Armored 
Medical Research Laboratory 
Fort Knox, Kentucky 
PROJECT NO. 6 - VISION IN TANKS 
Report On 
Sub-Project No, 6-1, Determination of the Visual Requirements 
for Various Tasks in Armored Vehicles 
'SUBJECT: STUDY OF ERRORS IN RANGE ESTIMATION WITH 
\ THE UNAIDED EYE 
Project No. 6-1 
30 April 1945 ARMORED MEDICAL RESEARCH LABORATORY • 
Fort Knox, Kentucky 
Project No, 6-1 
SFVEA 741 
30 April 1945 
1, PROJECT: No, 6-1 - Determination of the Visual Requirements for Various 
Tasks in Armored Vehicles; Subject: Study of Errors in Range Estimation with the 
Unaided Eye, 
a. Authority: Letter Commanding General, Headquarters Armored Force, 
Fort Knox, Kentucky, File 400,112/6 GNOHD, dated September 2k, 1942o 
b, Purpose: To study the accuracy of range estimation without visual 
aids among tank crew trainees with special reference to (1) degree of accuracy 
obtained, (2) possibility of preselection of the more competent individuals, (3) 
Improvement with training, and (4) influence of the character of targets and terrain,, 
2* DISCUSSION! 
a« Range estimation is one of the most Important and, at the same time, 
least accurate element in tank gunnery. Previous studies of visual range estimation 
performance have dealt with the average proficiency or lack of it in groups of sol- 
diers, the burden of the investigations being concerned primarily with the magnitude 
of the mean errors, and the influence of the distance of the target on the accuracy 
of its estimate. In substance, however, the outstanding finding has been that troops 
in general do not estimate ranges accurately enough to secure hits with the first 
round of fire, and that the errors, being of the order of 20% or more, necessitate 
bracketing with several rounds, before the target is struck. It seemed desirable, 
therefore, to inquire more fully into the measures that might be adopted to improve 
this situation. Such measures might be: 
(1) The incorporation of range finders as standard equipment for 
tanks. At the present time there are none in American tanks in 
combat theatres and the current opinion is that the time required 
to remove the range finder from its stowage, mount it, and use it, 
is not frequently available in combat. New types of range finders 
are, however, under consideration for limited issue, but it is 
generally agreed that the tactical, lighting, and terrain con- 
ditions are not always suitable for their use. Furthermore, 
special training may be a complicating requisite for their effect- 
ive employment. 
(2) Calculation of distance from the angle subtended bv the target 
on a mil scale. For this to be done accurately, tne optical 
instrument must be stable, as would be the case with the tank 
telescope if the tank were stationary, but not with binoculars; 
1 and the size of the target must be known. Time is required 
for the calculation; and, more important, the precise target 
sise will not be generally known. 
(3) Intensive training of tank crewmen in estimation with the 
unaided eye* As pointed out above, studies of the performance 
men have generally indicated relatively imperfect performanc 
One may therefore raise these questions: 
(a) Can the technique of instruction be improved, with con- 
sequent reduction in the error of estimate? 
(b) Would more intensive instruction of tank crewmen produce 
better results? (A recent ACEG report. No* 225 on The 
Accuracy of Visual Estimation of Initial Hange.».concludes 
that n.,*an improvement in performance by the NCO category 
of personnel is possible by continuous training and practice 
in range estimation"„) 
(c) Should tank crewmen, or more particularly tank commanders 
and gunners, be selected for their proved or potentially 
superior capabilities in range estimation? (The antiair- 
craft and coast artillery commands devote several weeks to 
training on height and range finders, of men preselected 
for their intelligence and the excellence of their vision0) 
bo The range estimation records of a large group of men for 32 targets on 
different ranges were obtained for study from the gunnery department of the 
This analysis undertakes to evaluate from these records five aspects of the 
estimation problem: 
three 
ARTCo 
range 
(1) The general level of competence of men in training as tank 
crewmen. 
(2) The level of excellence of the 10 most capable men of the group. 
