Freight car distribution


Anthony Thompson <thompson@...>
 

Dave Nelson wrote:
P.S. I miss Tim. He'd have willingly stepped into this discussion a long time ago. I don't much like to write about this anymore... it's a bit like explaining that the world is indeed round. It was a startling discovery and worthy of debate when first revealed but it has been explained for 10 years now. Search the archives. It's all still there.
Exactly. Wish I could have said it as well. And the world IS still round.

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@signaturepress.com
Publishers of books on railroad history


Tim O'Connor
 

Mike Brock wrote

... an error of 200% ... an error of 34% ... an error of 100%.
Many people who don't understand statistics think that any
variance from the MEAN indicates an ERROR. If everyone in
the U.S. had no money, and Bill Gates had a trillion dollars,
then the AVERAGE American would have $3,333 dollars. Then when
Mike visits his friends and finds that no one in the room has
any money, he'd call that an ERROR. Which it isn't.

Tim Gilbert did understand statistics. He knew that a random
distribution can result in some very interesting variations.
Considering the large numbers of freight cars in the U.S. those
samples Mike cites are completely consistent with a random
distribution.

Throwing dice is random too. If you throw 6 sixes in a row, is
that an ERROR? No. Is it unexpected? Not at all. Does it change
the chances of throwing 6 more sixes in a row? No. Is it likely
(probable) that you will? No.

Bottom line Mike: If you want to model ONE DAY on the UP main
line, you can run any thing you want. Because of all of the
possible permutations of freight car combinations that occurred
on the UP in 1952, chances are that your collection is not that
different than ONE day that occurred during that year.

Tim O'Connor


railsnw1 <railsnw@...>
 

Lately I have been doing research on the Yakima Valley Transportation
Co. and have been going through the 1967 Freight Abstracts for cars
received and forwarded for the year. While this date is out of scope
I found several instances of Canadian Pacific boxcars being loaded
with lumber at Northern California and Oregon sawmills and then
shipped to Yakima Pine Products in Yakima who further processed the
lumber.

Richard Wilkens

--- In STMFC@yahoogroups.com, Anthony Thompson <thompson@...> wrote:

Steve Lucas wrote:
Tax laws?? The CPR did purchase PS-1 boxcars from Pullman-
Standard
for this very reason. Yet, those tax and customs laws did not
prevent
US roads from keeping Canadian cars for their own use, simply
paying
the demurrage on them.
This is not an area of my own expertise, but we have been
told by
several people on this list in the past, that Canadian cars could
only
move in the U.S. to destinations to unload, then return empty, or
else
the Canadian owner would have to pay U.S. taxes on the cars if they
remained in use in the U.S. (if I'm remembering the story
correctly).
So yes, Canadian cars brought newsprint to Los Angeles, for
example,
but went straight back. I don't know about Canadian rules regarding
U.S. cars in Canada.
If this story is wrong, Steve, please enlighten us with the
right
story (or correct my wrong memory of what transpired previously).

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@...
Publishers of books on railroad history


devansprr
 

--- In STMFC@yahoogroups.com, Tim O'Connor <timboconnor@...> wrote:

Mike Brock wrote

... an error of 200% ... an error of 34% ... an error of 100%.
Many people who don't understand statistics think that any
variance from the MEAN indicates an ERROR. If everyone in
the U.S. had no money, and Bill Gates had a trillion dollars,
then the AVERAGE American would have $3,333 dollars. Then when
Mike visits his friends and finds that no one in the room has
any money, he'd call that an ERROR. Which it isn't.

Tim Gilbert did understand statistics. He knew that a random
distribution can result in some very interesting variations.
Considering the large numbers of freight cars in the U.S. those
samples Mike cites are completely consistent with a random
distribution.

Throwing dice is random too. If you throw 6 sixes in a row, is
that an ERROR? No. Is it unexpected? Not at all. Does it change
the chances of throwing 6 more sixes in a row? No. Is it likely
(probable) that you will? No.

Bottom line Mike: If you want to model ONE DAY on the UP main
line, you can run any thing you want. Because of all of the
possible permutations of freight car combinations that occurred
on the UP in 1952, chances are that your collection is not that
different than ONE day that occurred during that year.

Tim O'Connor
Tim is exactly correct. The science of operations research covers
deviations from average in great detail. It can have significant
impact on the design of many systems. The "Standard Deviation"
specifies the regular deviation from average, and I haven't seen any
standard deviation data published on home vs. foreign road cars. A
review of pictures from my era and railroad of interest suggests the
standard deviation for "home" box car percentages is quite high, less
so for gons, and much smaller for hoppers. And the occurance of one
foreign road's cars, on one train, can deviate a LOT from the averages.

I would think modelers who are modeling class I mainlines have quite a
bit of latitude in this area, but strings of TH&B cars in Wyoming
would be stretching it a bit.

