Most people see JMRI Operations and othr layout operation planning methodes (software assisted or manual) as a way to bring life to their layout. On the other hand, except a few extremely dedicated individuals, it is rarely used in the previous steps to get big picture of the prototype.
In my case, I'm a little bit late to the game with a layout almost set in stone. However, the last few week fine analysis of pictures and train movements on CFC helped me to create a JMRI prototypical freight traffic model. And it is an interesting because it better helps to understand how things moved over the rail and where some very specific task were performed.
After running more than 1000 simulations - which amount to three years of operation on Murray Bay - it became clear that replacing Montmorency with Wieland was a viable option. The traffic is stronger, car movements more frequent and a little bit more complex due to shuffling of cars between Clermont yard and there.
Villeneuve also took another twist by the addition of another prototypical movement that we never implemented on the layout: storing extra cars on a siding. CFC had the habit to leave cars there so they could pick them up if required. It had another layer of purpose and sense to this particular location. MoW material (ballast cars and snow plow) could also been spotted there before moving up to Clermont.
All in all, I'm glad to have put almost 4 complete days working out this JMRI model. Things are clearer and it will be easier to make choices later.It also feels there is a real outside world and dispatchers behind the trains instead of whimsically choosing cars out of the drawers. I'm looking forward in the next week when track will be back in action in Wieland to test these scenarios and further refine them.
While JMRI isn't a panacea and can't be the focus of layout planning, I still consider that for someone trying to model an entire subdivision, it can start to be another tool in the toolbox to determine how each location plays a part in a railway's daily life. Having to select only a very few elements, it is not bad to have some more factual data to help creating a realistic and engaging experience for operators.