(3) The degree of improvement with training. 
(4) The relationship of general vision and of previous civilian 
experience to accuracy in judging distance. 
(5) The influence of the character of tbs target on the estimate 
of its distance. 
3o CONCLUSIONS t 
a. The mean error for all 4472 estimates was This was considerably 
higher than the median error of 20# for the reason that one-sixth of all estimates 
were in error by more than U0%, On the other hand, slightly less than one-third 
of all estimates were exceptionally good, being incorrect by less than 10^o 
2 b, Some men possess outstanding skill in estimating distances. There 
were 10 men in the present group whose median error for 32 separate range esti- 
mates was 13% or less, the best having a median errof- of only 8%0 These men had 
had only several weeks basic military training prior to this test* 
c0 The failure of the group as a whole to improve in successive exercises 
suggests that the technique and extent of instruction does not substantially in- 
crease the range estimation proficiency of trainees. 
do The best and poorest range estimators of this and of another group of 
men, specially examined, could not be distinguished on the bases of civilian occupa- 
tion, participation in outdoor sport, urban or rural residence, or vision tests0 
e0 The tendency to underestimate the distance to targets with sharp, 
well-defined contours was consistent* Examples of such targets were: vehicles, 
coniferous trees, and range limit markers. The targets most frequently over- 
estimated were partly hidden objects and those with low contrast. Neither the 
target's azimuth nor distance up to 1800 yards appeared to influence the pattern 
of error* 
4o RECOMMENDATIONS; 
a* That identification of the more competent range estimators be regarded 
as an important consideration in the selection of tank commanders and gunners. 
b« That the reasons for the apparent superiority in range estimation 
possessed by some individuals be further investigated as a basis for improving 
range estimation instruction generally. 
Co That improved devices and techniques for range estimation instruction 
be sought. 
d. That the influence of the character of the target be pointed out in 
imparting range estimation instruction to men. In particular, the tendency to 
underestimate the distance to targets of sharp outline should be appreciated. This 
may be of importance when fire is directed at a terrain feature which is selected 
because of its inferred proximity to an enemy strongpoint. 
e0 That investigation and development of built-in tank range finder be 
continuedo 
NOTE: The recommendations as set forth in this project have been concurred in by 
Col„ Fred Wo Makinney, Chief of Staff, Armored Center. 
Submitted by: 
Arthur Freedman, Captain, MC 
k IncIsc 
APPROVED( 
#1 - Appendix I 
#2 - Appendix II 
#3 - Appendix III 
#4, - Figures 1, 2 and 3* to 3p 
WILLARD MACHLE 
Colonel, Medical Corps 
Commanding 
3 APPENDIX I 
.ACCURACY OF RANGE ESTIMATION 
lo Introductory. 
a« The present study of accuracy of direct range estimation with the un- 
aided eye makes use of data obtained from troops in the gunnery section of basic 
training at the Araored Replacement Training Center. There were 32 targets in all, 
and on these a. total of 4A72 separate estimates were made. The 32 targets were 
on 3 ranges, each range being employed in a separate exercise. 
Range §1 - Within the barracks area of the Post—10 targets, approxi- 
mately 175 men participatingo 
Range #2 - On Steele*s Firing Range #4—in open country—10 targets— 
approximately the same 175 men© 
Range #3 - On Steele*s Firing Range #3—in open country—12 targets— 
approximately 95 of the original 175 men© 
b0 Ranges are routinely estimated by the men in groups of 10 to 20, as 
part of their gunnery instruction course, under the supervision of an officer or 
NCO. For each range, the individual is given a blank form on which he is instructed 
to write a one-word description of the target, and his estimate of the distance in 
yards after the target is pointed out to him by the instructor. He is permitted to 
identify the target with binoculars, but the distance estimate is made with the un- 
aided eye© At the termination of the exercise the men are told the correct range 
to each target and they then compute their scores© It is of interest that although 
the program is intended for instruction, it is generally conducted more in the marine: 
of an examination, with no clues given by the instructors until the estimates have 
been made© 
20 Performance of the Group as a Whole. 
a, Figure 1 gives the overall accuracy of all the estimates. Fewer than 
a third were within 10? of the true range; approximately one-half were within 2056 
of the correct distance; and one-sixth were more than k0% in error. The mean error 
was 2606£, and the median error was 2056. Since this is the only range estimation 
practice which many of these men receive before going overseas as replacements, it 
would seem that they had not developed their skill to a high point of excellence. 