Conversely, there was a good TKM article a few months back on
estimating the roads for a branch outside of Pittsburgh based on the
actual industries served, their suppliers/customers, and photographs
of the area. Here it is hard to justify using national boxcar fleet
averages.

Nearness to interchanges is also a strong factor. During WWII the DRGW
and P&LE had almost identically sized box car fleets, but if modeling
the PRR in Pittsburgh, I doubt D&RGW box cars were sighted as often as
P&LE on the PRR main.

For me (WWII PRR main in central PA) the challenge for a boxcar fleet
is how many NYC cars would there be? Should they be weighted heavier
than the Reading (interchange at Harrisburg)? I would expect the PRR
agents might quickly send empty NYC cars back to interchange points
empty and load New England bound freight into PRR cars.

During WWII the PRR had 9.2% of the North American boxcar fleet, the
NYC 6.1% (1st and 4th, with CP and CN 2nd and 3rd). Pictures suggest
PRR boxcars vastly outnumber NYC box cars on PRR rails, and in fact
some of the midwestern roads, with boxcars fleets significantly less
than half the size of NYC's, are much more prevelant. So how many NYC
box cars should I have? I think I have to have... one? ;-)

And even though CP and CN boxcars, combined, outnumbered Reading 14 to
1, Reading boxcars are more prevelant in photographs from the area.

Clearly their is room for artistic license, and modelers should be
able to bias one's fleet to some extent based on the models available.
Fortunately for me there are several WWII PRR boxcars available, and
the limited number of accurate NYC WWII era cars is not a big problem
(seems like most NYC boxcar models are post-war?)

Dave Evans


Steve Lucas <stevelucas3@...>
 

I agree with both of you gents. But Dale, please don't extrapolate
my mention of one TH&B car showing up in Florida to mean that strings
of them would show up in Wyoming. This is about as likely as seeing
strings of UP boxcars in Halifax, NS!

Steve Lucas.

--- In STMFC@yahoogroups.com, "devansprr" <devans1@...> wrote:

--- In STMFC@yahoogroups.com, Tim O'Connor <timboconnor@> wrote:

Mike Brock wrote

... an error of 200% ... an error of 34% ... an error of 100%.
Many people who don't understand statistics think that any
variance from the MEAN indicates an ERROR. If everyone in
the U.S. had no money, and Bill Gates had a trillion dollars,
then the AVERAGE American would have $3,333 dollars. Then when
Mike visits his friends and finds that no one in the room has
any money, he'd call that an ERROR. Which it isn't.

Tim Gilbert did understand statistics. He knew that a random
distribution can result in some very interesting variations.
Considering the large numbers of freight cars in the U.S. those
samples Mike cites are completely consistent with a random
distribution.

Throwing dice is random too. If you throw 6 sixes in a row, is
that an ERROR? No. Is it unexpected? Not at all. Does it change
the chances of throwing 6 more sixes in a row? No. Is it likely
(probable) that you will? No.

Bottom line Mike: If you want to model ONE DAY on the UP main
line, you can run any thing you want. Because of all of the
possible permutations of freight car combinations that occurred
on the UP in 1952, chances are that your collection is not that
different than ONE day that occurred during that year.

Tim O'Connor
Tim is exactly correct. The science of operations research covers
deviations from average in great detail. It can have significant
impact on the design of many systems. The "Standard Deviation"
specifies the regular deviation from average, and I haven't seen any
standard deviation data published on home vs. foreign road cars. A
review of pictures from my era and railroad of interest suggests the
standard deviation for "home" box car percentages is quite high,
less
so for gons, and much smaller for hoppers. And the occurance of one
foreign road's cars, on one train, can deviate a LOT from the
averages.

I would think modelers who are modeling class I mainlines have
quite a
bit of latitude in this area, but strings of TH&B cars in Wyoming
would be stretching it a bit.

Conversely, there was a good TKM article a few months back on
estimating the roads for a branch outside of Pittsburgh based on the
actual industries served, their suppliers/customers, and photographs
of the area. Here it is hard to justify using national boxcar fleet
averages.

Nearness to interchanges is also a strong factor. During WWII the
DRGW
and P&LE had almost identically sized box car fleets, but if
modeling
the PRR in Pittsburgh, I doubt D&RGW box cars were sighted as often
as
P&LE on the PRR main.

For me (WWII PRR main in central PA) the challenge for a boxcar
fleet
is how many NYC cars would there be? Should they be weighted heavier
than the Reading (interchange at Harrisburg)? I would expect the PRR
agents might quickly send empty NYC cars back to interchange points
empty and load New England bound freight into PRR cars.

