[X] My Assistant

[SFJCody]

Something I have not posted online before...
Back in 2003, as four spacecraft approached Mars, I wrote down on a piece of paper my guess (based on nothing more than public information & gut feeling) at what each craft's chance of success (either at landing or orbital insertion) might be.
My guesses were:

Nozomi: 15%
Beagle 2: 20%
MER A: 60%
MER B: 60%
Mars Express: 85%

In 2005 I guessed that MRO had a 90% chance of success

Now, in 2008, I'm going to put a figure on Phoenix. That figure is:

55%


What do you think? Too low? Too high?

[tanjent]

This is what markets are good for. We can all exchange our best guesses, and we all feel a great emotional stake in the outcome. But a pecuniary stake tends to weed out the casual guessers and weight those remaining according to their confidence in a) the quality of their information and their having the experience and judgment to apply it effectively.

An outfit called Intrade has a similar market for the probability of the Google Lunar X prize being won by 2012.

http://www.intrade.com

It's probably not the only one. I think Lloyd's used to quote odds on a whole range of possible binary events.

It's too late to do anything about Phoenix over the weekend, but maybe we should encourage the Intrade people to float a similar issue based on the successful landing of MSL. If it seems crass to think about monetary payoffs when the really important value is something like "raising the consciousness of humankind" then pledge your future proceeds to the Planetary Society!

I think I'll be on the sidelines though, because my expertise is pretty much on the level of keeping my fingers crossed. Accordingly, if pressed, my best guess would have to be 50-50 with no apologies to either side. Anyway, it's just our left brains that feel compelled to quote odds about the probabilities. There is complete unanimity with respect to our hopes. 100% probability on that.

tanjent

[imipak]

An outfit called Intrade has a similar market for the probability of the Google Lunar X prize being won by 2012.

There's also LongBets, which I happen to have bookmarked for reasons we don't talk about here

Regarding the supposedly improved odds over MPL due to the failure modes having been studied and removed - alas it isn't quite that simple. Suppose you change an element which has a 5% chance of a mission-killing failure, in order to fix that. Now, your changes themselves must be analysed for the the new failure modes they introduce - and of course these components don't have to work in isolation, they must work in the particular environment - which we can only approximate and simulate before launch - and whilst many other components are busy doing their thing. The worst case is that the fixes end up making failure more, not less, likely.

(Edit: ...and whilst current and next-gen fly-by-wire military hardware are incredible things,
(i) a lot more is spent on them than on Mars probes!
(ii) they can be tested in the real flight environment, and
(iii) http://www.youtube.com/watch?v=faB5bIdksi8

I'd recommend RISKS Digest for anyone interested in real-world systems failures of all sorts - some tragic, some expensive, some amusing.

[mcaplinger]

The worst case is that the fixes end up making failure more, not less, likely.

True, but in the case of MPL, the landing-leg deployment switch bug was high-probability and its fix is obviously correct.

AFAIK there were no changes to the control algorithms. There may have been some made to the radar processing. http://www.nasa.gov/mission_pages/phoenix/...x-20080508.html