TCAS can get confused and confusing

If TCAS (traffic collision avoidance system) resolution advisories (RA) should be followed slavishly by pilots - and that is the official advice - some people have argued that the RA manoeuvre should be automated.

It seems their opinion - however logical it may seem - was premature.

It has been known for some time that certain combinations of closure vector, especially where one or both of the conflicting aircraft are in climbs or descents, can confuse the existing TCAS marque, and an inappropriate pilot reaction can make it worse.

Eurocontrol studies have now established that some of these weaknesses are embedded in the software, but that this can be overcome.

Until this is fixed however, Eurocontrol has issued a chilling piece of mathematics that describes the risk: 

"Aircraft equipped with TCAS II version 7.0 face a mid-air collision risk of   2.7x10^-8 per flight hour, corresponding to one collision every 3 years in European airspace. This exceeds the tolerable rate for catastrophic events related to equipment hazards by a factor of more than 25."

Gulp.

I would like to think that the number of times TCAS will get us out of simpler conflicts will exceed that factor. But would it? 

Eurocontrol says its new proposals would have prevented Uberlingen. But we still don't know what happened to the Gol 737-800 and the Embraer Legacy when they collided over Brazil. TCAS didn't prevent that, and it was a simple case of aeroplanes approaching each other on the same airway, same flight levels, but with reciprocal headings.