NAVAL ORDNANCE AND GUNNERY
VOLUME 2, FIRE CONTROL

CHAPTER 18
SPOTTING
HOME    INDEX
Chapter 18 Spotting
A. Laws of probability in their effect on gunfire and on spotting
B. The Spotter
C. Methods of spotting
                   A. Laws of Probability In Their Effect On Gunfire and on Spotting

18A1. Requirement

As the study of exterior ballistics has indicated, not all the factors which affect the flight of a projectile can be precisely evaluated in advance of firing. Even with the best fire control equipment available, experienced gun crews, and efficient fire control personnel, the opening shots may not hit the target. It is therefore necessary to apply corrections (or spots) to the initial firing data to bring the shots on the target. The corrections are applied to gun-laying data for subsequent rounds fired. This technique is called spotting.

18A2. Definitions

The following definitions relate to gunfire and to terms used in connection with spotting.

Salvo. A salvo consists of one or more projectiles fired simultaneously by the same battery at the same target.

Slow fire. Slow fire is that type of fire in which the fire is deliberately delayed to allow for the application of spots or to conserve ammunition.

Rapid fire. Rapid fire is that type of fire which is not delayed for purposes of applying corrections.

Continuous fire. Continuous fire is the firing of each gun without regard for the readiness of other guns in the battery. No real salvos exist in this type of fire, but a spotter may treat those shots which fall during an interval of a few seconds as a salvo for his purposes.

The MPI. The mean point of impact (MPI) is the geometric center of the points of impact of the various shots of a salvo, excluding wild shots.

Wild shot. A wild shot is a shot with an abnormally large dispersion in range, or deflection, or both.

In general the dispersion of a wild shot is too great to have been caused by any of the accidental errors mentioned in article 18A3, and is considered as due to a mistake rather than an error.

For example, a powder bag loaded with its ignition pad forward is a mistake and will probably cause excessive dispersion. Again, in pointer control, if a sight setter were to set a range of 15,000 yards instead of the correct range, 10,500 yards, the results would be considered a wild shot. In director control, an error in the train parallax correction or the roller-path inclination setting, both of which are applied at the individual turret, may have a similar result, as will become clear when these corrections are explained in chapters 19 and 21.

Pattern. The pattern of a salvo in range is the distance measured parallel to the line of fire between the shortest shot of the salvo and the longest shot, excluding wild shots. In deflection it is the distance measured at right angles to the line of fire between the shot falling or bursting farthest right and the shot falling or bursting farthest left, excluding wild shots.

Dispersion. The dispersion of a shot is the distance of the point of impact of that shot from the MPI of the salvo.

Dispersion in range is measured parallel to the line of fire, and in deflection at right angles to the line of fire in a horizontal plane.

Dispersion in range is positive when the shot falls beyond the MPI. Dispersion in deflection is positive when the shot falls to the right of the MPI.

The algebraic sum of the dispersions in range (or deflection) of the several shots of a salvo must equal zero. (See definition of MPI.)

Apparent mean dispersion. The apparent mean dispersion of a salvo in range (or deflection) is the arithmetical average of the dispersion in range (or deflection) of the several shots of the salvo, excluding wild shots.

True mean dispersion. The true mean dispersion is the arithmetical mean of the dispersions in range (or deflection) of an infinite number of shots, all assumed to have been fired under conditions as nearly the same as possible and excluding wild shots. (See art. 18A5.)

Hitting space. The hitting space (in range) for a target is the distance behind the target, measured parallel to the line of fire, that a shot striking the top of the target will strike the horizontal plane through the base of the target.

It includes the projection of the target’s vertical height upon the plane of the water and the target’s horizontal dimension in the line of fire (or depth). It may also include a distance in front of the target within which impacts are likely to produce under. water or ricochet hits on the target.

The hitting space in deflection is the width of the target.

The term, hitting space, as usually used, refers to the hitting space in range.

Danger space. The danger space for a material target is the distance in front of the target, measured parallel to the line of fire, that the target could be moved toward the firing point, so that a shot striking the base of the target in its original position would strike the top of the target in its new position.

Straddle. A straddle is obtained from a salvo in range (or deflection) when, excluding wild shots, a portion of the shots of that salvo fall or detonate short and others shots of the salvo beyond the target (right and left, respectively, for deflection).