b. Of the entire group,86 of the men estimated the distances to all 32 
targets. The distribution of the median estimation errors of these men is shown 
in Figure 2, For this group the mean error without regard to sign was 260 056 with 
a standard deviation of f 12 05% and a probable error of the mean of j- ©9156© 
3« Performance of Superior Individuals, 
Many of the scores in the group of 86 men indicated highly superior ability 
Inclo #1 
1 In range estimation. The best individual estimated all 32 targets with a median 
error of &% and a mean of 12D3%. The median error for the best 10 men of this 
group was 13% or less* Table 1 gives the individual scores of these 10 men, to- 
gether with a partial tabulation of the magnitude of their errors on all the 
targets. These are compared in the same table with the performance of the 10 
poorest performers. 
TABLE 1 
THE PERFORMANCE OF THE 10 BEST AND 10 POOREST RANGE ESTIMATORS 
ON 32 TARGETS ’ 
Name 
Median 
* Error 
Arithmetic > 
Mean 
* Error 
No. Estimates 
Accurate 
Within 10* 
No. Estimates 
In Error 
By 50* or more 
BEST 
X o E • «J • G© 
8 
12.3 
19 
0 
2, H.J.S. 
9o5 
13 o2 
16 
1 
3. F.R.J. 
10 
lUoU 
15 
0 
i», T.F.B. 
10 
17 o 5 
12 
2 
5» w.a.d. 
10.5 
12.7 
15 
0 
6g W.AoBo 
11 
18.5 
12 
2 
7o J• L.S«' 
12 
14.2 
14 
1 
8»> W. J,K• 
12 
19.9 
14 
4 
9o R.R.H. 
12.5 
19.6 
13 
2 
10„ E.F.K. 
13 
16.6 
11 
1 
POOREST 
Xo J♦X.D» 
39o5 
101 o 9 
6 
12 
2„ A.A.S. 
34.5 
35o5 
3 
6 
3. H.E.T. 
33 
29.3 
9 
9 
4o J.A.Zo 
32 
43.0 
5 
10 
5. R.K. 
31o5 
30„0 
6 
4 
6, H.L.C 
31 
AO. 6 
3 
9 
7o q.o.b. 
29 
32.1 
7 
7 
Be D.E.M. 
28.5 
46.8 
5 
11 
9o R.G.M. 
28«5 
30.3 
6 
7 
10. k.C.Bo 
28 
29.6 
6 
5 
(FOR CdlPAR- 
ISON) 
ALL HEN 
20 
26.6 
Inclo 01 
2 40 Influence of Training, 
a, The target distances on Range #1, §2 and #3 were estimated in that 
order with intervals of a few days between exercises. The scores indicate that 
the men did not improve in their performance on the successive ranges, but 
actually tended to do more poorly and accordingly did not profit from their 
instruct!on0 
TABLE 2 
PERCENT OF ACCURATE ESTIMATES AND MEAN ERROR BY RANGES 
% of Estimates 
Accurate with- 
Mean Error 
in 10* 
of all men 
Range #1 
40.2* 
18.3* 
Range #2 
26.2% 
31.9* 
Range #3 
22.6% 
27.6* 
ALL TARGETS 
26.6* 
b. The targets and locale of Range #1 were undoubtedly more familiar 
to the men than those on Range #2 and #3# since Range #1 is in the urban part of 
the Post, and Ranges #2 and #3 are in open country. This resulted in relatively 
greater accuracy of range estimation on Range #1 (Table 3), The targets on 
Ranges §2 and #3, on the other hand, were approximately equal with regard to 
difficulty of range estimation, as ah own in Table 3, wherein it is noted that 
the number of men making better scores on one of these two ranges is approxi- 
mately equal to the number of men making better scores on the other. 