During WWII the PRR had 9.2% of the North American boxcar fleet, the
NYC 6.1% (1st and 4th, with CP and CN 2nd and 3rd). Pictures suggest
PRR boxcars vastly outnumber NYC box cars on PRR rails, and in fact
some of the midwestern roads, with boxcars fleets significantly less
than half the size of NYC's, are much more prevelant. So how many
NYC
box cars should I have? I think I have to have... one? ;-)

And even though CP and CN boxcars, combined, outnumbered Reading 14
to
1, Reading boxcars are more prevelant in photographs from the area.

Clearly their is room for artistic license, and modelers should be
able to bias one's fleet to some extent based on the models
available.
Fortunately for me there are several WWII PRR boxcars available, and
the limited number of accurate NYC WWII era cars is not a big
problem
(seems like most NYC boxcar models are post-war?)

Dave Evans


Steve Lucas <stevelucas3@...>
 

I'm convinced that there was a lot of hoarding of good cars going on
on most roads, Canadian and American, in the timeframe of STMFC. And
it occurs to me that somewhere I read of US roads' reluctance to send
their cars to Mexico, for fear that the cars might remain for some
time on the Mexican rail system.

Yes, The AAR Car Service rules are very clear regarding use of empty
off-line cars, but were they followed?

Let me put it to you another way. Do you drive the speed limit and
stop at stop signs? Does your employer do everything that they are
supposed to? Railroads are buisnesses first, not charities. Their
goal is to make money. Following Car Service rules, well if it's
possible without losing money, I'm sure that the railroads followed
them. So a dirty car would very likely be sent back to its home
road. But a nice, new one that a customer wants, and may give the
local managers or switch crew some 80-proof consideration for at now
and/or at Christmas, hmmmmmmmmm....

Steve Lucas.

--- In STMFC@yahoogroups.com, "railsnw1" <railsnw@...> wrote:

Lately I have been doing research on the Yakima Valley
Transportation
Co. and have been going through the 1967 Freight Abstracts for cars
received and forwarded for the year. While this date is out of
scope
I found several instances of Canadian Pacific boxcars being loaded
with lumber at Northern California and Oregon sawmills and then
shipped to Yakima Pine Products in Yakima who further processed the
lumber.

Richard Wilkens

--- In STMFC@yahoogroups.com, Anthony Thompson <thompson@> wrote:

Steve Lucas wrote:
Tax laws?? The CPR did purchase PS-1 boxcars from Pullman-
Standard
for this very reason. Yet, those tax and customs laws did not
prevent
US roads from keeping Canadian cars for their own use, simply
paying
the demurrage on them.
This is not an area of my own expertise, but we have been
told by
several people on this list in the past, that Canadian cars could
only
move in the U.S. to destinations to unload, then return empty, or
else
the Canadian owner would have to pay U.S. taxes on the cars if
they
remained in use in the U.S. (if I'm remembering the story
correctly).
So yes, Canadian cars brought newsprint to Los Angeles, for
example,
but went straight back. I don't know about Canadian rules
regarding
U.S. cars in Canada.
If this story is wrong, Steve, please enlighten us with
the
right
story (or correct my wrong memory of what transpired previously).

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705
www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@
Publishers of books on railroad history


cvsne <cvsne@...>
 

I'll preface my remarks by stating that my primary goal is not to
create some sort of statisically valid model freight car fleet but
instead to recreate the visual aspects of the trains I'm modeling.
On my layout I want CV Train 490 to have FGE/WFE reefers on the head
end, since that's what the photos show were there etc . . . After
all, we look at plans and photos of prototype cars when we model
them - and don't decide to model a certain style of door or
brakewheel because they were the most "common."

That said, the study of the movement of cars across the country is a
fascinating topic well worthy of study in its own right, as Tim
showed.

Proximity to interchanges doesn't really matter - most of the time
the "home road" cars are out wandering the country hopefully making
money for their owners. The main reason you see so many PRR, ATSF,
CN etc . . . cars has nothing to do with the old model
railroad "Home road - 50%, Primary interchange roads 25% of the
fleet, "Secondary" interchange roads, 15% etc . . ." that some
modelers seem hung up on still is tied to the percentage of PRR,
NYC, CN, etc . .. cars in relation to the overall national car
fleet that some sort of traffic pattern.

I have a complete set of switchlists for White River Junction
Vermont in September-November of 1952. One of the guys in the CV
Historical Society compiled these into lists by reporting marks and
car types.

The findings - For the boxcars - about 3% of the cars were CV cars.
No surprise there, the CV hardly had a large fleet. More surprising,
B&M - the "interchange" road in WRJ - hardly any cars - something
like 8 over the entire 7 week period - not even 1%!!

Largest percentage, by an overwhelming majority? Canadian National.
One could argue that as the CV was a subsidiary of the CN, the CN
cars were "home road" in some form or another - The percentage of CN
cars? 50.5%

Maybe those old model railroaders were really onto something?

