Richard, this is about what some of us have been saying, although not so articulately as you. It is not a statistical game to be won, but it is about believability and approximation. The data talked about helps, but it is far from being the golden rule. And as you note, the fact is that we simply do not have a large percentage of the usable and desirable freight cars in reasonable production, and probably will not for years, if ever. And older cars were often used for the reasons you state. The variables are enormous. So the trick seems to be not having a fleet that looks out of place, that does not have any glaring errors, and gives the appearance of a reasonable facsimile of the typical, or even atypical put prototypical, freight train of the chosen period, knowing that it is probably off quite a bit given the constraints and variables, but still is in the ball park. I think that is all we can hope for, and some will be more successful than others. But the striving to get closer to nirvana is also an important factor and one some folks really enjoy. So it goes.

Have we completely stomped this thread to a pulp?

John Stokes
Bellevue, Wa





To: STMFC@yahoogroups.comFrom: rhendrickson@opendoor.comDate: Mon, 18 Aug 2008 15:43:17 -0700Subject: Re: A Purpose For Frt Car Distribution Studies. Was: Re: [STMFC] Re:Fwd: Re: Freigh




Sorry to be slow in responding; I was away over the weekend.On Aug 17, 2008, at 4:03 PM, Bruce Smith wrote:> On Fri, August 15, 2008 5:45 pm, devansprr wrote:>> > But I do not have good info on details such as tank car manufacturer> type (e.g. 11, 21, 27, etc). If that exists let me know and I'll > search> for it - there is a wealth of data in this site - it could use a full> word index system for quicker and more accurate searches.>> Tank cars - (Information from Tim Gilbert summarizing the January > 1943 ORER)> Union Tank Car Co. 38,707> General American Transportation Co. 27,867> Pennsylvania-Conley Tank - div. of GATX 10,327> Shippers Car Line 6,949> Sinclair Refining Co. 6,447> AT&SF 3,567> US War Department 2,475> SP - Pacific Lines 2,219> Gulf Oil Co. 1,551> UP 1,096> Dupont 1,068> Sun Oil Co. 1,035>> However, given that the ORER does not allow us to decipher types, > and tank> car companies were notorious for mixing cars from multiple makers into> series, the best I think we can do is try to convey the impression > of the> WWII fleet.>From my perspective, "conveying the impression" isn't just the best we can do, it's exactly what we're trying to do: convey the visual impression in miniature of the real thing. It seems to me that, in the process of flogging the subject of freight car distribution to death, some list members have lost sight of that fact. Building a realistic freight car fleet for a model that represents a specific railroad at a specific time and place is what most, if not all, of us want to do. Every kind of information that helps us do that is valuable: train and yard photos, conductor's train sheets, switch lists, and, yes, the insightful statistical analyses of Tim Gilbert, Dave Nelson, et. al. Depending on prototype RR, location, era, and other variables each of us has to interpret the available data for ourselves, with the objective of running trains which will be visually convincing not just to the average viewer but to the cognoscenti, i.e. the kind of people who subscribe to this list. Above all, we want to avoid the kinds of anachronisms or improbabilities that will make the illusion go "pop!" - e.g., in the case of WW II tank car trains, running RC's otherwise very nice models of postwar welded ICC-103Ws.>Having said that, let me elaborate a bit on Bruce's useful observations about WW II tank cars.> Some rules of thumb -> -UTLX had the biggest fleet so the X-3 would likely be the most common> tank (Lots of Sunshine kits!) and IIRC, the 8K size was NOT the most> common (10K?)>Yes, 10K gal. X-3s were far more numerous than 8K gal. X-3s, and should be ubiquitous. But remember that UTL also owned many cars that were older types (the frameless type Vs and the similar type Xs with underframes in 6, 8, and 10K sizes) as well as cars of other than UTL design (e.g., a whole bunch of early GATC radial-course 8K gal. cars which UTL purchased new in the early'20s as well as some 8K and 10K AC&F Type 21s, and even a few STC tank cars, all acquired second-hand).>>> -AC&F production of type 21 cars outnumbered type 27 by a significant> margin, and the 8K size was the most common of the type 21s.>More importantly, by the 1930s when the Type 27s had replaced the Type 21s, few oil companies were buying new tank cars, having turned to pipelines as a more efficient and economical way to transport petroleum products. Similarly, few GATC Type 30s were in petroleum service. The vast majority of the cars that were equipped for, and available for, crude oil shipments during WW II were built in the 'teens and '20s.>>> -GATC built tanks were fairly common and we have no reasonable > model in HO>Type 30 models are on the horizon, but - as noted above - the GATC tank cars we most need are earlier cars, many with radial course tanks. And we need a lot of those.>>> -"Oddballs" such as the UTLX "van Dyke", and earlier type 7 and 11 > tanks> should not appear in great numbers, but were certainly in use.>I agree that the Van Dykes would not have been numerous, though a surprising number of them did get resurrected for the WW II oil crisis. I don't agree about the earlier AC&F cars; Type 11s and Type 17s, and even some Type 7s, were very common in the fleets of many petroleum shippers, and the many WW II photos I have show them all over the place.>>> So, for now, a fleet of Sunshine and Walthers/P2K tanks with a few> Intermountain, Precision Scale, Speedwitch (NATX tanks), IM/Tichy > bashed> USG-A, Southern Car and Foundry (STC cars) and whatever else I've> forgotten (like the RC brass GATC cars and other brass cars) will > make a> reasonable fleet with the exception of missing GATC cars. Like > boxcars,> this approach will give you the varying sizes and features seen in > WWII> era "pipelines on rails">All true, as far as it goes (and Bruce goes about as far as he reasonably can at present). We certainly need GATC cars built ca. 1915-1930, but the problem is that GATC kept changing them so that some designs were built only for a very few years, so we probably won't see these in styrene any time soon. I would add that we very much need models of the UTL class X cars, which were built in large numbers and lasted, in many cases, into the 1960s. Something else we need are 3-course AC&F Type 21 tanks to go on the P2K underframes; the four course tanks modeled by L-L were built only in the early 1920s, and I have many photos of three course 10K gal. Type 21s in petroleum service. Once Jon Cagle gets his STC models in production, we could also use Pennsylvania Tank Car Co. underframes; PTC's plant was right next door to STC's in Sharon, PA and PTC bought tanks from STC to put on their own underframes, so a decent PTC underframe combined with Cagle's STC tanks and a few minor detail modifications would enable us to model another group of tank cars that were produced in large numbers in the 1920s. Other tank cars that have been modeled in brass over the years will add variety, if you can find (and, nowadays, afford) them. However, in the forseeable future, it's not going to be possible to model realistic WW II tank trains by popping RTR plastic models out of their boxes and putting them on the track.As for comprehensive prototype information on steam era tank cars, I'm working on it, but a couple of other books are currently higher on my priority list.Richard Hendrickson