Incl, ffl 
3 TABLE 3 
COMPARATIVE DISTRIBUTION OF THE MEAN ESTIMATION ERRORS 
OF 86 MEN FOR THE 3 RANGES ON WHICH THEY WERE 
SUCCESSIVELY TESTED AT ABOUT 3-DAY INTERVALS 
NUMBER OF 
MEN 
Lower Mean Errors on Range #3 as compared with Range #2 
40 
Higher « " ■ Range #3 " " " Range #2 
45 
No difference in Mean Errors on Range #3 as compared 
1 
with Range #2 
86 
Lower Mean Errors on Range #2 as compared with Range #1 
29 
Higher » " " Range #2 " • n " Range #1 
51 
86 
Lower Mean Errors on Range #3 as compared with Range #1 
23. 
Higher n " " Range #3 H " " Range #1 
63 
86 
NOTE - Range #1 in urban part of Post 
Ranges #2 and #3 in open country on firing ranges 
Inclo #1 APPENDIX II 
CONSIDERATIONS RELATIVE TO PRESELECTION 
1« Civilian Occupation and Residence., 
The men at the Replacement Center whose range estimates formed the basis 
of this study had gone overseas by the time analysis of the data was completed* 
Consequently, it was not possible to interview or further examine them to dis- 
cover any attributes which might be responsible for the performance of the best 
and poorest Individuals, The only information available was with regard to place 
Of residence and civilian occupation, which, as indicated in Table 4, does not 
appear to offer any basis for distinguishing between the superior and poorer 
groups* 
TABLE 4 
CIVILIAN OCCUPATION AND RESIDENCE OF TEN BEST AND TEN POOREST RANGE ESTIMATORS 
Name 
Civilian Occupation 
Civilian Residence 
BEST 
E. J.Go 
Farmer 
Rural 
Wis. 
H.J*S. 
None 
n 
Mont. 
P.R.J. 
Fanner 
M 
Wifio 
T.P.B. 
Mill Worker 
Urban 
Mich. 
W.A.D. 
Laborer 
n 
S * Ct> 
W.A.B. 
Mechanic 
R 
Uichc 
J a L.S . 
Electric ian 
n 
Florid* 
W.J.K. 
Clerk 
it 
Wia0 
H.R.H. 
Clerk 
M 
s. c. 
H.FoKo 
Mechanic 
ft 
ELL. 
POOREST 
■ -r:' 
J.I.D. 
Parmer 
Rural 
Wi«. 
A.A.S, 
Farmer 
n 
Mich. 
H.E.To 
Miner 
n 
Mont. 
J.A.Z. 
No record 
R.K. 
No record 
- 
H.L.C. 
Farmer 
n 
Moo 
G.0,3. 
Clerk 
Urban 
0. 
D.E.M. 
Farmer 
Rural Mo* 
R.G.M. 
Laborer 
it 
in. 
M.C.Bc 
Mechanic 
Urban Ill0 
Inclo #2 2* Vision and Participation in Sport0 
&« Accordingly more extensive examinations were made on other similar 
classes of men from the ARTC. The ten best and the ten poorest range estimators 
of two classes were given vision tests with the Keystone Telebinocular, and the 
ten best and ten poorest of one class were questioned regarding civilian occupa- 
tions and hobbies* The eye tests included examinations for acuity, astigmatism, 
binocular fusion and heterphoria, stereopsis, and color vision* The results of 
these tests may be summarized as follows: 
Of the 20 best range estimators; 
15 men had completely satisfactory vision 
1 man was color blind 
1 man had moderately diminished visual acuity 
3 men had multiple vision defects 
Of the 20 poorest r 
estimators: 
12 men had completely satisfactory vision 
2 men were color blind 
2 men were astigmatic 
4 men had multiple vision defects 
bc In view of the similarity between these two groups, it seems that 
neither the possession of normal vision nor the existence of visual defects deter- 
mined the rang-e estimating ability of this group of men. 
