Re: Numbers/percentages of important box car types

Gatwood, Elden J SAD <Elden.J.Gatwood@...>

Yes, I am! I think your points are well taken, and I did NOT mean those
small facts to turn into this debate once again, but it is always an
interesting topic.

A person intimately involved in statistical research would say that we need
way more data to be able to make real use of predictive models, but we don't
have all that consist data. A person might argue that we need at least 25
trains worth of data to develop a model with a certain confidence of being
statistically useful.

I would still like to see a group of fleet breakdowns for various folks made
available. Bruce's is a good one; what about some others? And what do you
think those statistics on car numbers mean for your fleets in regards to
timeframe and location?


-----Original Message-----
From: STMFC@... [mailto:STMFC@...] On Behalf Of Mike
Sent: Tuesday, June 13, 2006 12:20 PM
To: STMFC@...
Subject: Re: [STMFC] Numbers/percentages of important box car types

I know I should not comment on this but what the heck. You guys have all the
fun while I research viruses.

There are certain subjects that seem to come up for discussion every few
months. Hence, like lengthy debates about color, we once again see another
discussion about frt car...or box car...populations on our model RR's. These
discussions are entirely within scope and, I suppose, the only problem with
them might be that they might be copied and used by some enlightened
entrepreneur as a non drug response to insomnia. I do think, however, that
we should at least attempt to clarify certain the risk of
immediately putting some members into an unexpected coma. Thus:

Bruce Smith writes:

"Here we go again... That's a falsehood perpetrated by model
railroaders. The facts seem to show pretty clearly that with a very
few special exceptions regionality (location) and connecting
railroads have NOTHING to do with populations of boxcars. Connecting
railroads may affect specific train makes-up (like Mike's favorite
example the espee forwarder) but those are usually "balanced
translocations", as we geneticists call them."'s what we know from the tiny bit of data we have from the two
Wyoming Fraley's which have been analyzed...mine and Tim Gilbert's. The
following analysis was presented to the STMFC last Feb and March:

"The number of total cars to SP box cars in my report is 20.4 while 57.6 in
Tim's...4.9% in mine to 1.7% in Tim's. The number of SP box
9.89% of the total number of box cars. There were 233 UP box cars. The
number of SP box cars to foreign road box cars is 11.9%.

My Fraley shows the consists of 34 frt trains traveling between Rawlins and
Laramie, WY in 1949. Those trains include 2 WP box cars...#14334 and #40026.
The same trains show 53 definite C&NW box cars [ Tim Gilbert identified
55 ]. The C&NW % of the national fleet is about 2.79%. WP's % is 0.30%.
There are 1464 box cars in my Fraley. C&NW's 2.79% predicts 41 C&NW box
cars, 4.4 WP box cars. The error for C&NW is +14 cars...3.75% or an error of
about 34%. The 2 WP cars equates to error of about 50%. The % of
the national fleet for SP [ not T&NO ] is 3.1%. This predicts 46 SP box
cars. The actual number is 136...or an error of 90 cars...about a 200%
error. The national % for CB&Q is 2.57% which predicts 38 cars. The actual
number is 75 producing an error of about 100%.

Thus, we find some RRs with major connections to the UP's Wyoming/Nebraska
trunk line to have a significantly higher number of box cars present in UP
trains than that of the national fleet. These include C&NW, SP, and CB&Q.
OTOH, WP has a significantly lower number. While C&NW and SP are direct
connections to significant markets from the Wyoming/Nebraska trunk line [
central California and the midwest through Chicago ], neither the WP or Q
appears to perform such a connection. The Q did have significant
interchanges with UP in the Nebraska, eastern Wyoming and Colorado areas."

It seems clear from the 1949 Fraley data that box car populations of SP,
C&NW, and CB&Q on the UP between Laramie and Green River, WY do not follow
the national percentages during the spring of 1949. Period. We can speculate
why this might be true but, IMO, that is nothing but theory and it doesn't
really matter anyhow if one is modeling the spring of 1949. For this priod,
the population of frt cars on the UP between Laramie and Green River is
available...for 34 frt trains.

Other information, such as the video of the infamous UP train with 36+ SP
box cars shot in 1953 apparently near Buford, WY, and 4 or 5 additional
train footage shot during '53 confirms that large numbers of SP box cars
were accumulated in certain UP frt trains. Why this was so is again open to
speculation and theory but if one is modeling the area and time the reason
is not relevent. Such theory is, of course, relevent if one is trying to
draw conclusions in order to project populations in other areas or at
different periods of time. Unfortunately, IMO, such conclusions are simply
theory and the supporting data are not available. IOW, we don't have any
data...that I'm aware of...regarding the CB&Q in Nebraska, the Mopac and
Frisco from St. Louis to KC, ATSF from Chicago to LA etc. Knowing that the
actual numbers in the Laramie/Green River area are so much in error from the
national percentage prediction, how can one rely on such a theory to project
a population on the ATSF or SP in Arizona? Or even the L&N between Knoxville
and Cincinnati?

The Tim Gilbert/Dave Nelson theory that box car populations match...within
some "ball park"...the national percentage appears to be a "long term"
evaluation...say for a year. And, BTW, it is an interesting and useful
theory...definitely bringing light to an area filled with erroneous
conclusions from the Bruce notes. Anyhow, if one had frt conductor
books for an entire yr and IF the data showed a close match to the national
percentage, the modeler would still be left with the problem that he/she
doesn't model for the entire yr...unless they change the scenery about 4
times during the process. Yes...Wyoming scenery...think of the poor guy
doing any place north of south Ga...does change between Jan and June. IOW,
even if the population of a particular RR matched the national
percentage...and we don't know that it does...for a yr, it won't necessarily
in the shorter term.

So...what is one to do? Simple as can be. Don't sweat it. Acquire more cars
of the RRs with larger fleets, more cars of the RR you model, and one or two
cars of RRs with smaller fleets and move on. In fact, one could simply use
the national percentage. I's as good as any way to acquire more
cars of RRs with larger fleets, few cars of RRs with smaller fleets. That
will give you time to consider various responses to the Prototype Police as
to why you...modeling, say, the GN in Montana have a Seaboard hopper car or,
you, Armand, have an SP GS gon up in Vermont. The answer, BTW, is also
simple. Just shrug your shoulders and say..."I have no just showed
up". If the Prototype Police begin writing a ticket...quickly produce your
photo of the UP train in Wahsatch, UT [ The Streamliner, Vol 18 no 2, pg
17 ] with the 2 Mopac hoppers and 2 Rock Island GS gons ]. I don't, however,
suggest running 2 SP GS gons, Armand. Don't want to push the issue you

Anyone still awake?

Mike Brock

Yahoo! Groups Links

Join to automatically receive all group messages.