Sorry to be slow in responding; I was away over the weekend.

On Aug 17, 2008, at 4:03 PM, Bruce Smith wrote:

On Fri, August 15, 2008 5:45 pm, devansprr wrote:

But I do not have good info on details such as tank car manufacturer

type (e.g. 11, 21, 27, etc). If that exists let me know and I'll
search
for it - there is a wealth of data in this site - it could use a full
word index system for quicker and more accurate searches.

Tank cars - (Information from Tim Gilbert summarizing the January
1943 ORER)
Union Tank Car Co. 38,707
General American Transportation Co. 27,867
Pennsylvania-Conley Tank - div. of GATX 10,327
Shippers Car Line 6,949
Sinclair Refining Co. 6,447
AT&SF 3,567
US War Department 2,475
SP - Pacific Lines 2,219
Gulf Oil Co. 1,551
UP 1,096
Dupont 1,068
Sun Oil Co. 1,035

However, given that the ORER does not allow us to decipher types,
and tank
car companies were notorious for mixing cars from multiple makers into
series, the best I think we can do is try to convey the impression
of the
WWII fleet.

From my perspective, "conveying the impression" isn't just the best
we can do, it's exactly what we're trying to do: convey the visual
impression in miniature of the real thing. It seems to me that, in
the process of flogging the subject of freight car distribution to
death, some list members have lost sight of that fact. Building a
realistic freight car fleet for a model that represents a specific
railroad at a specific time and place is what most, if not all, of us
want to do. Every kind of information that helps us do that is
valuable: train and yard photos, conductor's train sheets, switch
lists, and, yes, the insightful statistical analyses of Tim Gilbert,
Dave Nelson, et. al. Depending on prototype RR, location, era, and
other variables each of us has to interpret the available data for
ourselves, with the objective of running trains which will be
visually convincing not just to the average viewer but to the
cognoscenti, i.e. the kind of people who subscribe to this list.
Above all, we want to avoid the kinds of anachronisms or
improbabilities that will make the illusion go "pop!" - e.g., in the
case of WW II tank car trains, running RC's otherwise very nice
models of postwar welded ICC-103Ws.

Having said that, let me elaborate a bit on Bruce's useful
observations about WW II tank cars.

Some rules of thumb -
-UTLX had the biggest fleet so the X-3 would likely be the most common
tank (Lots of Sunshine kits!) and IIRC, the 8K size was NOT the most
common (10K?)

Yes, 10K gal. X-3s were far more numerous than 8K gal. X-3s, and
should be ubiquitous. But remember that UTL also owned many cars
that were older types (the frameless type Vs and the similar type Xs
with underframes in 6, 8, and 10K sizes) as well as cars of other
than UTL design (e.g., a whole bunch of early GATC radial-course 8K
gal. cars which UTL purchased new in the early'20s as well as some 8K
and 10K AC&F Type 21s, and even a few STC tank cars, all acquired
second-hand).