c0 Of the 10 best range estimators who were questioned concerning 
civilian sports activities; 
4 men were farmers and had hunted extensively 
4 men had lived in the country but had hunted little 
2 men were city dwellers with no hunting or golf expereince 
Of the 10 poorest r 
estimators so questioned: 
5 men were farmers and had hunted extensively 
2 men had lived in the country but had hunted little 
3 men were city dwellers with no hunting or golf experience 
d. It seems unlikely from these comparisons that previous civilian exper- 
ience exerts a consistent influence on range estimating ability, although some men 
who may have been surveyors, hunters, or golfers, undoubtedly do profit by these 
activities 
6o By exclusion, the inference seems to be that Judgment of distance is & 
special skill which is either inherent or can be learned through diligent application] 
that it may not be much impaired by moderate vision defects; and that no method is 
apparent as yet for selecting in advance of actual trial, those individuals whose 
performance is likely to be superior to that of other men0 
Incl® 02 
2 APPENDIX III 
INFLUENCE OF CHARACTER OF TARGET AND TERRAIN 
1. Certain factors which may conceivably influence the accuracy of range 
estimates are thoserelated to the nature of the target and its surrounds, in- 
cluding illumination and contrast. 
a0 In this regard instructions in the Basic Field Manual, U. SD Carbine, 
Caliber o30 M, FU 23-7 state that: 
Objects seem nearer; 
(1) When the object is in bright light* 
(2) When contrast is sharp; 
(3) When the line of sight is over a uniform surface; 
(4) When the line of sight is downward; 
(5) When looking over a depression most of which is hidden; 
Objects seem more distant; 
(1) When looking over a depression most of which is visible; 
(2) When there is poor light or fog; 
(3) When only a small part of the object can be seen; 
(4) When the line of sight is upward from low ground; 
b0 An analysis of the estimates on the 32 targets confirms the principles 
outlined above, to the extent possible in a study of this limited extent„ In most 
instances the errors in estimating the distances to targets which were clearly seen 
and sharply outlined were characteristically underestimates. Conversely, some of 
the targets which were overestimated presented indistinct outlines. However, 
neither the targets most frequently estimated correctly (i0e0 within nor 
those with no consistent pattern of error, possessed attributes in common to suggest 
an explanation for the pattern of the distance estimates0 
Co Table 5 gives a description of the 32 targets, with their distances 
and azimuths, the percentage of accurate estimates, and the percentage of in- 
accurate estimates, partitioned for over- and underestimation. 
Inclo #3 
1 TABLE 5 
DESCRIPTION OF TARGETS 4 
ARRANGED IN ORDER OF DECREASING PERCENTAGE OF UNDERESTIMATES 
RANGE #1 - Urban Area of Fort Knox; Observation Point on Knoll0 
RANGE #2 - Firing Range in Rolling Country; Observation Point, Top of Slight Rise0 
RANGE #3* - Firing Range in Partially Cleared Rolling Country; Observation Point 
in Flat Clearing, 
(Correct estimates are those within of true range; 
erroneous estimates are those more than 10J6 in error0) 
Tar 
Dis- 
TTorT 
rect 
% Erroneous 
Estimates 
Range 
No. 
get 
No0 
Description 
tance 
Yds. 
Azi- 
muth 
Remarks 
Esti- 
mates 
over 
short 
3 
4 
Range marker 
365 
E 
Sharply outlined, 
brightly striped. 
16 
9 
75 
2 
3 
Trunk of dead tree 
699 
NNS 
Framed by clearing in 
woods, but beyond the 
woods0 
18 
11 
71 
3 
U 
Jeep near snail 
shed 
149 
W 
Sharply outlined. 
12 
17 
70 
3 
6 
Range marker 
653 
SE 
Sharply outlined, 
brightly striped. 
