Re: Why not model actual train consists?

Robert kirkham

I'm not doing exactly that - I have no train consists from my modelling area and era to work from (the nearest consists are dated some 25 years later).

However, on the advice of a mathematically minded friend, I do try to model the cars shown in photos taken in my modelled location and eta (a five mile stretch of the CPR's waterfront trackage in Vancouver through 1946). I waiver away from this approach to allow inclusion of cars shown in photos taken from 1944 through 1948, just to beef up the total numbers. That has its risks as the port had different traffic during the war and immediate post war periods, though it is hard to know how much anything changed. O well. (I do know of one shipyard ceasing construction. We stopped shipping supplies to the USSR near the end of the war. That's about all I really know so far.)

With a lot of the photos I've collected being taken at a distance and with variable quality lenses and ability of photographers, I often cannot make out car numbers. Sometimes all I can make out is that a car is steel and the same height as the car next to it. Or steel frame (like a fowler car). Or a gon with 13 panels.

My intent is to document the observable traits of the cars shown in the photos I have collected into a spread Excel sheet. Each photo is a spread sheet. In some cases the observable data about a car will be only one or two traits; in others it is everything you would want for modelling down to the specific car number.

For modelling purposes, I will develop my fleet based on the distributions that can be discerned from the photos. So, for example, of cars where the owner is known, I will develop ratios for car ownership. For basic car types (box, automobile, gon, flat, reefer, etc) I can develop proportions based on the cars that have those traits. For wood, steel and composite construction types, I can again discern ratios from the known cars in the photos. Likewise for home road versus off road cars.

Often the limitations in what one can see in a photo kick in. So I can see that a string of Fowler stock cars sit on a track, but cannot tell what type of roof, or the number of end posts, or door type. In those cases, the method I am using dictates that I can model any fowler stock car to fill that slot in the overall distribution and still be consistent with the observed data. The same approach is applied to modified 1937 steel boxcars, UTLX tank cars, etc., etc.

At a certain point I have to account for the cars in the photos that I cannot discern much about (usually the basic car type is evident, but even that could be subtly wrong - eg. if all I can see is a roof on a house car, I probably know it isn't a reefer, but it may be a boxcar, automobile car or even a ventilator; usually I can also discern approx. length and height). The unknown cars do form part of the overall car distribution I am trying to model. But when I turn to model that part of the distribution, I am "allowed" to include cars that are chosen from what is available on the market without being too picky. Given that the models available to the hobby tend to be those that were most common, using such cars to represent the unknowns in my roster isn't going to create too many obvious anomalies. I'll save scratch-building and modifying kits for those cars I can clearly identify!

The result should be a fleet that resembles the actual data I have.

Also, to the extent certain combinations of cars are on certain tracks, I can discern some limited things about the use of the yard tracks, and the order of cars on any given track could be used as a basis for making up trains (i.e. since trains were received into yard tracks and pulled from yard tracks, a yard track equals part or all of a past or future train).

The reality is that the amount of information discernable from photos is likely to exceed my model building time as well as my needs. When I eventually start to approach the ratios evident in the photos, the less-filled slots in the roster will be more evident and allow me to focus my work. For now, I can pretty much target anything of interest without fear of distorting the eventual roster.

Rob Kirkham

From: "Jim Betz" <jimbetz@...>
Sent: Monday, May 23, 2011 7:52 AM
To: <STMFC@...>
Subject: [STMFC] Why not model actual train consists?


We keep re-visiting the topic of freight car distribution ...
and discussing how to represent the freight cars for a point
in time or an "era" ... and then, presumably, adjusting the
mix of freight cars on our layouts ...

So I'm prompted to ask "Why not model actual trains?" As in
find a train sheet you like and go for it - with selective
compression - of course. And then extend it to several trains.

Is anyone out there doing this? Thinking seriously about it?
Tried it and found they couldn't field enough models to fill
the bill (how close did you get)? Selectively compress out the
models you don't have as a first cut? No, substitutions allowed
(same number series but different number)? One box car - from
the correct era - is just as good as the next?

You could even preserve the order of the cars in the train
even if you aren't modelling all of them. And 'just' adjust
the waybills for the layout?
- Jim Why Not

Join to automatically receive all group messages.