-AC&F production of type 21 cars outnumbered type 27 by a significant
margin, and the 8K size was the most common of the type 21s.

More importantly, by the 1930s when the Type 27s had replaced the
Type 21s, few oil companies were buying new tank cars, having turned
to pipelines as a more efficient and economical way to transport
petroleum products. Similarly, few GATC Type 30s were in petroleum
service. The vast majority of the cars that were equipped for, and
available for, crude oil shipments during WW II were built in the
'teens and '20s.

-GATC built tanks were fairly common and we have no reasonable
model in HO

Type 30 models are on the horizon, but - as noted above - the GATC
tank cars we most need are earlier cars, many with radial course
tanks. And we need a lot of those.

-"Oddballs" such as the UTLX "van Dyke", and earlier type 7 and 11
tanks
should not appear in great numbers, but were certainly in use.

I agree that the Van Dykes would not have been numerous, though a
surprising number of them did get resurrected for the WW II oil
crisis. I don't agree about the earlier AC&F cars; Type 11s and Type
17s, and even some Type 7s, were very common in the fleets of many
petroleum shippers, and the many WW II photos I have show them all
over the place.

So, for now, a fleet of Sunshine and Walthers/P2K tanks with a few
Intermountain, Precision Scale, Speedwitch (NATX tanks), IM/Tichy
bashed
USG-A, Southern Car and Foundry (STC cars) and whatever else I've
forgotten (like the RC brass GATC cars and other brass cars) will
make a
reasonable fleet with the exception of missing GATC cars. Like
boxcars,
this approach will give you the varying sizes and features seen in
WWII
era "pipelines on rails"

All true, as far as it goes (and Bruce goes about as far as he
reasonably can at present). We certainly need GATC cars built ca.
1915-1930, but the problem is that GATC kept changing them so that
some designs were built only for a very few years, so we probably
won't see these in styrene any time soon. I would add that we very
much need models of the UTL class X cars, which were built in large
numbers and lasted, in many cases, into the 1960s. Something else we
need are 3-course AC&F Type 21 tanks to go on the P2K underframes;
the four course tanks modeled by L-L were built only in the early
1920s, and I have many photos of three course 10K gal. Type 21s in
petroleum service. Once Jon Cagle gets his STC models in production,
we could also use Pennsylvania Tank Car Co. underframes; PTC's plant
was right next door to STC's in Sharon, PA and PTC bought tanks from
STC to put on their own underframes, so a decent PTC underframe
combined with Cagle's STC tanks and a few minor detail modifications
would enable us to model another group of tank cars that were
produced in large numbers in the 1920s. Other tank cars that have
been modeled in brass over the years will add variety, if you can
find (and, nowadays, afford) them. However, in the forseeable
future, it's not going to be possible to model realistic WW II tank
trains by popping RTR plastic models out of their boxes and putting
them on the track.

As for comprehensive prototype information on steam era tank cars,
I'm working on it, but a couple of other books are currently higher
on my priority list.

Richard Hendrickson

Walter M. Clark writes:

"What we have is the problem that always occurs when using statistics.
Most people won't accept that a relatively small sample size can
forecast much larger populations with extreme accuracy. Just listen
to the arguments after a national poll is released, where they polled
only a few thousand people and used that data to project within just a
few percentage points of error what the entire Nation's population thinks."

Well...it all depends upon the selection of data process. IOW, do you believe that if you chose your entire sample from either of the attendees at one of the political conventions that you would get usable results?

"Tim's and Dave's analysis is for an entire year."

It is? That's news to me...although you may be correct. I've been asking about that. Tim's 1947 Fraley, however, contains the UP trains from only about a month...mine, which he says blew his theory to "hell" covers a month and a half.

"I had a professor in Graduate School who, in the Statistics class,
always brought a cloth drawstring bag with 100 red and 100 black poker
chips to class. No matter what he was teaching, and no matter how big
or how small the numbers he was working with, a handful of chips from
that bag was ALWAYS extremely close to what the calculations said."

Ah. Reminds me of my first encounter with a crap table in Vegas. The Showboat Casino out on Henderson Hy...1964. While I knew 2 weeks later that the dice have no memory [ at least they weren't sposed to ]...I noticed some guy making quite a successful run. Soon I was betting against him. Soon I was broke. I have no idea how many passes the guy made...27 or so was supposed to be the known record...but he came close to breaking it. Had I come to Vegas 2 weeks earlier I would not afterward have to work for 33 yrs for NASA. Incidentally, I sat through a similar statistics meeting. In this case, it was a Lockheed presentation in which the presenter claimed that their software was 85% successful. My division chief [ he spoke with a German ascent for some reason...one of his friends had a problem that when he got excited his right arm flew up as if he was pointing toward a distant mountain ] noted that we didn't really think 85% was very good and that success was a 0 or 1. Sorta like trying to drive from Kennedy Space Center to St. Louis only to have your car explode while crossing the Tennessee River in Cairo, IL. 85% success?