For me, I opted to crack open another bottle of Scotch and not worry
too much about it . . .

Marty


Steve Lucas <stevelucas3@...>
 

Tony--

You are very correct about the tax implications for Canadian roads
when their equipment (locos and cars) were used in the US for a
longer time than necessary. CN, CP, and BC Rail all had cars built
in the US after the timeframe of STMFC to try to deal with this.

But the Canadian roads had the same problems that the US roads did in
getting their cars back. And probably paid a lot of tax and duty
because of it. Not sure about freight cars, but certin CN passenger
cars were noted as having "duty paid" on them for international use.

Maybe I'm missing something, but under Car Service rules, could empty
cars not be loaded "in the direction of their home road"? If so,
this would have been a licence to do practically what the railroad
wanted to with an empty car. And would explain a lot of anamolies.

Steve Lucas.

--- In STMFC@yahoogroups.com, Anthony Thompson <thompson@...>
wrote:

Steve Lucas wrote:
Tax laws?? The CPR did purchase PS-1 boxcars from Pullman-
Standard
for this very reason. Yet, those tax and customs laws did not
prevent
US roads from keeping Canadian cars for their own use, simply
paying
the demurrage on them.
This is not an area of my own expertise, but we have been
told by
several people on this list in the past, that Canadian cars could
only
move in the U.S. to destinations to unload, then return empty, or
else
the Canadian owner would have to pay U.S. taxes on the cars if they
remained in use in the U.S. (if I'm remembering the story
correctly).
So yes, Canadian cars brought newsprint to Los Angeles, for
example,
but went straight back. I don't know about Canadian rules regarding
U.S. cars in Canada.
If this story is wrong, Steve, please enlighten us with the
right
story (or correct my wrong memory of what transpired previously).

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@...
Publishers of books on railroad history


Steve Lucas <stevelucas3@...>
 

I never had the pleasure of meeting Tim, but it's patently obvious
that his work was very thorough. What is reasonably typical, on the
average? He's probably closer than any of us ever will be. And it's
easy enough to reconcile it with the Jan., 1953 ORER reprint. In
fact, I'm going to use his work to help me understand freight car
distribution.

To say that his work is ignored? Not by me. I appreciate your
referencing it.

Steve Lucas.

--- In STMFC@yahoogroups.com, Anthony Thompson <thompson@...> wrote:

Steve Lucas wrote:
I'd like to say that there is a method that works for every time
and
location, but the fact is that with two million cars in the North
American rail system, just about any combination of railroads'
cars
are possible at a specific time and location.
I'm glad Tim Gilbert isn't with us any longer, to see his
work
ignored. The issue is not, and has not been, what is POSSIBLE. The
issue is, what is reasonably typical, on the average. Tim's work
addressed that in considerable detail, and I for one found it
persuasive.

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@...
Publishers of books on railroad history


Bruce Smith
 

On Aug 12, 2008, at 8:13 AM, devansprr wrote:
For me (WWII PRR main in central PA) the challenge for a boxcar fleet
is how many NYC cars would there be? Should they be weighted heavier
than the Reading (interchange at Harrisburg)? I would expect the PRR
agents might quickly send empty NYC cars back to interchange points
empty and load New England bound freight into PRR cars.

During WWII the PRR had 9.2% of the North American boxcar fleet, the
NYC 6.1% (1st and 4th, with CP and CN 2nd and 3rd). Pictures suggest
PRR boxcars vastly outnumber NYC box cars on PRR rails, and in fact
some of the midwestern roads, with boxcars fleets significantly less
than half the size of NYC's, are much more prevelant. So how many NYC
box cars should I have? I think I have to have... one? ;-)
Dave,

If you've read my article in TKM on the PRR fleet ;^) you know that the PRR averaged around 50% home road, but that includes about 75% home road hoppers, 50% home road gons and 25% (estimated) home road boxcars. Those numbers also reflect cars stored or awaiting repairs, so the actual number in trains is slightly less.

For the rest of your boxcar fleet, the NYC cars should be roughly 6% of your total fleet. Thus if you have 100 non-PRR boxcars, six should be NYC. Of those cars, a substantial portion should be the USRA steel boxcar (Westerfield, and styrene?(someday).


