Math Errors Not Limited to NASA

March 15th, 2009

Posted by: admin

In what reminded me of the 1999 conversion error that led to the loss of the Mars Climate Orbiter, Scientific American’s 60 Second Science Blog noted a math error that contributed to the shuttering FutureGen, a clean-coal test-bed project, in 2008.  The Government Accountability Office released a report last week noting that the cost assessments for the project failed to consistently account for inflation.  This led to cost figures that appeared higher than they actually were, and the perceived cost overruns led to the cancellation of the project.

It’s entirely possible that FutureGen may not be able to deliver on what is promised.  But that gamble is better made when cost estimates, and other related math, are done properly.

6 Responses to “Math Errors Not Limited to NASA”

    1
  1. donmon Says:

    Several news stories give this item the same spin that you do: a silly math error causes an otherwise important project to be canceled.

    The House
    Press Release,
    however, states unequivocally that DoE manipulated the numbers deliberately in order to kill a project the leadership didn’t like:

    In an effort to kill the FutureGen project, top officials at the Department of Energy knowingly used inaccurate project cost figures and promoted an alternative plan that career staff repeatedly warned them would not work….

    The report from the Majority Staff of the Subcommittee on Investigations and Oversight of the Committee on Science and Technology also states

    It is difficult to believe that anyone working at the top levels of DOE or the White House … did not know the difference between “constant” and “as spent” dollars

    The cancellation also came about a month after Illinois was chosen over Texas as the site for the demonstration project.

    Don Monroe

  2. 2
  3. David Bruggeman Says:

    I tend to discount Congressional press releases – from both parties – as they often read like attack ads than fodder for news items. I defer to the GAO report, which doesn’t share Rep. Gordon’s apparent interest in finding hidden motives behind every screw-up.

    While I won’t dismiss the possibility you suggest, I think it unlikely. While we don’t expect agencies to make stupid mistakes like NASA and its contractors mixing up Metric and Standard units, they do. There is no uniform treatment of cost figures, in the sense that there is a rule to use constant dollars from a specific year, correcting for inflation.

    If the Bush Administration really didn’t want to develop a clean-coal plant, they could easily have directed its closure, independent of rising costs. I’d be a bit more skeptical if this FutureGen project involved renewables.

  4. 3
  5. donmon Says:

    Thanks for explaining why you didn’t find the release from the Democrats on the subcommittee credible. I don’t know enough about government accounting practices to judge their statement that this was such an obvious apples-and-oranges comparison that the misdirection had to be intentional.

  6. 4
  7. David Bruggeman Says:

    The report, as I read it, only implied that the misdirection had to be intentional. And if you wish to rely on someone for assessment of government accounting practices, the folks at GAO are the better source.

    Back when it was a minority, the Democrats on the House Science Committee (or at least the ranking member, now-Chairman Gordon) were trying to make political points out of an offshoring report developed by the Technology Administration. As Roger described back then, the hunt for a smoking gun against the opposition was pretty empty. While there’s more here in terms of bad management, my tendency is to assume incompetence before malicious intent.

  8. 5
  9. Oliver Says:

    This is a minor point, but important in such contexts. The reason that MCO was lost was not the error in units (though that was in some sence the ultimate cause). It was IMO bad management. The error led to navigation being off whack throughout the fliught, with the observed trajectories not converging on the predicted trajectories after every course correction. This should have shown pepole that something was wrong, and I believe people did take note — but the worry never reached the intensity/level needed to track it back to the fact that the probe was delivering pounds not newtons.

    The point of making this distinction is that you can’t eliminate errors — but you can seek to develop a management structures and ethos which recognises and catches them early. It was the lack of that structure and ethos, I believe, that doomed MCO.

  10. 6
  11. David Bruggeman Says:

    Oliver, I don’t think the point is minor, and appreciate you emphasizing it. Consider the conversion error (and the corresponding calculations on FutureGen) as a symptom more than a cause.