23 
13 
63 
1 
2 
Large lone oak 
tree 
710 
NNS 
On upward slope of 
clearing. 
39 
4 
57 
2 
6 
Pile scrap metal 
543 
ENE 
Below horizontal line 
of sight—no obstruc- 
tion to vision. 
23 
21 
56 
3 
1 
Lone pine tree 
526 
NNE 
Up gradual slope to- 
ward horizon. 
15 
30 
53 
2 
7 
Jeep body 
333 
S 
Clearly seen—contrast 
low. 
24 
24 
52 
3 
i 
3 
Broken truck body 
1045 
ENE 
Top of body visible 
through brush—dead 
ground intervening be- 
yond crest of upward 
slope. 
24 
25 
50 
2 
Incl, B TABLE 5 (cont'd) 
Range 
No. 
Tar- 
Dis- 
Azi- 
muth 
rcor'-' 
rect 
Esti- 
mates 
% Erroneous 
Estimates 
g«t 
F°t 
Description 
tance 
Ids. . 
_ Remarks ... 
over 
phort 
1 
4 
Corner of large 
bldg. 
513 
ENE 
Sighted downward, no 
obstruction. 
31 
19 
49 
3 
8 
Jeep under cedar 
tree 
495 
WSW 
Tree symmetrical and 
sharply outlined. 
23 
28 
48 
1 
1 
Barracks bldg. 
843 
N 
Among trees, elevated 
above group of bar- 
racks In hollow. 
39 
18 
43 
2 
2 
Broken down truck 
456 
NNE 
Incompletely seen due 
to irregular ground. 
29 
29 
42 
2 
10 
Small shed 
775 
KSE 
Seen through trees 
near permanently 
located tanks. 
28 
32 
42 
3 
9 
Black truck 
603 
W 
Seen across hollow 
partly obstructed 
by brush. 
27 
31 
41 
2 
4 
Rusty broken down 
track 
732 
NNE 
Incompletely seen 
due to irregular 
ground. 
26 
34 
40 
3 
7 
Telephone pole 
702 
SW 
Low contrast. 
Against woods. 
20 
39 
40 
2 
8 
Truck body 
922 
£ 
Partly hidden by foli- 
age—^indistinct—con- 
trast very low. 
27 
37 
36 
I 
6 
Comer of 
barracks 
558 
S 
Sighted over high 
coal piles. 
35 
29 
36 
2 
9 
Safety marker 
805 
£ 
Brightly striped— 
but seen on edge. 
33 
32 
35 
3 
2 
Broken truck body 
785 
NE 
Little more than cab 
visible due to brush, 
terrain irregularity 
and L.w contrast. 
28 
37 
34 
Incl. #3 
3 TABLE 5 (eont»d) 
Tar- 
Die- 
% Cor- 
rect 
% Erroneous 
Estimates 
tenge 
No. 
get 
No? 
Description 
bance 
Asi- 
muth 
Remarks 
Esti- 
mates 
over 
short 
3 
5 
i 
Battered truck body 
825 
ESS 
Obscured by brush0 
-Contrast low. 
33 
34 
32 
1 
9 
Corner bldg0 
850 
W 
Only part of bldg, vis- 
ible along pared 
street. 
45 
23 
32 
3 
10 
Battered peep 
630 
w 
Obscured by brush and 
low contrast. 
22 
45 
31 
2 
1 
Cedar tree— 
symmetrical 
117 
11 
On slightly higher 
ground than 0oPo with 
moderate depression . 
intervening—seen 
against sky0 
12 
58 
30 
3 .= 
12 
Badly battered 
peep 
1880 
WWW 
On eroded slope across 
broad ravine. 
28 
44 
27 
2 
5 
Bare spot on 
distant ridge 
1760 
NE 
No specific object as 
target. Distance is an 
even mile. 
42 
32 
25 
1 
10 
Barracks 
840 
HNW 
On eminence partly 
hidden by trees. 