Mike Brock

Tim

Where else would you hit a nail?<

Sometimes on my thumb or forefinger

Aidrian

--- In STMFC@yahoogroups.com, "Jack Burgess" <jack@...> wrote:

<snip> Does anyone know if the rate has been set now?

Jack Burgess
www.yosemitevalleyrr.com

A recently received mailing from Sunshine listed room rate as $92.00.
Gene Green

Except for all those times the polls are wrong, for varying reasons. As for the ORER, as previously noted, those can be way off in terms of actual cars in service, so another variable.

So what you are saying Tim and Dave are saying is that their stats show us precisely how many box cars from each of the railroad in the US at the time being modeled should be present in your personal fleet of models in order to run consists that, over the year, match the national averages? And if you are only modeling one small segment of the larger railroad, say the GN in 1951, the percentages still count, you just reduce your box car fleet in numbers by the percentages so that you stay constant in the differences? So if there would be 100 CN cars on the GN system at that time, and you only have 90 cars total, you reduce the number proportionally? But what if your line is not part of the main line, but an important branch working the logging and mining routes in the Cascade foothills? I may be completely missing all this, but I am trying to understand how this would work on the average person's home layout, or is that not of any consequence?

John Stokes
Bellevue, WA



To: STMFC@yahoogroups.comFrom: walterclark@earthlink.netDate: Mon, 18 Aug 2008 21:52:50 +0000Subject: [STMFC] Re: Freight car distribution




--- In STMFC@yahoogroups.com, "Bruce Smith" <smithbf@...> wrote:>> Folks,> > I grow weary of this. No one has answered Tony's challenge to provide> evidence that supports a "regional bias in boxcars" position. It seems> that any evidence that does not support this position is dismissed as> "absurd" or "biased" and opinion and hyperbole are used in lieu of data.> Many of the arguments that have been used about statistics are not based> in a sound understanding of the field. For example, deviations from the> mean are EXPECTED and do not invalidate the mean. Certainly,modeling the> deviations would be.... deviant!> > What I find truly remarkable is the wealth of DIFFERENT types ofdata that> seem to support the national fleet as a starting place for a model> representation. The Charles data is flawed at best, yet it supports the> model. Wheel reports from different locations support the model...> > -It ISN'T about individual trains (How often do we need to say this?)> -Individual trains may have very specific make-ups that varysignificantly> from the national averages.> -It is about a "fleet"> -That fleet will then fluctuate on the layout of the owner, providing> deviations from the mean... and modeling those deviations that almost> certainly occur in real life.> -It is a STARTING place - I have yet to hear anyone offer a different> starting place based on data other than their own opinion, and frankly,> I'd rather take my chances with the data ;^)> -The model DOES NOT preclude the use of oddballs, but it does INFORM the> modeler that these are oddballs.> > Over and out> Bruce> > Bruce Smith> Auburn, AL>Bruce, and all,What we have is the problem that always occurs when using statistics.Most people won't accept that a relatively small sample size canforecast much larger populations with extreme accuracy. Just listento the arguments after a national poll is released, where they polledonly a few thousand people and used that data to project within just afew percentage points of error what the entire Nation's population thinks.Tim's and Dave's analysis is for an entire year. Not every train, andnot every train in a single month. Neither Tim nor anyone else whoaccept Tim's and Dave's calculations are claiming that. It is thetotal of all box cars over an entire year. This can (not must, we allcan do what we darn well please on our own model railroads) help medecide which box car kits, and in what quantities, I need toreasonably represent the possible activity in my month of November1941 on my mythical short line that connects to the Southern Pacificin northern California. I had a professor in Graduate School who, in the Statistics class,always brought a cloth drawstring bag with 100 red and 100 black pokerchips to class. No matter what he was teaching, and no matter how bigor how small the numbers he was working with, a handful of chips fromthat bag was ALWAYS extremely close to what the calculations said. And grabbing 10 separate handfuls (counted and replaced in the bagbetween grabs) was ALWAYS RIGHT ON! This can be shown in any group of30 people. Even though there are a total of 366 possible birth dates,any somewhat random group of 30 people WILL ALMOST ALWAYS have twowith the same birth date.Intuitively it doesn't make sense, and that's what the naysayers arearguing. Yes, there are exceptions, and situations that don't matchthe calculations, but those exceptions and outlying situations aresmoothed out over time.Between Tim's and Dave's analysis, and Larry's Excel spreadsheet,along with a close-to-my-modeling-period ORER I can determine how manybox cars from what railroads I need to plan for, then adjust that asnecessary based on what kits are available to me.Is it perfect, and can I generate a single freight train thatperfectly matches a single freight train from Fraley? No, but thereis no way to come up with anything more accurate. Perfect? No, butbetter than the rest.Time stopped in November 1941Walter M. ClarkPullman, Washington, USA

You are missing the whole point.

National polls are based on a well defined disribution of people to call, not on a random selection.