Error of the MPI. The error of the mean point of impact is the distance of the MPI from the target or other reference point such as the center of the hitting space, measured parallel to the line of fire for range and at right angles to the line of fire for deflection.

18A3. Accidental errors causing dispersion

The problem of spotting is complicated by dispersion. If a battery of guns is fired at the same instant with the same settings in range and deflection, the projectiles will not all land at the same point.

If the battery of guns were stationary and rigidly fixed in elevation and train, variations in range and deflection would be caused by:

(1) differences in weight and temperature among individual powder charges;

(2) differences in projectile weights;

(3) variations in angles of projection-the axes of projectiles diverging, in varying amounts, from the continuation of the bore axis as they leave the guns;

(4) differences in projectile seating, causing variations in density of loading and initial velocity;

(5) differences in erosion among the several guns, with corrections not precisely made;

(6) differences in droop among similar guns, and unlike variations in droop with temperature changes;

and

(7) variations in amount the gun mount will yield and irregularity in action of recoil mechanisms.

These are sufficient to justify acceptance of the fact that, even under ideal conditions, dispersion in the points of fall of projectiles from several guns, or in several shots from the same gun, may be expected.

If this battery of guns be mounted aboard ship, and each be individually positioned by a pointer and trainer, the rolling and pitching, and the yawing, of the ship itself will further cause dispersion.

The motion of the ship may cause the pointers and trainers of the several guns to misalign their sights on the target when the guns are fired.

The same effect may result from failure to fire exactly simultaneously, causing different guns to fire at slightly different points in the roll, and thus at different velocities of roll. Director-controlled gunfire, although not subject to the same characteristic errors as pointer fire, is subject to its own set of characteristic errors. These are discussed in some detail in articles l8Al1 through 18A15.

The errors mentioned above as characteristic of pointer fire, and many of those which are characteristic of director fire, come under the general classification of accidental errors. They are revealed by analyses of firings, and their effects are governed by laws of probability, as will be explained.

18A4. Determining the location of the MPI

The definitions in the above paragraph will now be illustrated by an example, representing a salvo from ten 5”/38 caliber guns fired at range 8,500 yards against a target 40 feet high, 600 feet long, and with a beam of 90 feet. The target is situated with its length at 90 degrees to the line of fire.from ten 5”/38 caliber guns fired at range 8,500 yards against a target 40 feet high, 600 feet long, and with a beam of 90 feet. The target is situated with its length at 90 degrees to the line of fire.
The danger space in range from column 7 of the range table is 40/20 X 25 = 50 yards. Since the “depth” is 90 feet, or 30 yards, the actual danger space is thus 50 + 30 = 80 yards. As discussed in article 17B11, for most battle ranges the value of hitting space is the same as that for danger space. Hence, in this case the hitting space can be considered 80 yards. Figure 18A1 represents the plan view of the target shown in terms of hitting space, an area 80 yards in range and 200 yards in deflection. The center of the hitting space is at C. (The fact that the target’s deck is not rectangular in shape, causing a narrowing of the hitting space at the extremities, will be ignored.) The points of impact of the several shots are indicated by numbered dots and may be identified by referring to the adjacent table, which gives the location of each impact from the reference axes, in this case taken as intersecting at the center of the target’s waterline.

The location of the MPI is determined by measuring the distance, in range and deflection, of the point of impact of each shot from convenient coordinate axes, and finding the mean of these distances. It is convenient to refer impacts to axes intersecting at the center of the target’s waterline, as was done in the table. The mean of these distances is 80 yards over and 20 yards right, which locates the MPI 80 yards beyond and 20 yards right of the center of the waterline.

If we assume that the error of the MPI is its distance from the center of the hitting space, this is seen to be 40 yards over and 20 yards right.

Inspection of figure l8Al, or of the table, shows that the pattern of this salvo is 260 yards in range and 130 yards in deflection. (None of these figures, nor any figures in this chapter, are to be considered as typical.)

18A5. Determining apparent mean dispersion

In determining the apparent mean dispersion it must be borne in mind that it is the arithmetical mean of the dispersions of the several shots, without regard to sign. The position of the MPI having been plotted, the several points of impact are now referred to the MPI to determine the individual dispersions and the apparent mean dispersion.