63 
14 
23 
1 
3/ 
Truck body with 
no wheels 
831 
NS 
On upward slope of 
grass covered clearing 
—outline sharp. 
53 
33 
13 
1 
5 
One room bldg0 
635 
E 
Along straight paved 
road. 
30 
59 
11 
1 
7 
Corner large 
bldg a 
325 
wsw 
Lone bldgo on hill- 
sighted upward 
against sky* 
30 
60 
9 
1 
a 
Upper eaves of 
bldgo 
371 
w 
Only tipper corner 
of bldg, visible 
along paved street. 
37 
56 
V 
6 
Indc .#3 
4 dc Arranged In similar of*der, Figure 3# a to £, shows graphically 
the distribution of errors on these targets. On 7~of them, or more of the 
estimates were within 10% of the true range0 Six of these were on the first 
range, which together with the lower average error on this range suggests that 
targets in urban areas present less difficulty than those in open country. There 
was a tendency to underestimate the distances on 19 targets, to overestimate on 8, 
while on 5, although the errors were numerous, no patter of error was discernible. 
Table 6 lists those targets which were most strikingly over- or underestimated, 
and presents the predominant character of error on them for all the men, and for 
the 10 beat men. Since the errors of both groups of men is in the same direction 
on both sets of target*, the impression is gained that the type of error is a 
function of the target rather than of the individuals doing the estimating. Nor 
did it appear that the pattern of error for any one man was characteristically 
over or short. 
•o These observations might be applicable to specific combat situations 
confronting armored crews* When the range must be estimated to a terrain feature 
in the vicinity of which an enemy strongpoint is suspected, the estimate is very 
likely to be more than 10% short if the object is one which has sharp outlines0 
If the target be a vehicle, or a tree, for example, its distance will tend to be 
underestimated if its contours are sharp, and overestimated if only part of it 
is seen, and that Indistinctly* 
Incl* #3 
5 TABLE 6 
COMPARISON OF PERFORMANCE OF ALL MEN WITH THAT 
OF 10 BEST MEN ON SELECTED TARGETS. 
Errors of all men*I 
Errors of best 10 men* 
tenge 
No. 
Target 
No0 
Targets 
Consistently 
Underestimated 
% of 
Total 
% 
Over 
% 
Under 
No0 of men 
over 
No* of men 
under 
3 
4 
Range marker, 365 yds. 
84 
9 
75 
0 
5 
2 
3 
Trunk of dead tree in 
clearing, 699 yds. 
82 
11 
71 
0 
4 
3 
11 
Jeep at close range, 
149 yds0 
77 
17 
70 
0 
6 
3 
6 
Range marker, 653 yds0 
76 
13 
63 
0 
2 
1 
2 
Large lone oak, 
710 yds0 
61 
4 
57 
0 
3 
2 
J s’ 
6 
Pile of scrap metal 
in field, 543 yds* 
77 
21 
56 
1 
4 
Targets 
Consistently 
Overestimated 
1 
a 
Upper front part of bldg 
visible along street, 
371 yds* 
. 62 
56 
6 
5 
1 
1 
7 
Corner large bldg, on 
hill, 325 jrds. 
69 
60 
9 
4 
1 
1 
5 
Shed visible at end 
of street, 635 yds0 
70 
59 
11 
5 
0 
* Differing from true range by more than 10JC 
lncl„ #3 
6 FIG. I 
DISTRIBUTION OF 4472 RANGE ESTIMATES MADE BY 
TRAINEES IN GUNNERY AT THE A.R.TC. 
(a) 20 TARGETS ON 2 RANGES ESTIMATED BY ALL MEN. 
{ b ) AN ADDITIONAL 12 TARGETS ON A 3 RD. RANGE 
ESTIMATED BY HALF THE MEN. 
PERCENT OF ALL ESTIMATES 
MAGNITUDE OF ERRORS 
WITHOUT REGARD TO SIGN 
Incl, #4 
FIG. I FIG. 2 
MEDIAN RANGE ESTIMATION ERRORS OF 86 
INEXPERIENCED GUNNERY TRAINEES AT A.RIC. 