Regards
Bruce

Bruce F. Smith
Auburn, AL
http://www.vetmed.auburn.edu/index.pl/bruce_f._smith2

"Some days you are the bug, some days you are the windshield."
__
/ &#92;
__<+--+>________________&#92;__/___ ________________________________
|- ______/ O O &#92;_______ -| | __ __ __ __ __ __ __ __ |
| / 4999 PENNSYLVANIA 4999 &#92; | ||__||__||__||__||__||__||__||__||
|/_____________________________&#92;|_|________________________________|
| O--O &#92;0 0 0 0/ O--O | 0-0-0 0-0-0


Anthony Thompson <thompson@...>
 

Steve Lucas wrote:
I never had the pleasure of meeting Tim, but it's patently obvious that his work was very thorough . . . To say that his work is ignored? Not by me. I appreciate your referencing it.
Glad you've looked it over. I think it's quite impressive. We really need to get it into print somewhere, since it's natural that not everyone has access to, or wants to access, the archives of this particular list. Maybe RP CYC? Ed or Pat, are your scanning this?

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@signaturepress.com
Publishers of books on railroad history


Anthony Thompson <thompson@...>
 

Steve Lucas wrote:
You are very correct about the tax implications for Canadian roads when their equipment (locos and cars) were used in the US for a longer time than necessary. CN, CP, and BC Rail all had cars built in the US after the timeframe of STMFC to try to deal with this.

Maybe I'm missing something, but under Car Service rules, could empty cars not be loaded "in the direction of their home road"? If so, this would have been a licence to do practically what the railroad wanted to with an empty car. And would explain a lot of anamolies.
Yes, the Car Service rules were really consensus standards, and we know from the testimony of many people who worked in past eras that the rules were NOT followed if not convenient. But remember, the responsible entities on tax issues were not Car Service but governmental bodies. I don't know about you, Steve, but my experience with the tax people is that they frown on doing things which are against the rules but "convenient." Any taxation authority which got the bit in their teeth might well be quite aggressive on this point, though I have no idea if this actually did happen with freight cars in the steam era.

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@signaturepress.com
Publishers of books on railroad history


benjaminfrank_hom <b.hom@...>
 

Ton Thompson wrote:
"We really need to get it into print somewhere, since it's natural
that not everyone has access to, or wants to access, the archives of
this particular list. Maybe RP CYC? Ed or Pat, are your scanning
this?"

Tony, Tim and I had discussed before he passed away. Tim had drafted
a paper with his findings, but it was extremely detailed and by his
own admission something we probably could not have pushed in the
hobby press, even in a publication such as RP Cyc. Our biggest
hangup was how to express the information in terms that most
hobbyists could understand without losing the heart of his analysis.
It's my desire to follow through on this, but it's competing with
many other projects, not to mention the day job and the reserves.

Unfortunately, this thread has shown this work to be sorely needed.
I share Dave Nelson's frustation - we seem to keep coming back to
this subject, and people continue to fail to look in the archives and
continue to ignore Tim's work. It's true it's not the be-all and end-
all on this subject, but it has far more analytical rigor than the
unsupported opinions that have been presented over the past few
days. Tony's challenge that the burden of proof rests with those
making the argument is dead on - if you think the "Boomer Pete/Chubb
Neighbors Weigh Heavier" is a more valid model, then back it up with
some data. If you can't, then you're just wasting all of our time.


Ben Hom


cvsne <cvsne@...>
 

Ben,

Contact me off list on this - I think the new e-magazine "Model Rail
Hobbyist," may be an ideal outlet for publishing some of Tim's work.

I planned to share some of Tim's LCL research (a whole other kettle of
fish) in some way in that magazine.

Marty


Cyril Durrenberger
 

The mix of freight cars to use on a model railroad is indeed a complex topic. I have a few comments.

1. The mix can be expected to vary significantly from year to year and from railroad to railroad, so general statistics may not provide a good answer for the specific railroad and year or era that one is modeling.

2. The mix of cars on a model railroad is going to be greatly influenced by the industries the model railroad "serves", that is the industries that have placed on the model railroad. If the model railroad has through freights then the mixture of cars on these freights can be more like what was found on the railroad at that time and place. Industry and railroad wide data can portray a very different picture than the cars that were used at a specific location or segment of the line.

For example the GN and NP had fleets of ore cars used in Minnesota, but it is not likely that these ore cars showed up in the states to the west of Minnesota. Also during the time period of interest to this list, the ore traffic was very seasonal. In most years they did not have much ore traffic when the ports where not open due to ice. In some cases they did have a few all rail trains.

3. The annual reports that each railroad sent to the state railroad commission contain data that should be considered. I am familiar with the reports sent to Texas and Minnesota during the time period of most interest to me -1900 to 1920. The exact data varied from year to year, but they had statistics on the tonnage or car loads of a number of different commodities carried on the railroad. For many years this was divided into classes - loaded on road and unloaded on road, loaded off road and unloaded on road, loaded on road and sent off road. Loaded off road and unloaded off road (sometimes called bridge traffic).

