Real stats on engine failures?

David Megginson wrote
I'm curious where the statistics are that show that most pilots cannot
handle an AI failure in IMC. This FAA report

http://www1.faa.gov/fsdo/orl/files/advcir/P874052.TXT

states that vacuum failures are a factor in an average of 2 accidents per
year, and that there is an average of one vacuum-related accident for every
40,000 to 50,000 GA IFR flight plans filed. That doesn't tell us much,
though, since we don't know how many non-fatal vacuum failures occurred
during those flights.

I have about 700 hours flying behind a dry pump, and one catastrophic
failure. I also have about 1400 hours flying planes with gyros (some
of my time is in gliders and no-gyro taildraggers) and at least three
gyro failures. I have to believe that vacuum or gyro failure occurs
AT LEAST once every 1000 hours.

Assuming that the average GA IFR flight plan leads to 30 minutes of
IMC (I know a lot of them are filed procedurally so I'm being
pessimistic) that still sounds like 1 accident in 20,000 hours. So it
sounds to me like 95%+ of the pilots who experience vacuum or gyro
failure are handling it without an accident.

From what I've seen of GA IFR pilots, at most 10% are getting
recurrent training in partial panel operations to PTS standards.

Michael