The apparent mean dispersion, based on 10 shots, is thus 62 yards in range and 32 yards in deflection. Being based on such a limited number of observations, it is not a true measure of the accuracy of fire. The true measure of the accuracy is the mean dispersion of an infinite number of shots, all fired under the same conditions as those under consideration, and is called the true mean dispersion. Although it is obviously impossible to measure the value of the true mean dispersion experimentally. a theoretical value is given by the relation.
The true mean dispersion, in the above case, is found by multiplying the apparent mean dispersion by 1.054. Hence the true mean dispersion in range, denoted by Dr, is 62 X 1.054 = 65 yards, and in deflection, denoted by Dd, is 32 X 1.054 = 34 yards.

18A6. Law of probability applied to dispersion

After determination of the true mean dispersion of shots fired by a battery of guns, the next step is the investigation of the manner in which the accidental errors, which resulted in this dispersion, will affect the chances of hitting a target, and how the control of a battery of these guns will be influenced by a knowledge of these errors.

In this study use is made of the laws of probability, which deal with the prediction of future occurrences on the basis of information gained from past occurrences. The laws of probability are the basis of the science of statistics, and may apply to almost any event or occurrence.
The law which pertains to symmetrically distributed accidental errors is known as the normal probability law. The accidental errors of gunfire follow a fairly symmetrical distribution, and this law is applied to them. It is based upon the following considerations:

1. That positive and negative errors of the same size are equally probable, and hence will occur with equal frequencies.

2. That small errors are more probable than large errors, and hence will occur with greater frequency than large errors.

3. That very large errors will not occur (or, more specifically, that very large errors will probably be due to mistakes and not be classifiable as accidental errors).

These considerations all pertain to the probable frequency of occurrence of errors of certain size and sign. Translated into a mathematical equation, the curve representing that equation will have the following characteristics (the abscissas measure the size of the errors; the ordinates, the frequency of their occurrence):

1. Since positive and negative errors of the same size occur with equal frequency, the curve must be symmetrical to the right and left of the point that denotes zero error (x=0), which is the origin.

2. Since small errors occur more frequently than large errors, the maximum ordinate must occur at the point that denotes the least error (x=0), which is the origin.

3. Since large errors rarely occur, and beyond some limit do not occur at all, the right- and left-hand branches of the curve must each approach the horizontal axis rapidly and meet it at some point.

In figure 18A2 are shown two such curves, as they apply to the accidental errors of gunfire. Curve A is based on a relatively small mean dispersion. (Actually, curve A represents a mean dispersion just two-thirds as great as is represented in curve B, both curves being drawn to the same scale.)

In the determination of dispersion for a gun the student will recall that wild shots are excluded. It is, therefore, apparent that the curve will meet the horizontal axis at the limit, on each side of the origin, beyond which shots are excluded as wild shots.
18A7. Determining probability of hitting within a certain error

The true mean dispersion and the probability curve having been determined, it is possible to determine the probability of hitting within a certain distance of the origin, that is, within a certain error, plus or minus.

A table has been prepared in which this limit, denoted by a, is expressed as a percentage of the true mean dispersion D. The entering argument is a/D, the ratio of the particular error (±a) under consideration to the true mean dispersion D. The quantity P found from the table is the probability that an error not greater than ±a will occur, i. e., that the error will lie within the limits +a to -a. (The value of D’, from which D results, must be obtained from target practices on which the points of impact of all shots in each salvo are measured or photographed.)

It must be understood that this table may be used for the probability of occurrences in general, which follow the normal probability law. In the general case, a represents the error while D represents the index of precision as determined by past observations. (The index of precision refers to the degree of accuracy which has been observed in the past. In the case of gunfire the index of precision is the value of the true mean dispersion.)
From an inspection of the table two facts are readily apparent:

1. For a probability of 0.500, the error will lie between ±0.846D. Or, the probability of hitting between limits separated by l.692D is 50 percent.

2. The probability of hitting within limits of error equal ±4D is 0.999. It follows that a dispersion greater than 4D will fall without the curve. In connection with problems in this course the student should consider a dispersion greater than 4D as due to a wild shot.

18A8. Probability of hitting when MPI is at center of hitting space

The case pictured in figure 18A1 will now be employed to illustrate the use of the probability table, in the case where the MPI is located at the center of the hitting space. If the entire salvo be shifted in range and deflection to bring the MPI to coincide with C, it will be seen that shots 4, 5, 6, and 7 are material hits, i. e., hits within the material target. The table will be employed to determine the percentage of an infinite number of shots which may be expected to hit this target.