4. Prior to 1920 there was a list of mileage (or fees paid related to mileage) for private owned cars. Since in many cases the refrigerator and tank cars were privately owned, this can provide information on these cars that would be seen on the railroad. In general there are patterns, but these vary from year to year. For example there were no PFE cars prior to 1906, since that is when they began business, and the number of miles increased as their fleet became larger. The mileage for tank cars grew greatly in the teens and the number of companies shipping became much larger each year. Many times there would be a car from a private owner that would show up for one trip for one year, and never again.

5. I used the ORER data for the year of concern to me, 1910 to develop the fleet of SP Atlantic Lines box cars and SP Pacific Lines box cars to use on my railroad. Basically these were based on the percentage of ownership for this year, assuming that the cars used would likely reflect these data. The Atlantic Lines owned a large fleet of flat cars to serve the wood products industries so this was taken into account.

6. I have also used a process to determine the cars to use from other railroads, but will not go into that here.

I can provide more information if desired, but remember we are talking about 1910. The process would be applicable to almost any era. I had written a detailed article on this process for the tank car fleet I am using and sent it to Scale Rails. They wanted to publish it, but lost all of the photos. I was not included to go through all of that work again.

Cyril Durrenberger


Bruce Smith <smithbf@auburn.edu> wrote:
On Aug 12, 2008, at 8:13 AM, devansprr wrote:
For me (WWII PRR main in central PA) the challenge for a boxcar fleet
is how many NYC cars would there be? Should they be weighted heavier
than the Reading (interchange at Harrisburg)? I would expect the PRR
agents might quickly send empty NYC cars back to interchange points
empty and load New England bound freight into PRR cars.

During WWII the PRR had 9.2% of the North American boxcar fleet, the
NYC 6.1% (1st and 4th, with CP and CN 2nd and 3rd). Pictures suggest
PRR boxcars vastly outnumber NYC box cars on PRR rails, and in fact
some of the midwestern roads, with boxcars fleets significantly less
than half the size of NYC's, are much more prevelant. So how many NYC
box cars should I have? I think I have to have... one? ;-)
Dave,

If you've read my article in TKM on the PRR fleet ;^) you know that
the PRR averaged around 50% home road, but that includes about 75%
home road hoppers, 50% home road gons and 25% (estimated) home road
boxcars. Those numbers also reflect cars stored or awaiting repairs,
so the actual number in trains is slightly less.

For the rest of your boxcar fleet, the NYC cars should be roughly 6%
of your total fleet. Thus if you have 100 non-PRR boxcars, six
should be NYC. Of those cars, a substantial portion should be the
USRA steel boxcar (Westerfield, and styrene?(someday).

Regards
Bruce

Bruce F. Smith
Auburn, AL
http://www.vetmed.auburn.edu/index.pl/bruce_f._smith2

"Some days you are the bug, some days you are the windshield."
__
/ &#92;
__<+--+>________________&#92;__/___ ________________________________
|- ______/ O O &#92;_______ -| | __ __ __ __ __ __ __ __ |
| / 4999 PENNSYLVANIA 4999 &#92; | ||__||__||__||__||__||__||__||__||
|/_____________________________&#92;|_|________________________________|
| O--O &#92;0 0 0 0/ O--O | 0-0-0 0-0-0


Anthony Thompson <thompson@...>
 

Ben Hom wrote:
Tony, Tim and I had discussed before he passed away. Tim had drafted a paper with his findings, but it was extremely detailed and by his own admission something we probably could not have pushed in the hobby press, even in a publication such as RP Cyc.
Ben, I would be willing to help, perhaps with editing it somewhat to fit what a publisher such as RP CYC might want. I would advocate checking with Ed and Pat before assuming that it couldn't be done.
I note that several list members are chiming in with tangential issues: assigned service cars, local and regional freight operations, etc. These are certainly interesting and for some modeling needs, would overwhelm other considerations. And Mike Brock's desire to model specific trains is also a challenging problem for anyone who deals with a specific time and place. But this is not what Tim's focus was. It was on the free-running cars, primarily box cars. It is important to realize that those cars DO NOT generally follow the patterns of those other cars just referenced.
I am not criticizing the discussion of other issues, which as I said are definitely important and need research. What I am criticizing is anyone who thinks they can fling out an opinion on box cars without some solid data. If you're one of them, be aware that Tim raised the bar far over your head.

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@signaturepress.com
Publishers of books on railroad history


Mike Brock <brockm@...>
 

My, my...lots of interest in frt car distribution.

Dave Nelson writes:

"Here's a simple explanation of the issue:
---------
Safeway sells all sorts of things, one of which is Yoplait Yogurt.
Railroads move all sorts of freight cars, one of which is a SP 40', etc,
etc, etc boxcar.

On average, not every shopping cart has any Yoplait Yogurts.
On average, not every train has any SP boxcars.

When Yoplait yogurt is on sale, one of Dave Nelson's shopping carts once had
20+ Yoplait yogurts in it.
When UP is pulling empties across Sherman Hill one of UP's train once had
20+ SP boxcars in it."

