would not be obliged to change course14 in order to launch, which would be a most important advantage.
The form of attack indicated would always leave clear water ahead of each torpedo-boat, would render it possible for all to see the target continually and would permit a good dispersion in order to diminish the probability of all being discovered at the same time and the probability of effective gun-fire; finally it would present the great advantage of being able to launch simultaneously or at minimum intervals.
14 The torpdeo-boat in changing course through 90° must remain exposed to fire for 35 seconds more, during which the relative movement is minimum. (Author's note.)
It seems to me that the ideas set forth may represent how much of what is desirable might be accomplished under the best conditions, which does not exclude the fact that in practice we must deny ourselves some things, and often be content with what is possible, without thinking too much of what is desirable.
Therefore, the formations mentioned represent an ideal limit toward which we should tend; and this does not imply the necessity of always having a similar disposition also before the phase of the attack, although it is certainly logical in every phase of the maneuver to occupy oneself principally with the dispositions that may be best for the attack, which is the supreme object, rather than with those that may facilitate the previous maneuvers.
In the ideal attack, torpedo-boats, in hurling themselves against ships, should be so arranged as to approximate the nearest possible to the said formations, and, therefore, during the search they will be the better prepared for the attack the simpler and quicker the passage to the formation for attack can be made.
Hence it will be seen how important may also be a good initial disposition of the units of each section during the period of search, having in view good preparation for the attack; since, if facility of maneuvering in the phases preceding the attack is desirable, on the other hand, all those conditions that facilitate good dispositions for the attack are to be held as much more important.
Evidently, during the search, the various scouting groups will not traverse the sea at hazard, but will follow pre-established routes based upon the data of the problem, and such as to insure complete co-ordination of the partial scouting services of the single groups, and, therefore, the maximum probability of success.
In this phase, what is chiefly important is the assurance of reciprocal contact, the prompt transmission of news among the various units of the group, and the fact of not being found in relative positions too different with respect to the ship, in case it is discovered, and in respect to that position forward of the beam which it will always be necessary to gain. Being found in positions considerably less advanced with respect to the leader of the column, from which may spring the necessity of increasing the speed in order to close up, is very disadvantageous. Therefore, it seems to me that the order most suitable for this phase should not present facilities for excessive elongation, as in general is realized at night in the simple column of vessels.
The advisability of diminishing the length of the formation specially exists in the case of modern torpedo-boats with greater tonnage than those of the past.
All these considerations lead me to find in the lines abreast by sections, such as are shown in Fig. 9, the order that under certain conditions might lend itself sufficiently well to this phase, especially in view of the attainment of the ideal conditions for attack.
The triangular formation, in the case of a group composed of six units, is similar to the one advocated by Labres, and to the one which according to Chandler is used in Germany.
{figure}
FIG. 9
Formations abreast are the ones that present greater safety in navigation at night with lights obscured, especially when, as is the case in every search, there are pre-established courses to be followed; they permit of the better exercise of vigilance in the sense of the direction followed; maintaining them is not more laborious than maintaining columns of vessels, nor is it as difficult as it may seem to those unaccustomed to their use. In columns of vessels, on the other hand, with high speed, it is easy to lose contact or vice versa, to have dangerous close approaches, especially in running with lights obscured, that is to say, when a distance of about a hundred meters suffices to cause one to lose sight of the stern of the boat immediately preceding, while in lines abreast or oblique it is much more difficult to lose sight of the dark lines of the entire length of the adjacent torpedo-boats. All recall what intense attention and what continual variations of speed the formations in columns of vessels require, which, therefore, soon bring about great fatigue.
In formations abreast the squadron is presented well subdivided into two tactical groups, ready to be separated for the opportune maneuvers.
Thus, for example, it is evident that in case of encountering the ships on opposite courses, the arrangement abreast would lend itself very well to being so disposed as to attack from both sides (see Fig. 3). And it would lend itself equally well in the case of meeting with courses normal to that of the ship as is seen from Figs. 6 and 7. If the ship, even in this case, should maneuver so as to keep its bow toward one of the units sighted, there would result therefrom a situation similar to that already considered, and for which I have already pointed out the opportune counter maneuver. If instead it should change course with the helm hard
{figure}
FIG. 10.
over, given the distance of sighting the boats, it might easily succeed in turning its stern in good time to the torpedo-boats, which in line abreast would be in an advantageous position for subdividing and spurting ahead in order to attack from both sides (see Fig. 10), or for maneuvering so as to try the attack from one side only.
I have principally considered the typical cases of attacks from positions included within the sectors forward of the ship's beam, in which the advantage of the lines of dispersion abreast for attack appears evident. I have scarcely mentioned attacks from positions abaft the ship's beam since, save in special cases they must be held to be attacks of expediency, and for them there might be repeated considerations similar to those set forth.
It is not contended that the line abreast may not present inconveniences; but in this connection it must be remembered that while the passage from line abreast to column of vessels can be executed with great ease and without loss of time, certainly the same cannot be said of the contrary maneuver.
After all this, without stopping to make a minute analysis of all the possible cases of relative positions in meeting—which would be lengthy—it seems to me that, solely on the basis of what I have synthetically pointed out, it is possible to believe that, on the whole, the line abreast is the formation that can more easily conduce to the best conditions for attack, which, I repeat, does not exclude passing easily and without inconvenience from line to column.
I have said that the formations must not be understood as the means of arranging and conducting the various units before the attack, in order afterwards to hurl them against the enemy from determined positions; therefore, the various units in attacking, while not being kept strictly bound to the formation, but only by the principle of being spread out in scattered order sufficiently ahead of the ship, must keep well in mind that good success depends upon the maximum co-operation, and that, therefore, not only must they not interfere with each other, but they must also be able to discern how they may aid each other as much by seeking to keep the attention of the ship continually directed toward the same partial objective, as by seeking to keep themselves on the same line so as to arrive at the same time at the launching position, etc.
Naturally, attacks conducted according to similar standards would be regulated by special practical rules which I do not think it necessary to analyze, having proposed to limit myself to considerations of a general nature concerning the fundamental principles which determine the best conditions. My intention is more to bring out the object of the maneuver than the maneuver itself.
In practice the adoption of these criteria would perhaps, in its widest sense, signify the dispersion of the various units which had remained grouped together up to the moment of sighting the ship, with the maximum liberty of maneuver, and so as to hurl themselves against the enemy from different directions, seeking to arrive at launching position at one and the same time, following only the indispensable rules in order to avoid interfering with each other when turning to withdraw; thus for example: no torpedo-boat must ever launch before the next nearer to or farther from the beam of the ship, according as the attack is developed in the bow or the stern sector profitable for launching.
It seems to me that these forms of attack may represent the rational theoretical solution which sufficiently reconciles all the varied and complex exigencies of the problem; forms that in certain cases might be practically realizable under conditions of darkness not excessive and against a squadron or a numerous convoy.
The results of numerous and accurate experiments, properly organized, might certainly prove to be much more interesting and convincing than all the considerations of a theoretical character that I have set forth.
In any event, however, whatever may be the judgment or the result of experience, the tendency to give so much value to certain too compact or excessively long orders of attack, which only permit successive launchings by the various units and at intervals not minimum, on the principle of sacrificing everything to the ideal angles of impact of 90°, seems to me illogical, and, above all, I do not understand the unwillingness sufficiently to recognize the value of the scattered order, and the taking into consideration only those forms that, while they may perhaps present possibilities of some advantages in the worst cases, certainly sacrifice the good conditions of those best cases that it must always be the intention to realize.