Ed:
I am a certified aerospace cost estimator. Color me skeptical.
After 10,000 years of building 'em, nobody knows how to accurately estimate
the cost of a custom house. Indeed, the only reason housing tracts tend to
come in close to the overall project estimate is that they first build 3-4
"model homes" and then apply production learning effects to those actuals.
SpaceX's development costs appear to have been about 1/4 the "commercial"
government estimate for the initial version of Falcon 9; I've no problem w/
that; it is similar to the about 1/3 we see in other "commercial"
government estimates. Maybe it is so; maybe others can duplicate it...I'll
remain skeptical.
The future cost of stuff has--in my experience--to many variable effecting
the answer to ever allow accurate estimation until you have firm bids on
all the piece parts; even then the chances are high something unexpected
will happen, good or bad.
And then there's the reliability conundrum: did you unwittingly impose
greater risk on your system by being aggressive about costs? I don't know
and I don't know anybody who does know....
Bill
On Sat, Nov 19, 2016 at 12:36 PM Ed LeBouthillier <codemonky@xxxxxxxxxxxxx>
wrote:
On Sat, 2016-11-19 at 17:17 +0000, William Claybaugh wrote:
Careful.
Development costs are irrelevant when they are written off, as SpaceX
has done (and as is the case in all government developments).
Sure, after the project is done, development is as you say.
My interest in bringing this whole issue up is the up-front view of
development cost (e.g. estimating scope and cost before you engage in
development).
The only factor determining their competitive position is their
production cost.
I agree.
Many of us on this list are, generally, engineers and not managers or
sales people. Our focuses tend to be the up-front estimation of cost,
scope, performance and things like that. So, in that way, I guess I too
am focused on that earlier part of projects, than the big picture
business realities.
I am also--from experience--extremely skeptical of top level cost
modeling of what might have been; such efforts never have anything
close to the fidelity of a serious project cost estimate...which are
themselves characteristically too low.
I guess that's what I'm trying to focus on: serious project cost
estimation (where "cost" includes scope and time). As I said in earlier
posts, there have been, historically, a number of failures (and a few
successes) in rocket development projects. I was wondering whether
those might be due to pre-assesssing a lower difficulty level than they
actually were and whether we might be able to educate ourselves to be
able to better assess scope and cost up front.
As I mentioned to one fellow, there's nothing wrong with taking on a
big project that you know you don't have resources to complete, as long
as you recognize that fact up front. It is often useful to test oneself
to see what is possible, or just to lay a foundation for others, or
just to accomplish as much as one can.
But, sometimes, we need to know how to think about, plan, and estimate
projects that we expect to accomplish. Understanding the methods and
tools that allow us to do that is useful.
Finally, there appears to be a discrepancy between the $300-400
million development cost number and the amount of cash the company
burnt--including the non-public funding--in the relevant timeframe;
not to mention all the money that has gone into further development
since these estimates were made many years ago.
I understand. My statement was not meant to be a definitive statement
of what I believe such values might be. I was trying to give an off-
the-cuff image of the possible differences. In general, all I was
saying is that it is possible to develop things (substantially?)
cheaper than the cost models suggest. That doesn't mean the cost models
are worthless; they help give a first-order, order of magnitude
estimate of scope that can be useful, I think.
