So, y'know, what's necessary to do one of these "studies&#

Amtrak Unlimited Discussion Forum

Help Support Amtrak Unlimited Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

neroden

Engineer
Joined
Feb 23, 2014
Messages
9,548
Location
Ithaca, NY
When I read the studies (performance improvement plan, proposals for new services, etc.) they don't look that fancy. Is there a cut and dried procedure for them? A standard method of estimating ridership, a standard method of estimating revenue, a standard method of estimating costs? I know that they just make wild guesses as to what the host railroads will charge since they always turn out to be wrong.

What I'm wondering is... a lot of these studies seem to cost absurd amounts of money considering what they are. Hundreds of thousands of dollars, or even millions. That's suitable for a full EIS or a full set of engineering diagrams or negotiated prices with cities and railroads and property owners, but not for these rough ballpark studies.

If I, for instance, provided free volunteer labor to study the financial status of a "two a day on the Water Level Route" proposal, how hard and how expensive would it really be? That's the problem for me at the moment with making such a pitch: I don't know where to look for a ridership or revenue model which would be accepted, or a cost model which would be accepted. But it can't be that hard, can it?
 
Good knowledge of Excel/PowerPoint, financial modeling skills, and access to a good database/Internet would seem to be the basic requirements.

The real trick may be having the business connections or sales pitch to get yourself properly paid by those parties handing out the cash. :)
 
When I read the studies (performance improvement plan, proposals for new services, etc.) they don't look that fancy. Is there a cut and dried procedure for them? A standard method of estimating ridership, a standard method of estimating revenue, a standard method of estimating costs? I know that they just make wild guesses as to what the host railroads will charge since they always turn out to be wrong.

What I'm wondering is... a lot of these studies seem to cost absurd amounts of money considering what they are. Hundreds of thousands of dollars, or even millions. That's suitable for a full EIS or a full set of engineering diagrams or negotiated prices with cities and railroads and property owners, but not for these rough ballpark studies.

If I, for instance, provided free volunteer labor to study the financial status of a "two a day on the Water Level Route" proposal, how hard and how expensive would it really be? That's the problem for me at the moment with making such a pitch: I don't know where to look for a ridership or revenue model which would be accepted, or a cost model which would be accepted. But it can't be that hard, can it?
If you're talking about the internal route proposals, service changes and studies, they are forecasts based upon the information available at the time of the study. They are not designed to be fancy and vary in scope depending on what is being researched. They are deep and detailed but remain forecast.

I wish I could show you one. Actually, I wish I could show them to the vast majority of the board. A lot of the posts/posters wouldn't have much to post about. BTW, who said they cost a lot of money? Where are you getting these figures from?
 
Last edited by a moderator:
Every time a *state* or *city* orders this sort of study, they get charged a hell of a lot for it! And theirs are no more detailed than Amtrak's.

Amtrak's internal costs for doing studies appear to be much lower. I to have to say, about the internal proposals, I know "they are forecasts based upon the information available at the time of the study."... the question is, where would one get the information and what forecasting models are considered acceptable? Because I certainly have all the skills to do such a study. And the interest to do it for free. What I do not have is access to the data *or* the preferred forecasting models. I already know that when dealing with governments you typically don't get listened to unless you use the model preferred by the people you're pitching to, or at least a variant of it (if there's really stupid stuff in their model which needs to be fixed).

It would really help if when one of these studies said "We expect that this action would lead to an increase of X riders", it explained WHY they expect that. What model are they using? What data were input into the model?

I mean, I know the basics of ridership modeling, but it's an art, not a science. You make choices about what model you use, and some work better in different times and places than others do. I could do a simple gravity model any day, or use the rough and well-tested rubric that going from 1-a-day to 2-a-day increases ridership by more than a factor of 2, but Amtrak is clearly using slighly fancier models.

And revenue modeling is even more of an art with even more choices in the model.

Cost modeling should be a science, but I'd love to know how Amtrak figures the costs for it.

There's a GIGO aspect to the modeling: a while back Seattle did a "study" on whether it would be cheaper to operate new electric trolleybuses or buy diesels. It claimed that diesels were cheaper. It was complete nonsense, and people quickly pointed out that they assumed the operating and maintenance costs of the OLD (1970s) trolleybuses, not of brand-new trolleybuses; while also assuming that diesel prices would never go up. After a public campaign, Seattle backed down and bought new trolleybuses, and yes, they are more cost-effective.

[Aside: I strongly suspect Amtrak of making mistakes like this occasionally (everyone does), but it's impossible to identify these mistakes if they don't show their work.]
 
Last edited by a moderator:
Back
Top