Doing what is being done with the freight cars is really a random selection of a very limited number of samples, not a set of samples selected a prior to provide a good representation of the population being studied.

With a bag of two colors of items each of the same amount, yes it is right that you will always get about the same distribution no matter the sample size.

But we have a bag of a large number of railroads, not just two and they are not equally distributed. There is a basis for the distributon that is very complex as discussed already. So this is not the same problem.

On classic example of the birthdays add the year to the day and month and see what happens.

The more sets of data studied, the more likely we are to obtain a good answer.

Cyril Durrenberger

wmcclark1980 <walterclark@earthlink.net> wrote:
--- In STMFC@yahoogroups.com, "Bruce Smith" <smithbf@...> wrote:

Folks,

I grow weary of this. No one has answered Tony's challenge to provide
evidence that supports a "regional bias in boxcars" position. It seems
that any evidence that does not support this position is dismissed as
"absurd" or "biased" and opinion and hyperbole are used in lieu of data.
Many of the arguments that have been used about statistics are not based
in a sound understanding of the field. For example, deviations from the
mean are EXPECTED and do not invalidate the mean. Certainly,

modeling the

deviations would be.... deviant!

What I find truly remarkable is the wealth of DIFFERENT types of

data that

seem to support the national fleet as a starting place for a model
representation. The Charles data is flawed at best, yet it supports the
model. Wheel reports from different locations support the model...

-It ISN'T about individual trains (How often do we need to say this?)
-Individual trains may have very specific make-ups that vary

significantly

from the national averages.
-It is about a "fleet"
-That fleet will then fluctuate on the layout of the owner, providing
deviations from the mean... and modeling those deviations that almost
certainly occur in real life.
-It is a STARTING place - I have yet to hear anyone offer a different
starting place based on data other than their own opinion, and frankly,
I'd rather take my chances with the data ;^)
-The model DOES NOT preclude the use of oddballs, but it does INFORM the
modeler that these are oddballs.

Over and out
Bruce

Bruce Smith
Auburn, AL

Bruce, and all,

What we have is the problem that always occurs when using statistics.
Most people won't accept that a relatively small sample size can
forecast much larger populations with extreme accuracy. Just listen
to the arguments after a national poll is released, where they polled
only a few thousand people and used that data to project within just a
few percentage points of error what the entire Nation's population thinks.

Tim's and Dave's analysis is for an entire year. Not every train, and
not every train in a single month. Neither Tim nor anyone else who
accept Tim's and Dave's calculations are claiming that. It is the
total of all box cars over an entire year. This can (not must, we all
can do what we darn well please on our own model railroads) help me
decide which box car kits, and in what quantities, I need to
reasonably represent the possible activity in my month of November
1941 on my mythical short line that connects to the Southern Pacific
in northern California.

I had a professor in Graduate School who, in the Statistics class,
always brought a cloth drawstring bag with 100 red and 100 black poker
chips to class. No matter what he was teaching, and no matter how big
or how small the numbers he was working with, a handful of chips from
that bag was ALWAYS extremely close to what the calculations said.
And grabbing 10 separate handfuls (counted and replaced in the bag
between grabs) was ALWAYS RIGHT ON! This can be shown in any group of
30 people. Even though there are a total of 366 possible birth dates,
any somewhat random group of 30 people WILL ALMOST ALWAYS have two
with the same birth date.

Intuitively it doesn't make sense, and that's what the naysayers are
arguing. Yes, there are exceptions, and situations that don't match
the calculations, but those exceptions and outlying situations are
smoothed out over time.

Between Tim's and Dave's analysis, and Larry's Excel spreadsheet,
along with a close-to-my-modeling-period ORER I can determine how many
box cars from what railroads I need to plan for, then adjust that as
necessary based on what kits are available to me.

Is it perfect, and can I generate a single freight train that
perfectly matches a single freight train from Fraley? No, but there
is no way to come up with anything more accurate. Perfect? No, but
better than the rest.

Time stopped in November 1941
Walter M. Clark
Pullman, Washington, USA

Well, I got thoroughly chastised and sneered at by the resident statistical intellectuals because I questioned the state of undress, then along come some other people asking cogent questions about all this, and Mike asks logical and pertinent questions again. It still seems to me that this is all hogwash about the national statistics as they may apply to and be useful for predicting how many SP box cars will be seen on the Inside Gateway on April 23, 1956 (stayed within the magic time frame, Tim). Random doesn't mean predictable, except that you can predict that it will be random. Dictionary definition of random is "lacking aim or method; purposeless; haphazard." In statistics it means "of statistical sample selection in which all possible samples have equal probability of selection." Maybe we are applying random walk here, Tim, or random variables, where the variable's values are determined independently according to a probability distribution? Predictability means capable of being predicted, which means to say in advance what one believes will happen. Yes, you can predict that the percentages of box cars in a given freight train will be random, but you say they will be in a set percentage that does not vary. Round and round we go.