Oooops. On April 16, 1949 a UP frt train moved East pulled by Big Boy 4004 [ which still exists in Holiday Park in Cheyenne and BLI now makes a model of it...hmmm ] with 31 SP box cars in tow. MT? Nope. Most carried lumber but a few had other things. This train was one of five "lumber" trains in Fraley's book. The first train on Mar 1, had 39 cars carrying lumber, 7 were SP box cars. Train two on Mar 3,'49, was 77 cars in length and contained 27 SP cars carrying lumber. Of the 27, 25 were box cars and 2 were flats. Compare this to a train on Apr 7, '49, which contained 98
cars. Of these, 58 carried lumber and 4 were SP box cars. One more SP box car was in the train giving SP box cars a 5% presence....closer to the national average. The train on April 20 contained 32 cars of lumber, one of which was an SP car. The video tape The Big Boy Collection contains several trains containing many SP box cars, the most significant one being the "infamous" one pulling 36 PFE reefers followed by at least 36 SP box cars. I have speculated...but cannot confirm...that the shot was near Buford, WY, and might be westbound. Tim Gilbert proposed that this train was one of MT's but we have no proof.

Dave continues with:

"So which do you want to talk about -- one of my one-of shopping carts when
Yoplait is on sale or the overall average of what Safeway sees day in and
day out?"

Well, with regard to the UP frt trains, we really don't know if the lumber was on sale or not. Meaning, we don't know how common the lumber trains were. Their consist with regard to SP box cars is inconsistent as well. Nevertheless, there were certainly UP frt trains in the Super Nova class of containing a hugely larger number of SP box cars than the national % would indicate.

"IMO only the overall average has any informative value as you can
create from that any combination of items for the individual events as you
see fit. OTOH, following the one-of event **as-if** it had informative
value does nothing but create identical events, repeated ad-nauseum."

Meaning...I think...that unusual trains were...well...unusual? Actually, I don't think the data shows that. In fact, it is surprising to me anyhow to note the variety of the consists. For example, the train on 3-11-49 contained 83 cars....26 carrying ore in box cars. Not a UP car in the train...although some PFE reefers were present.

Tim O'Connor wrote:

Mike Brock wrote

... an error of 200% ... an error of 34% ... an error of 100%.
"Many people who don't understand statistics think that any
variance from the MEAN indicates an ERROR."

Well...I thought the purpose of the theory was for it to project a number for box cars of various RR's to be found on the rails of other RR's. Having lived in Las Vegas for six months and having received training at some of their best crap tables I do understand that the laws of probability do predict but don't guarantee. Nevertheless, the theory was/is only a theory and I think its complete validity is still open to discussion. IOW, I didn't accept that it did not treat closely associated RR's differently from others...i.e., for UP, SP as compared to NP or FEC. Tim Gilbert and I discussed this at length.

"If everyone in
the U.S. had no money, and Bill Gates had a trillion dollars,
then the AVERAGE American would have $3,333 dollars. Then when
Mike visits his friends and finds that no one in the room has
any money, he'd call that an ERROR. Which it isn't."

No. I'd call that strange [ not unlike those 36 SP box cars in the train ] given that Bill Gates is an American and lives in the U.S.

To me, the issue is that a theory was put forth...based on wheel reports...that would give some basis for determining the distribution of box cars on other RR's. The theory...as I understand it...was that the % of foreign RR box cars would match the % of the foreign RR's box cars in the national box car fleet. It sounds worthy but, IMO, the proof of the puddin' is in the eatn'. IOW, how does it work with the available data? Not bad but apparently it doesn't work for box cars of the home RR on its own rails. Tim Gilbert and Dave Nelson apparently observed this and reacted accordingly. Thus, home RR box cars are expected to be present in about twice their national avg on their own rails. OK. I don't think it works well for box cars of closely associated RR's either...as in the case of SP, CB&Q and C&NW cars with respect to UP. Why would this be? Frankly, I don't care. It's what the available data shows. It's as simple as that.

Even assuming I am correct, how does any of this pedict or establish a frt car population for the modeler? Well...it helps. OTOH, as I have shown above, it doesn't work well in determining the population of individual trains. It helps in that we know that box cars went everywhere...as did stock cars and to some extent gons. However, at least on the UP, we cannot ignore the fact that of 34 frt trains in the spring of 1949, 15 had one or less SP box cars, 9 had none at all while two had astronomically high numbers of 27 and 31 [ all loaded ]. Almost as strange as the fact that no one in the U.S. had any money [ they bought too many frt car models? ] while one of those in the U.S. had a trillion dollars [ and no frt cars? ].