The probability of hitting in range, Pr, is determined by entering the table with a value of ±a/Dr equal to ±40/65=0.62. (The hitting space being 80 yards in range, ±40 yards encompasses the hitting space. The true mean dispersion in range, Dr, from article 18A5, is 65 yards. The ratio will be carried to two decimal places.) The value of Pr is 0.379, to three decimal places.

In deflection the hitting space is 200 yards and the true mean dispersion in deflection, Dd, is 34 yards; thus ±a=l00 yards. ±a/Dd=l00/34= 2.94, and the probability of hitting in deflection, Pd, is 0.981.

It is an axiom of the laws of probability that if an event is independently controlled by more than one set of conditions, the probability of its occurrence under all of these sets of conditions is equal to the product of the several probabilities of its occurrence under the several sets of conditions. Thus, in a pack of playing cards, the probability of drawing is 1/13, and the probability of drawing a spade is ¼. The probability of drawing the king of spades is then the product of these two probabilities, or 1/52. In this case, the probability of hitting the target depends upon the probability of hitting in range and the probability of hitting in deflection. Therefore, the probability of hitting, P, is the product of these two probabilities, or 0.379X0.981=0.372. Then for a salvo of 10 shots about 4 hits should be obtained. (The fact that more or less than this number may be obtained on any one salvo is of no moment at this time, because these laws are based upon a great number of occurrences and do not directly apply to any one.)

Figure 18A1 also shows that the densest portion of the salvo is centered upon and relatively close to the MPI. In article 18A7 it was shown that one-half of the shots, or five in this case, should fall within ±O.846D of the MPI, or within 0.846X65=55 yards in range and 0.846X34=29 yards in deflection. In figure 18A1 it is seen that 5 shots are within this limit in range; 4 shots are within this limit in deflection, with 2 more just 1 yard outside the limit.

The above paragraph shows the importance of maintaining the MPI at the center of the hitting space. As the MPI moves away from the center, it carries with it the area of greatest density of impacts, and the probability of hitting falls off very rapidly. For comparison, the probability of hitting this same target will be determined for the case where the MPI lies just one-half pattern, or 130 yards, beyond the center of the hitting space.will be carried to two decimal places.) The value of Pr is 0.379, to three decimal places.

In deflection the hitting space is 200 yards and the true mean dispersion in deflection, Dd, is 34 yards; thus ±a=l00 yards. ±a/Dd=l00/34= 2.94, and the probability of hitting in deflection, Pd, is 0.981.

It is an axiom of the laws of probability that if an event is independently controlled by more than one set of conditions, the probability of its occurrence under all of these sets of conditions is equal to the product of the several probabilities of its occurrence under the several sets of conditions. Thus, in a pack of playing cards, the probability of drawing is 1/13, and the probability of drawing a spade is ¼. The probability of drawing the king of spades is then the product of these two probabilities, or 1/52. In this case, the probability of hitting the target depends upon the probability of hitting in range and the probability of hitting in deflection. Therefore, the probability of hitting, P, is the product of these two probabilities, or 0.379X0.981=0.372. Then for a salvo of 10 shots about 4 hits should be obtained. (The fact that more or less than this number may be obtained on any one salvo is of no moment at this time, because these laws are based upon a great number of occurrences and do not directly apply to any one.)
Figure 18A1 also shows that the densest portion of the salvo is centered upon and relatively close to the MPI. In article 18A7 it was shown that one-half of the shots, or five in this case, should fall within ±O.846D of the MPI, or within 0.846X65=55 yards in range and 0.846X34=29 yards in deflection. In figure 18A1 it is seen that 5 shots are within this limit in range; 4 shots are within this limit in deflection, with 2 more just 1 yard outside the limit.

The above paragraph shows the importance of maintaining the MPI at the center of the hitting space. As the MPI moves away from the center, it carries with it the area of greatest density of impacts, and the probability of hitting falls off very rapidly. For comparison, the probability of hitting this same target will be determined for the case where the MPI lies just one-half pattern, or 130 yards, beyond the center of the hitting space.