Quantum physicists know about this. A random event cannot be predicted or duplicated, it's a Surprise! Almost everything is predictable, but many outcomes are very difficult to predict because the variables that drive the outcome are either unknown or difficult to measure. That is precisely what we are dealing with here. While we can get the stats on the nationwide freight car fleet, and somehow come to the conclusion that this percentage holds true as the box cars travel around the nation, each following as if by magic its random predictable pattern and percentage, the fact is that there are a whole host of variables that drive the outcome and we either don't know them all or we don't have enough information to do anything with them.

This is like playing a video game, it exercises the mind and the keeps one's juices flowing, but in the end it is virtually meaningless and not necessarily a good way to spend one's time, especially when one realizes that there are so many models to build and run and so little time to do it in.

Bye bye,

John Stokes
Bellevue, Wa





To: STMFC@yahoogroups.comFrom: brockm@brevard.netDate: Mon, 18 Aug 2008 17:20:08 -0400Subject: Re: [STMFC] Re: freight car distribution - rejecting the equal distribution hypothesis.



We are getting to the point...not unlike discussions about color...where weare not seeing anything new. So...we are getting to the point where thethread will need to be terminated. Obviously some members are convincedregarding the Nelson/Gilbert theory's validity and others are not. Before the thread is terminated...until new data becomesavailable, I would appreciate seeing some clarification on a few points.Tony Thompson writes:"I personally think Tim's data say very clearly that theappearance of free-running cars like box cars DID follow,statistically, not absolutely with mathematical precision (Tim neversaid anything like that, so let's drop it now), the proportions in thenational car fleet."What does that mean? In the 1947 data covering Laramie to Green River, thetheory predicts [ I guess that's a good word ] 28 SP box cars. The actual number was 34...a 20% error. In the 1949 data, the theory predicts 52 SP box cars. The actual number was 136...or an error of 161%. Now...when I have mentioned this before, the answer was...nooo problem. This is statistics. OK...fine. No argument. Suppose that damned UP train with the 36 SP box cars was in Fraley's sample. Now the error would be 230%. What if 5 more such trains showed up? 576%. What if it were 1000%? Or 10000%? When does it become a problem...or are the violating SP numbers just thrown away? If the reply to this is that an error of 161% is OK, why bother with the individual national %? Just take the acceptable SP number of 136 = .01 (Y) (1325 box cars), Y = 10.2% and use it for all RR's? After all, SP's national % of 3.6% is fairly representative of all RR's except for PRR and NYC. Just add 5% more for them. I guarantee that the "error" between the actual national % for CGA, Rutland or FEC won't produce a worse error than using the actual SP national % does with the Fraley 1949 data. And, it will be a lot easier to do...don't have to look up anything. Of course, the same thing can be achieved by just acquiring the same number of cars for every RR except PRR and NYC. Get two of each of those. Then do the same thing 3 more times until you have 4 cars of every RR except PRR and NYC which you will have 8 of. I guarantee that will get you in the envelop of statistical success just as much as taking ther national % for each RR."The point is that it's theunderlying reality. Anyone who doesn't have better data than Tim's willjust have to get used to it."As far As I know, Tim's data was the 1947 Fraley and the Southern RR data on a train in Asheville. I gave Tim a copy of my Fraley [ 1949 ]. Did he use any other data of actual car reports?Mike Brock

Tim O'Connor writes:

"True, but I didn't include MP since I don't think it was much of an interchange
partner in that area -- it went south to KC, and so did the UP via the Kansas
Pacific. Wabash got very little traffic as well but UP may have been kinder to
them since they were a beneficial passenger train partner. I dunno. I think
C&NW always got the lion's share."

I've recently completed an article on the Overland Route's Fast Mail. I was surprised to find that CB&Q handled the mail between Omaha and Chicago while C&NW handled express...at least during the early 50's.

"> No. OTOH, I'm not seeing blocked eastbound cars in trains in Wyoming.

SP blocked cars for Chicago and other eastern points in Roseville. These
may show up as smaller blocks on the UP. Or maybe UP just shuffled them
around in Ogden for grins.

You ever think that Fraley may have just saved stuff he thought was NOT
typical? Like a big block of SP box cars? :-)"

Sorry. Nice try though.

I'm going to analyze my 1956 data and we'll see what that shows. Fortunately
for other members on the group, it will probably take me 5 yrs to do it.

I'm suspect many here are relieved to hear it."

I would say that your term "many" should be replaced by "most" or even "all".

Mike Brock

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

Folks,

I grow weary of this. No one has answered Tony's challenge to provide
evidence that supports a "regional bias in boxcars" position. It seems
that any evidence that does not support this position is dismissed as
"absurd" or "biased" and opinion and hyperbole are used in lieu of data.
Many of the arguments that have been used about statistics are not based
in a sound understanding of the field. For example, deviations from the
mean are EXPECTED and do not invalidate the mean. Certainly,

modeling the

deviations would be.... deviant!

