Re: Freight car distribution
toggle quoted messageShow quoted text
"He's probably closer than any of us ever will be."
I think I strongly disagree with that statement but need to find what
he actually said.
Ok, lets go over the same old ground ONCE AGAIN. There will be a bit more
of what I did as I can recall what I did. On the whole, Tim put in more
effort and posted more frequently.
About 10 or 11 years ago, on the old FCL, I started posting comments about
how data published by the AAR **suggested** that if there was an even
distribution of house cars across the nation that non-home road cars would
appear at any given location in numbers approximating the percentages each
road's fleet was in the national pool. Unfortunately, at that time I had no
detailed data to back up the claim and the idea was spurned.
Shortly after that I purchased 3 dozen wheel reports from the SRR (Asheville
NC to Spencer NC) and began transcribing some of them into a database so I
could do a detailed analysis. From that effort came the first hypothesis,
which was, as I recall, Non-Home Road US boxcars were distributed fairly
uniformly across the nation and that if one wanted to build a roster of cars
that reflected that, buy according to the percentages each roads boxcars
were relative to the boxcar fleet. I also posted those averages and backed
up the claim w/ examples. There was a lot of debate and IIRC most people
found the idea interesting but remained skeptical.
About 2 years later, Tim Gilbert joined the list and shortly after that
contacted me and said he too was working the same topic, using a wheel
report from the SRR (IIRC, Potomac Yard to some point RR west). We compared
notes and agreed more data supported what had already concluded. Both Tim
and I continued to accumulate wheel reports and do analysis on the data we
At some early point in time Tim and I agreed there was a need for certain
qualifiers... So what had started out as a general statement got reduced a
bit by the qualifiers "Post WWII", "Class 1", and "US Boxcar fleet". We did
that because we had no data in hand to indicate what happened before WWII,
what might be found on class II, III, Short line, and Terminal roads, and
for reasons unclear, Canadian boxcars were not appearing in expected
At some later time Tim took an interest in freight car mileage and expressed
a number of hypothesis from that data. This was something I never could get
my head around so I can't explain it here.
At some later time Tim concluded that flatcars behaved the same as boxcars
-- a fairly even distribution across the nation. This came from wheel
report data and ICC car mileage data.
At some later time Tim concluded general economic trends would influence the
distribution of cars by virtue of contracting or expanding the number of
cars sent offline. I had drawn the same conclusion earlier but being focused
on just 1950 I had not taken the time to transcribe and quantify it; Tim
had. (n.b., a lesson there: there is always too much data, never enough
At some later time I think that Tim concluded that gondolas might behave the
same as boxcars and flatcars -- a fairly even distribution across the
nation. I'm just not sure anymore.
At some later time I obtained data from the Canadian Bureau of railroad
statistics and concluded from that data that Canadian marked boxcars
appeared in greatly reduced numbers than their fleet size suggested, roughly
10%, because Post WWII Canadian freight car shipments were 90% within
Canada. Somebody else (name forgotten) chipped in that there were certain
laws the explained why this was the case. This brought out all sorts of
counter arguments from fans of the CV. That was the last work I did on the
subject. Tim continued plugging away and, if you look thru the archives,
you will see that Tim posted much more frequently on the topic than I did
(he was retired, I wasn't.. He cared a lot, I was a bit indifferent).
Tim, Mike Brock, and I all obtained wheel reports from the UP. So entered
Mike into the debates.
About this time I suggested that the distribution of home road boxcar and
flatcars would vary on the home road according to whether a given location
tended towards mostly car shipments (high numbers of home road cars) or
mostly car receipts (low number of home road cars). I also suggested that
the distribution of foreign-road boxcars and flatcars
All thru this period I continued to do analysis on the overall freight car
fleet using a digital ORER, as well as commodity reporting based on very
detailed ICC data; Tim continued to explore the ICC blue book data. All of
this material often got included in posts on freight car distribution.
And then Tim died.
So FWIW, here's where I believe the hypothesis (several) stands today:
1) In the Post WWII years, Class I foreign-road, US marked boxcars and
flatcars were distributed fairly uniformly throughout the nation and each
type of car would be seen in numbers approximating each roads contribution
to the overall US fleet of that car type. IOW, if the NYC owned X% of the
US fleet of boxcars, then approximately X% of all foreign-road boxcars seen
on other railroads would be from the NYC.
2) Canadian Class I marked boxcars and flatcars would appear on US rails at
roughly 10% of the numbers approximating each Canadian roads contribution to
the overall US fleet of that car type. Because of regulation requiring the
prompt return of Canadian cars to Canada it was reasonable to expect a
higher concentration of Canadian marked cars on lines crossing the US
boarder than would be case if such cars wandered freely within the US.
3) The ratio of home to foreign road cars one would find on any given road
would change with overall US economic conditions, with the number of foreign
road cars rising in good economic times and falling in recessions. Actual
data of online and offline, US marked house cars and open top cars can be
found in the ICC Blue books for any given year.
4) The smoothness of the distribution of foreign road boxcar and flatcars
**may** be related to the number of and distance to the nearest interchange
points. IOW, the distribution might not be as even at corners of the US
(e.g., consider Bangor ME; Miami FL; San Diego CA; and Bellingham WA)
relative to, say, Chicago. How much the distribution changes as the number
of interchange points declines and distance to interchange points rises is
5) The number of home road boxcars and flatcars seen on the home road will
vary according to whether the location observed relative to the ratio of
outbound and inbound shipments. For instance, if Ditchwater has a 9:1 ratio
of shipments to car receipts one might expect to see home road cars in
protective service at Ditchwater as the railroad could not count on inbound
foreign road cars to appear frequently enough. OTOH, if Gotham City has a
1:9 ratio of shipments to car receipts one might expect to see very few home
road cars in protective service at Gotham City as there would be plenty of
foreign road MTY's that could be pressed into service.
6) The data does not support any hypothesis on the distribution of hoppers.
I'm not sure what Tim thought about gons or stock cars.
7) The hypothesis do not suggest anything about what one might see in any
given train but instead indicate what one might see in multiple trains over
a period of time. How many trains... How much time... The data does not
provide a definitive answer but does appear to hold true above 1000 cars.
8) The data indicates there is a **slightly higher occurrence** of adjacent
road foreign cars near interchange points with that foreign road but given
that most roads percentage of the nation fleet are very a small percentage,
in most cases the increase in actual numbers is nominal.
9) For purposes of building up a roster of model cars, the buyer would do
well to buy 20-25% home road and 75-80% foreign road boxcars and flatcars,
with the foreign road cars in rough proportion to what each foreign road
contributed to the US fleet. Ideally the buyer would acquire a large number
of additional cars for roads that contributed fewer cars to the US fleet and
using them as a pool, to cycle individual road names in and out to represent
the last 5-10% of his foreign road cars.