Referring to figure 18A3, the center of the hitting space, which is represented by the shaded area, is at C. The MPI is 130 yards beyond and in line with C. In order to find the probability of hitting within the hitting space, in range, it is necessary first to determine the probability of hitting the total area ABCD, using ±a equal to 170 yards, which is seen to be equal to the error of the MPI plus one-half the hitting space, S’. This probability is denoted by the symbol Pr1. Next the probability of hitting within the area EFGH is determined using ±a equal to 90 yards, which is equal to the error of the MPI  minus one-half the hitting space, S’. This probability is denoted by Pr2. Now, if Pr2 is subtracted from Pr1 the result will be the probability of hitting within area ABCD minus EFGH, or within the two areas ABFE and HGCD. These two areas are equal, so that the probability of hitting within the single area HGCD is one-half this result, or

Pr= 1/2 (Pr1 - Pr2).

This problem will now be solved as explained above, first for Pr1, then for Pr2, and finally for one-half their difference. True mean dispersion in range, Dr, as before, equals 65 yards.

For Pr1, ±a/Dr=170/65=2.62; and Pr1=O.963.

For Pr2, ±a/Dr=90/65= 1.38; and Pr2=0.728.

Then Pr=½(0.963_0.728)=0.l 18.

But P=PrXPd. The value of Pd will not differ from the value found above, since the error of the MPI in deflection is zero.

Hence P=0.1l8x0.981=0.l16.
18A9. Spotting MPI to center of hitting space

The two cases illustrated in figures 18A1 and 18A3 will be analyzed to see what information may be obtained of interest to a spotter on shipboard. In the first case, with the MPI at the center of the hitting space, the laws of probability show that about four hits (i. e., 0.379X10) in range should be obtained in a 10-shot salvo. Therefore, about six shots will miss the target; and since the first consideration of the normal law states that positive and negative errors will occur with equal frequency, it follows that about three [(10-3.79)÷2] of these misses will be “shorts.”

Probably the most common fault of spotters is the feeling that shorts are wasted, since it is so readily apparent that they have not hit the target, and that ensuing salvos must be spotted up until no shorts are seen. Consideration of relative sizes of pattern and hitting space will show, however, that if the MPI is to be anywhere in the hitting space, some shots must be observed short of the material target, even with the symmetrical shot distribution assumed in this discussion.

Again, remembering that the greatest density of impacts is grouped within a relatively short distance of the MPI (±0.846D), it is obvious that if no shorts are seen the MPI must be well beyond the center of the hitting space and, at any likely range, well outside the hitting space. Thus the greater part of the salvo, and particularly the densest portion of the salvo, is actually being wasted. Also, as the range increases, the hitting space decreases, the probability of hitting decreases, and the number of shorts which should be seen increases. (Of course, the converse is true; the range is finally reducing to a point where the maximum ordinate of the trajectory is less than the vertical height of the target. in this case the danger space equals the range; the range at which this condition occurs is called the danger range.)

In the second case, the error of the MPI in range is one-half the pattern size. (It should be mentioned that this pattern size does not represent the pattern size of an infinite number of shots, since it was based on the points of impact of only ten shots.) The MPI is 90 yards beyond the material target and the densest portion of the salvo, included within the distance ±0.846Dr, is entirely beyond the target. The probability of hitting in range is reduced from 0.379 (for error of the MPI equal to zero) to 0.118, or by more than two-thirds. These values clearly demonstrate the necessity of spotting the MPI to the center of the hitting space, which in turn requires a considerable number of observed shorts.

An officer detailed to duties which include spotting should know the probability of hitting at various ranges up to the extreme range of his battery against the various types of target which might be encountered. This information will tell him the number of shorts which should be expected or striven for under all possible conditions.

18A10. Accidental errors causing shift of MPI

Another phase of the problem of interest to the spotter is a knowledge of the amount which the MPI’s of successive salvos may be expected to vary in location. Since the MPI is the center of the points of impact, which are the results of unpredictable accidental errors, it is logical to suppose that no two salvos will be exactly alike. Or, since successive shots from the same gun, fired under the same conditions, may be expected to vary, the MPI’s of successive salvos will also vary.

The MPI of 100 shots (although fired as 10 salvos of 10 shots each) is the geometrical center of the points of impact of these 100 shots. The MPI’s of the individual salvos will differ from this aggregate MPI by varying amounts which, however, are still in accordance with the normal probability law. The distance each MPI is located from the aggregate MPI is the dispersion (in range and deflection) of that MPI. The arithmetical mean of the dispersions of several MPI’s (with respect to the aggregate MPI) is called the mean dispersion of the MPI. This value may be obtained by the equation
While the spotter cannot estimate the MPI of a salvo with any such degree of accuracy, he should expect the number of shorts seen on successive salvos to vary as the MPI’s of these salvos move about the center of the hitting space. He should therefore not be too quick to spot when only one or two salvos seem to wander off the target during a string that is, in general, satisfactory.

