Once more with feeling

Sonar vessels searching for Malaysia Airlines Flight MH370 in the southern Indian Ocean may have missed the jet, the ATSB’s Chief Commissioner Martin Dolan has told News Online. he went on to point out the uncertainties involved, the difficulty of terrain that could mask the signature of wreckage and that therefore problematic areas would need to be re-surveyed. Despite all that the Commissioner was confident that the wreckage site would be found by June. Me I’m not so sure.

The commissioners remarks actually points to the continued failure to apply Bayesian search techniques to planning the search operation. You see in a Bayesian search having calculated the probability that the wreck site might lie within a particular cell in the search zone (which has been done) you also estimate the probability of finding it if it were within a particular search cell, based on dept, currents and bottom terrain. Then you multiply the two probabilities together to get the cumulative likelihood of finding the wreck site in a particular cell. Finally, you start looking in the sites where that combined likelihood is the greatest and work your way down the probabilities, updating the remaining cells probabilities with your search results as you go.  What you don’t do is go looking in the cells were it’s most likely that the aircraft ended up.

The advantages of Bayesian searches are that all the information available is used coherently and you can produce estimates of the cost for a given success probability at any point, but apparently the ATSB is happy to forgo such advantages despite the proven track record of this technique, you know like finding H Bombs the USAF accidentally dropped in the bay of Naples, or  the wreck of AF447 after three prior unsuccessful searches. If you think all that sounds a little snarky, you’d be right, I’m pretty sure we’re going to get to June and be none the wiser as to what happened to MH370.

### 2 responses to MH370 Update: We may have missed something

1. 

Greetings, and, as always, interesting observations . . .
However, I wonder if there may be an oversight in the analysis, in that the determination of the searches to be conducted DID consider the Bayesian aspects, via “Bayesian methods in the search for MH370”, per DST, link: https://www.atsb.gov.au/media/5733804/Bayesian_Methods_MH370_Search_3Dec2015.pdf
Or maybe the issue you raise is that the track plot was determined using Bayesian methods, whereas the actual planning and conduct of the search itself may not have been?

• 
Matthew Squair 03/02/2016 at 6:56 pm

That’s exactly, and somewhat frustratingly, the case. Hats off to the DSTO boffins for the track study, but no score for applying Bayesian techniques to planning the search.