Mike Brock


Malcolm Laughlin <mlaughlinnyc@...>
 

--- In STMFC@yahoogroups.com, "Steve Lucas" <stevelucas3@...> wrote:

I never had the pleasure of meeting Tim, but it's patently obvious
that his work was very thorough. What is reasonably typical, on the
average? He's probably closer than any of us ever will be.

======

I'll have more to say on this later, but first maybe we need to review
what Tim actually said. Yesterday morning, after urging from Mike
Brock, I did a search in our archives on "freight car distribution"
and got 266 messages to look at. During the course of the day I
browsed the 150 or so from before I joined this list. I'm still
confused as to what Tim really concluded about distribution by
ownership. Is there anyone who has those details or can point us to
the messages that have the conclusions,.

In regard to the above snippet from Steve.

My key point is that "reasonably typical" "on average" is not a really
useful concept. Like of my many drives between Boston and New York it
would be extraordinary for one to be the average time, and don't forget
that in the car stuff we're talking about distributions with all kinds
of skewness and normal distributions aren't demonstrably common.

"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.


Anthony Thompson <thompson@...>
 

Malcolm Laughlin wrote:
My key point is that "reasonably typical" "on average" is not a really useful concept. Like of my many drives between Boston and New York it would be extraordinary for one to be the average time, and don't forget that in the car stuff we're talking about distributions with all kinds of skewness and normal distributions aren't demonstrably common.
What's your point? That there is variability and skewness does not make the average uninteresting; in fact, it emphasizes the average as the central point of the data. No one, least of all Tim, ever said that the average was the entire story. But if all I think about is the variability, and ignore the average, I really have no idea what I'm talking about.

Tony Thompson Editor, Signature Press, Berkeley, CA
2906 Forest Ave., Berkeley, CA 94705 www.signaturepress.com
(510) 540-6538; fax, (510) 540-1937; e-mail, thompson@signaturepress.com
Publishers of books on railroad history


devansprr
 

--- In STMFC@yahoogroups.com, Bruce Smith <smithbf@...> wrote:

Dave,

If you've read my article in TKM on the PRR fleet ;^) you know that
the PRR averaged around 50% home road, but that includes about 75%
home road hoppers, 50% home road gons and 25% (estimated) home road
boxcars. Those numbers also reflect cars stored or awaiting repairs,
so the actual number in trains is slightly less.

For the rest of your boxcar fleet, the NYC cars should be roughly 6%
of your total fleet. Thus if you have 100 non-PRR boxcars, six
should be NYC. Of those cars, a substantial portion should be the
USRA steel boxcar (Westerfield, and styrene?(someday).


Regards
Bruce
Bruce,

Pretty sure I have read ALL your TKM articles, and greatly appreciate
your efforts.

But I'm still not convinced on NYC boxcars. I certainly plan on more
than 1, but looking at pictures I rarely see NYC box cars in through
freights on the PRR Middle and Pittsburgh divisions. But I think I
recall seeing several at the Altoona freight house.

I think it is legitimate to bias populations based on the local
situation. Isn't it likely that nearly all NYC box cars loaded in New
England, with destinations in the upper west - west of the
Mississippi, would be routed over the NYC's water level route if they
touched NYC rails before PRR rails? Or would most shippers specify the
entire route of their shipment?

And east bound MT NYC boxcars from the west (admittedly VERY rare
during WWII) would probably be interchanged onto the NYC well west of
Pittsburgh.

To me, that would make west bound NYC box cars through Huntingdon, PA
less frequent than EB CB&Q loads, which is what the photographs seem
to indicate, even though the NYC fleet was bigger than the CB&Q.

The Altoona NYC sightings also make sense. As one of the world's
largest railroad car and locomotive shops (the largest during WWII
era?), there must have been a steady stream of parts and materials
flowing into Altoona from all over the North-East, and many of those
suppliers would have been served by the central, hence some NYC box
cars end up at the Altoona freight house.

Clearly Mike's UP reports show SIGNIFICANT deviations from the
averages. Lacking wheel reports, I'm more inclined to work off of
photographs from the locations being modeled rather than national
averages. National averages can help fill in the gaps where no other
data is available, but the science behind operations research would
suggest that significant deviations from the average are the NORM, not
the exception. Freight car movements were not a random process, and
therefore should not be expected to exhibit standard deviations
typical of Gaussian distributions.

I realize I'm not at the modeling level of this group's membership - I
monitor for the wealth of information provided, but I have limited
time, and frankly limited modeling skills, that are necessary to build
a bunch of resin NYC box cars, so if NYC ends up being 1-2% of my box
car fleet instead of 6%, I'm not losing any sleep over it, and if
anyone complains about a shortage of NYC box cars during a future op
session, I can confidently reply "not today", and know that it is a
"prototypical", and statistically reasonable response.

Respectfully,
Dave Evans