I8A11. Control errors

The only accidental errors so far considered have been those which affect individual guns of a battery. These are known as gun errors and should be distinguished from another class of error which affects the battery as a whole. These are known as control errors and include those made in range-keeping, transmitting of data to the guns, and, in director fire, director pointing errors. They are not reflected in increased pattern sizes (i. e., increased dispersion among the guns) but in increased dispersion of the MPI’s themselves.

Before the first salvo is fired at a target, as many effects likely to cause error of the MPI as practical are accounted for. If this salvo has an error of MPI, it is the result of control inaccuracies. There are four general control inaccuracies which may cause error of the MPI:

1. Rangekeeper set up with incorrect range, courses, and speeds.

2. Ballistic corrections based on conditions not existing at the time of firing (initial velocity, wind, air density, and others).

3. Battery not properly aligned with director.

4. Indeterminate errors (Class B and personnel errors).

18A12. Incorrect rangekeeper set-up

The present range to the target is valid only to the extent that its measurement is accurate. An error in this basic range measurement is directly registered as an error in MPI.

Measurements of own-ship course and speed usually are reasonably accurate, but any inaccuracies result in an error of MPI.

The determination of the target’s course and speed is made directly from the spotter’s estimate of target angle and speed, from the radar plot in CIC, or by rate controlling. Correct values of these two variables are most difficult to determine; they are the chief cause of an incorrect rangekeeper set-up, and thus the chief source of M PI error.

18A13. Inaccurate ballistic corrections

The computer determines corrections necessary to compensate for variations from standard conditions. If the determination of these corrections is based on incorrect values of ballistic wind, initial velocity, air density, and others, the total ballistic correction will be in error and will result in a corresponding error in MPI.

18A14. Improper battery alignment

The alignment referred to here is not intended to mean improper alignment between the guns of a battery. Such misalignment results in greater dispersion and larger pattern sizes, but does not materially affect the error of the MPI of a salvo. The improper alignment referred to here means the misalignment which may exist between the controlling director and the battery as a whole. This type of error is generally caused by failure to director-check the battery and all directors. Thus, a battery which is aligned with one director is not necessarily aligned with another which may be in control. Frequent director checks can assist in elimination of this cause of error in MPI.

18A15. Indeterminate errors

One of the two classes of indeterminate errors results from the fact that some of the computations by fire control instruments are only approximations of the true solutions. They are acceptable because they are susceptible of easier mechanization. These approximations result in Class B errors, which are small for normal ranges and therefore cause minimum error in MPI. However, at extremely short or long ranges the errors may become large, depending upon the instrument concerned, and thus may seriously affect the MPI. When Class B errors are known to be large, they are not admissible as accidental, and steps must be taken to make correction for them.

The other class of indeterminate errors is assignable to control personnel. An example would be the director pointer or trainer being off the point of aim when the salvo is fired. Only training and experience can prevent the occurrence or reduce the magnitude of such mischance. Small errors of this type merely cause a slight shift or dispersion of MPI and should not be corrected by the spotter. The director operator should inform the spotter of large discrepancies in the point of aim, in order that the spotter may distinguish this error from others.

18A16. Summation of errors

During actual firing a spotter cannot analyze each fall of shot to determine the exact geometrical MPI of a salvo. Splashes last only a matter of seconds, and the firing situation is probably changing even during that short interval. The spotter must have previously analyzed all probable errors and the reasons for them. This knowledge, coupled with practical training, will enable him to make instant and intelligent decisions as to the proper correction or spot necessary to hit the target.

To summarize, the general errors attending shipboard gunfire are:

1. Accidental gun errors causing dispersion of shots. These errors are only compensated to the extent of achieving desired pattern sizes. Most of the errors are eliminated by careful design, frequent checks of battery alignment, normal upkeep of the battery, and the training of gun crews.

2. Accidental gun errors causing a shift in the MPI of successive salvos. The shift in range is usually small; a noticeable shift in deflection is observed when using radar aim.

3. Control inaccuracies causing error of MPI. It is the primary duty of the spotter to “spot” the corrections in range and deflection necessary to bring the MPI of a salvo to the desired point of impact. He should recognize the first two classes of errors in order to spot the error of MPI caused by control inaccuracies.