(32 TARGETS ON 3 RANGES) 
NO. OF MEN 
MEDIAN % ERROR FOR 32 ESTIMATES. 
WITHOUT REGARD TO SIGN 
FIG. 2 
Incl* FIG. 3 A-P 
DESTR1BUTI0N OF RANGE ESTIMATION 
ERRORS FOR EACH TARGET. 
INDICATING BOTH THE DIRECTION AND MAGNITUDE 
OF ERROR 
THE CHARTS ARE ARRANGED IN THE 
SAME ORDER AS IN TABLE 5. 
RANGE*!- 10 TARGETS- APPOX. 175 MEN 
*2-10 TARGETS-APPOX. 175 MEN 
*3-12 TARGETS-APPOX. 86 MEN 
FIG. 3 FIG.3-A 
RANGE-3 
TARGET-4 
FREQUENCY 
RANGE-2 
'TAR GET-3 
FREQUENCY 
PERCENTAGE OF ERROR 
Incl. A 
FIG.3-A F1G.3-B 
RANGE-3 
TARGET-11 
FREQUENCY 
RANGE-3 
TARGET-6 
FREQUENCY 
PERCENTAGE OF ERROR 
Incl. 
FIG. 3-B F1G.3-C 
_ RANGE-1 
TAR GET-2 
FREQUENCY 
RANGE-2 
TARGET-6 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3-C 
Incl. #4 FIG.3-D 
RANGE -3 
TARGET-1 
FREQUENCY 
RANGE -2 
TARGET- 7 
FREQUENCY 
PERCENTAGE OF ERROR 
[ncl. A 
FIG.3-D FIG.3-E 
_ RA N 6E - 3 
TA R G E T - 3 
FREQUENCY 
RANGE -1 
TARGET-4 
FREQUENCY 
PERCENTAGE OF ERROR 
Incl. 
FIG.3-E FIG.3-F 
RANGE-3 
TARG ET-8 
FREQUENCY 
RANGE-1 
-TARGET- I 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG 3-F 
Incl. #k FIG.3-G 
RANGE-2 
'TARG E T-2 
FREQUENCY 
RANGE-2 
TAR6ET-10 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3-G 
Incl. ftU FIG.3-H 
RANGE-3 
TARGET-9 
FREQUENCY 
_RANGE-2 
TARGET-4 
V 
o 
z 
LU 
3 
O 
LU 
or 
Ll. 
PERCENTAGE OF ERROR 
FIG.3-H 
Incl. M FIG.3-1 
RANGE-3 
TA R G ET-7 
FREQUENCY 
RANGE-2 
TARGET-8 - 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3-1 
!ncl. jfk FIG.3-J 
RANGE-1 
TARCET-6 
FREQUENCY 
RANGE-2 
TARGET-9 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3-J 
Incl. tik FIG.3-K 
RANGE-3 
TARGET -2 
FREQUENCY 
RANGE-3 
TARGET'S 
FREQUENCY 
PERCENTAGE OF ERROR 
Incl. 
FIG. 3-K FIG.3-L 
RANGE-1 
TARGET-9 
FREQUENCY 
RANGE-3 
TARGET-10 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3-L 
Incl. #k FIG.3-M 
RANGE-2 
TARGET- I 
FREQUENCY 
RANGE-3 
TARGET-12 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3 “M 
Incl. $4 FIG.3-N 
RANGE-2 
TARGET-5 
FREQUENCY 
RANGE- I 
TARGET-IO 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3-N FIG.3-0 
RANG E I 
TA R G E T- 3 
FREQUENCY 
RANGE I 
TARGET-5- 
FREQUENCY 
PERCENTAGE OF ERROR 
FIG. 3-0 
Xncl, A FIG.3- P 
RANGE I 
TAR 6 ET-7 
FREQUENCY 
RANGE I 
TARGET-8 
FREQUENCY 
PERCENTAGE OF ERRORS 
Incl. #4 
FIG.3-P