What I find truly remarkable is the wealth of DIFFERENT types of

data that

seem to support the national fleet as a starting place for a model
representation. The Charles data is flawed at best, yet it supports the
model. Wheel reports from different locations support the model...

-It ISN'T about individual trains (How often do we need to say this?)
-Individual trains may have very specific make-ups that vary

significantly

from the national averages.
-It is about a "fleet"
-That fleet will then fluctuate on the layout of the owner, providing
deviations from the mean... and modeling those deviations that almost
certainly occur in real life.
-It is a STARTING place - I have yet to hear anyone offer a different
starting place based on data other than their own opinion, and frankly,
I'd rather take my chances with the data ;^)
-The model DOES NOT preclude the use of oddballs, but it does INFORM the
modeler that these are oddballs.

Over and out
Bruce

Bruce Smith
Auburn, AL

Bruce, and all,

What we have is the problem that always occurs when using statistics.
Most people won't accept that a relatively small sample size can
forecast much larger populations with extreme accuracy. Just listen
to the arguments after a national poll is released, where they polled
only a few thousand people and used that data to project within just a
few percentage points of error what the entire Nation's population thinks.

Tim's and Dave's analysis is for an entire year. Not every train, and
not every train in a single month. Neither Tim nor anyone else who
accept Tim's and Dave's calculations are claiming that. It is the
total of all box cars over an entire year. This can (not must, we all
can do what we darn well please on our own model railroads) help me
decide which box car kits, and in what quantities, I need to
reasonably represent the possible activity in my month of November
1941 on my mythical short line that connects to the Southern Pacific
in northern California.

I had a professor in Graduate School who, in the Statistics class,
always brought a cloth drawstring bag with 100 red and 100 black poker
chips to class. No matter what he was teaching, and no matter how big
or how small the numbers he was working with, a handful of chips from
that bag was ALWAYS extremely close to what the calculations said.
And grabbing 10 separate handfuls (counted and replaced in the bag
between grabs) was ALWAYS RIGHT ON! This can be shown in any group of
30 people. Even though there are a total of 366 possible birth dates,
any somewhat random group of 30 people WILL ALMOST ALWAYS have two
with the same birth date.

Intuitively it doesn't make sense, and that's what the naysayers are
arguing. Yes, there are exceptions, and situations that don't match
the calculations, but those exceptions and outlying situations are
smoothed out over time.

Between Tim's and Dave's analysis, and Larry's Excel spreadsheet,
along with a close-to-my-modeling-period ORER I can determine how many
box cars from what railroads I need to plan for, then adjust that as
necessary based on what kits are available to me.

Is it perfect, and can I generate a single freight train that
perfectly matches a single freight train from Fraley? No, but there
is no way to come up with anything more accurate. Perfect? No, but
better than the rest.

Time stopped in November 1941
Walter M. Clark
Pullman, Washington, USA

Cyril Durrenberger says:

"One of the main problems with all of this data analysis is that you are trying to take a very small sample size (really insignificant) and extrapolate to the whole fleet. This also assumes that the any sample you analyze is representative of the whole, which is not very likely."

Amen.

"1. Use data that is likely to reflect trains that would be hauling a representative sample. This would have to be through freights, not local freights.

2. Then compare train data set to roster data set for that year in a very easy to follow format (I suggest an excel spread sheet). This would be done for each train data set. Post the results in a data base for the group.

3. The groupings of small railroads would be necessary to simplify the process.

4. Then ask the question - does the train data compared to the roster data support that the cars are distributed according to the size of the national roster information."

You are saying to compare the real data to the national %? Good grief. Use real data? <G>. BTW, here's Tim's comments regarding such:

"In 1947, the ownership of foreign boxcars aggregated into eight ICC
Geographic Regions correlated pretty well with the percentage those
regions owned of the National Boxcar Fleet. In 1949, that correlation
was blown to hell."

I have no analysis on individual RR's other than SP from my 1949 Fraley. I seem to recall, however, being bitterly disappointed that NP box cars were much less represented than I had hoped...refuting my Fifth Rule of Frt Cars. Sort of like being hoisted on my own Fraley.

"If Tim has already done this, then present the data in such a format. What I have seen does not do this. Please do not tell us that is what it shows, show us the data then each person can determine if this applicable to their situation and then they can decide how to use the information."

As far as I know, Tim only grouped RR's by regions except for those closely associated to UP...SP, Milw, CB&Q, and C&NW as is shown in his message of Feb 3, 2006, which I republished here last week.

Mike Brock

Tim O'Connor says:

"In another post you said the PNW cutoff handled 12 trains a day. Using
your 35 trains a day average, that amounts to... over 33%.
Just trying to be consistent."

Tim, you have to read more carefully.

"> Yes, about 15% went on the cutoff through Granger to Pocatello, ID."

About 15% went the other way. 15% + 15% = 30%.

Mike Brock...got to be quick these days

Here's another note of discord.

