Spinners for making destination selections for switch lists

[shortliner]

Video clip on using the "spinner system" for routing/placement of cars on Robert Todd's layout.

[dave1905]

You would get the same results over time if you just made a list and when down it in sequence.

 

I often wonder if you had two identical layouts and on one the owner just picked the destinations and on the other you used some "randomization" scheme, if over time there would be any significant difference between the operation.

[Nearholmer]

There's something about "exterior purpose" going on with randomisation system too, though.

 

It feels right that, as the operator of a railway, you are told by your customers, Mr Chance, Mr Risk, The Spin Brothers etc, where to send their goods, rather than "just making it up" ......... Its so much more railway-like.

 

Kevin

[dave1905]

I understand the desire to not feel "in control" and to not know what is going to happen before it does.

 

The reality is that much of the "randomness" of railroad movements is because the patterns aren't apparent or the pattern has enough variation that it doesn't appear to be a pattern.  Railroad local jobs tend to work at the same time, customers tend to use the same commodities inbound and ship the same commodities outbound.  They purchase supplies from a limited number of vendors and ship to a limited number of customers.  Many things that are "random" to the modeler aren't random to the railroad. 

 

There was a paper mill in Arkansas, they used 50 ft plain boxcars.  The pool of cars were made of IPD (short line) cars, RBOX and cars from the two railroads that served the plant.  From the customer's and railroad's perspective all of those cars were equal, they were the same car.  It didn't matter which of the cars was used.  Modelers see a "random" rainbow of cars being spotted, the railroad sees a set block of 5 boxcars being spotted each day.  Another paper mill had a wood chip train that operated on about a 3 day cycle.  To the causual user the train showed up at different days of the week each week, to the railroad it was a loading every third day.

 

If somebody wants to "randomize" the movements, have at it, I understsnd the deisre.  But "randomization" isn't required or in some cases, less prototypical.  YMMV

[Talltim]

I'm unconvinced by the realism of spinning a wheel. Real traffic flows aren't random, although they might be intermittent.

For outgoing goods a industry will have a contract with a particular car supplier, be it a RR, private leaser or themselves (i.e. they own the cars) so any empty cars going to them will tend not to be random (although they could be quite mixed)

For incoming goods there may seem to be more randomness to the casual viewer, but the cars will depend on who they are receiving from, and most industries will only have a few suppliers.

I've been watching videos recently to work out what cars were common in the train passing through Lawrence. I've noticed a lot of cars come in rafts, e.g. 3 BAR 50' boxcars together, 2 Railbox boxcars together, 5 leaaser gons together. I don;t know for certain, but I'm pretty sure that each raft will be going together to a particular industry.

[BG John]

Isn't it a question of whether you're operating the layout to achieve realism, or for the enjoyment and challenge of a shunting puzzle?

[Talltim]

How about the enjoyment and challenge of a realistic shunting puzzle?

[Nearholmer]

I can imagine that, on those vast US layouts that represent great swathes of railroad, including traffic origins and destinations, and are operated by about ten people at once, then not much randomness is needed to avoid boring repetition - it can be very like a real railway - but on a shunting plank???

 

Realistic micro-yard operation would probably involve spotting a couple of cars a day, or fewer, in pretty routine ways to serve particular traffic flows, with a "special" about once every three months, and would quickly become seriously boring without randomisation.

 

K

[Rivercider]

I have always been facinated by shunting and freight trains in general.

It is interesting to see how different approaches are taken to recreating shunting in model form.

 

Would I be right to assume that traffic flows and shunting in a North American yard complex would be somewhat different

from a British yard complex? By my time on BR starting in 1977 there were many less active private sidings and goods yards,

so there were few places where a yard pilot might be berthing a single wagon or two at each of a dozen locations.

 

I understand operating sessions need to be enjoyable, but a completely random set of shunting instructions would not seem right to me.

 

On BR I worked for many years as a TOPS clerk and often did enquiries to follow freight traffic flows and it was interesting to see how they changed.

For example firm A might regularly dispatch 3 vehicles a day to depot B so you would expect depot B to normally receive 3 vehicles a day.

If the journey involved remarshalling at several points it may be that a connection is missed somewhere along way so depot B receives no vehicles on day 2, but then 6 turn up on day 3. Or one of the batch of three may get red-carded and be delayed for several days.

I dont use dice to decide what runs on my shunting layout, but if I did I would work out a system that added variation to a steady pattern of arrivals.

So instead of wagons arriving 3,3,3,3,3,3 a more random pattern might be 3,3,2,4,3,0,6 etc.

 

I did like the idea of randomising the release of vehicles or re-spotting them if the vehicle has not been unloaded or perhaps awaiting repair before moving,

 

cheers

[BG John]

I think the spinner idea takes away much of the unrealistic randomness, because the cars only get allocated to industries that use them, rather than going to anywhere on the layout. With a simple inglenook, any car/wagon needs to go anywhere to make it work, so may end up in a place where it wouldn't go in real life. I've got a set of cards to shuffle for mine, and the whole puzzle is about moving them into particular positions, then dealing again and moving them all somewhere else. Hardly like operating a real railway though.

[Glorious NSE]

On BR I worked for many years as a TOPS clerk and often did enquiries to follow freight traffic flows and it was interesting to see how they changed.

For example firm A might regularly dispatch 3 vehicles a day to depot B so you would expect depot B to normally receive 3 vehicles a day.

If the journey involved remarshalling at several points it may be that a connection is missed somewhere along way so depot B receives no vehicles on day 2, but then 6 turn up on day 3. Or one of the batch of three may get red-carded and be delayed for several days.

I dont use dice to decide what runs on my shunting layout, but if I did I would work out a system that added variation to a steady pattern of arrivals.

So instead of wagons arriving 3,3,3,3,3,3 a more random pattern might be 3,3,2,4,3,0,6 etc.

And that ^ is where the random factor is good - no matter how good your railroad is at planning that will happen to some extent, trains will get delayed or cancelled, connections missed, equipment will have faults, and an utterly perfect pattern will not exist.

 

(Not to mention curved balls like CSX running their railway to a week containing 6x 28 hour days as well, which must be fun for making regular connections to railways that run with 7 days in a week!)

 

With a big layout (or the real thing) you don't have to plan for that randomness, it just happens. With a small layout I do think you need to generate it externally...

[dave1905]

So instead of wagons arriving 3,3,3,3,3,3 a more random pattern might be 3,3,2,4,3,0,6 etc.

 

 

And that is the difference between "random" and "variation".  Random means the events are unrelated and all possibilities have an equal chance.  Variation means that the operation tends to operate between some upper and lower bound and they may be connected.

 

To operate prototypically the trick is how to add/simulate variation without being totally random.