Distribution of the boxcar fleet, 1948-50

Dave & Libby Nelson <muskoka@...>

Having had the topic of home vs. foreign cars come up again I decided to
crunch a few numbers & see whether the data from my conductors books was in
any way similar to an ORER from about the same time. The focus is on
boxcars. The first number in the table below is percentage each roads own
roster of boxcars is of the April, 1950 North American boxcar fleet. My
database lacks the military roster, but as I used a fleet total of 834664
cars I doubt the missing cars would skew the results I show. The second
number is the percentage of total spottings that road has in my 1948
conductors books. 1231 boxcars were recorded by the conductor, making them a
sample size of 0.15% of the total boxcar fleet.

I cut off the listing at the accumulated 80th percentile of the North
American Fleet (i.e., the remaining 20% would have taken up another 80

If the numbers are close the boxcars are geographically evenly
distributed;if they are not close, some undetermined factor is causing more
or less cars to have been recorded by the conductor.

Results are listed below, sorted by expected % in descending order.

road expected actual
CN 7.97% 0.24%
PRR 7.95% 8.20%
NYC 7.70% 7.96%
CP 6.55% 0.08%
ATSF 4.20% 4.47%
MILW 3.84% 2.84%
BO 3.43% 3.82%
SP 3.33% 2.27%
SOU 3.13% 11.94%
UP 2.82% 3.01%
CNW 2.78% 3.98%
GN 2.68% 2.11%
CBQ 2.59% 3.01%
IC 2.52% 2.27%
NP 2.34% 2.44%
RI 2.18% 2.84%
MP 2.17% 2.11%
CO 1.69% 2.52%
SLSF 1.58% 1.79%
ERIE 1.44% 2.19%
LN 1.39% 3.01%
WAB 1.39% 0.57%
ACL 1.36% 1.79%
SAL 1.13% 1.38%
NW 1.06% 1.46%
SOO 1.06% 1.06%

As one can see, the actual matches the expected very closely, with 3 large
exceptions: Canadian roads are under-represented and the home road (SOU) is
over represented. None of these should be a surprise. One of the smaller
exceptions can also be easily explained: The route from which the conductors
books are taken is a short distance from Knoxville TN, a major gateway
between the Southern and the L&N and the city where most of these cars have
or will pass through. This may also be a factor in the C&O numbers.

The reader is left to draw their own conclusions on what this might mean, if
anything, for building up their roster of boxcars.
Dave Nelson