Photos are also biased evidence. You can’t tell which are equipped box cars. To give one example of the bias introduced, consider auto parts cars. These were assigned to parts plants by each railroad in some kind of proportion to revenue received. There was no requirement to consider ownership in loading them. These cars would tend to appear on all railroads serving the auto industry in proportion to ownership. Another such fleet would be DF cars assigned to General foods and other large grocery manufacturers. This was less so in the 40’s than in the 60’s, use of equipped cars having been an increasing trend over several decades.

The auto parts traffic would tend to make lines between parts plants and assembly plants have observed box car marks closer to the national proportion than would other lines.


Malcolm Laughlin, Editor 617-489-4383
New England Rail Shipper Directories
19 Holden Road, Belmont, MA 02478

[Non-text portions of this message have been removed]

We are getting to the point...not unlike discussions about color...where we
are not seeing anything new. So...we are getting to the point where the
thread will need to be terminated. Obviously some members are convinced
regarding the Nelson/Gilbert theory's validity and others are not. Before the thread is terminated...until new data becomes
available, I would appreciate seeing some clarification on a few points.

Tony Thompson writes:

"I personally think Tim's data say very clearly that the
appearance of free-running cars like box cars DID follow,
statistically, not absolutely with mathematical precision (Tim never
said anything like that, so let's drop it now), the proportions in the
national car fleet."

What does that mean? In the 1947 data covering Laramie to Green River, the
theory predicts [ I guess that's a good word ] 28 SP box cars. The actual number was 34...a 20% error. In the 1949 data, the theory predicts 52 SP box cars. The actual number was 136...or an error of 161%. Now...when I have mentioned this before, the answer was...nooo problem. This is statistics. OK...fine. No argument. Suppose that damned UP train with the 36 SP box cars was in Fraley's sample. Now the error would be 230%. What if 5 more such trains showed up? 576%. What if it were 1000%? Or 10000%? When does it become a problem...or are the violating SP numbers just thrown away? If the reply to this is that an error of 161% is OK, why bother with the individual national %? Just take the acceptable SP number of 136 = .01 (Y) (1325 box cars), Y = 10.2% and use it for all RR's? After all, SP's national % of 3.6% is fairly representative of all RR's except for PRR and NYC. Just add 5% more for them. I guarantee that the "error" between the actual national % for CGA, Rutland or FEC won't produce a worse error than using the actual SP national % does with the Fraley 1949 data. And, it will be a lot easier to do...don't have to look up anything. Of course, the same thing can be achieved by just acquiring the same number of cars for every RR except PRR and NYC. Get two of each of those. Then do the same thing 3 more times until you have 4 cars of every RR except PRR and NYC which you will have 8 of. I guarantee that will get you in the envelop of statistical success just as much as taking ther national % for each RR.

"The point is that it's the
underlying reality. Anyone who doesn't have better data than Tim's will
just have to get used to it."

As far As I know, Tim's data was the 1947 Fraley and the Southern RR data on a train in Asheville. I gave Tim a copy of my Fraley [ 1949 ]. Did he use any other data of actual car reports?

Mike Brock

My point exactly. If you miss the nail, then you're not hitting it.

Well, that's not what my thumb thought. <g>

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

Dennis Storzek wrote:

This seems to be the hardest thing for modelers to grasp; that there was a significant portion of the rail shippers who DID NOT want fast service. They routed their rollers via the sloooow route, with many stops and poor connections. It gave them more time to sell the load, therefore they could start trying to sell it in the west, and work eastward as the car passed different diversion points. If the car zigzagged all over the country before they finally found a buyer in NYC, or Naples FL, they didn't care . . .

Were these diversions unlimited? PFE provided up to three of them for free, after that the shipper paid for each one. But that's perishables, and the "imperishables" may be different.

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

On Aug 18, 2008, at 2:27 PM, Malcolm Laughlin wrote:

-It is a STARTING place - I have yet to hear anyone offer a different

starting place based on data other than their own opinion, and frankly,
I'd rather take my chances with the data ;^)

Did you read my post about a distance based algorithm. Selecting the weighting factors is a guess, but if I had to bet, I'd say it will give me something closer to real distributions on railroads like FEC, B&M-MEC-BAR, WP and many ithers, if they were observable.

Malcolm,

I did and I can suggest one weighting factor that works for this approach... zero ;^) Since this approach appears to be based on your own speculation and you have yet to produce any supporting data set or refute the conclusions of the existing data sets, it doesn't seem like a particularly valid model. The word fantasy comes to mind.

-The model DOES NOT preclude the use of oddballs, but it does INFORM the

modeler that these are oddballs.

But it's indication of oddballs can be very wrong.

Huh? I don't understand what you are trying to say. Are you saying that boxcars predicted by the model as oddballs aren't? Any DATA for that? Or is this just speculation based on gut feelings?

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

My point exactly. If you miss the nail, then you're not hitting it.