Statistically speaking, how reliable are Lycoming and Continental general aviation piston engines? What I'd really like to either find or build is a table showing the probability of experiencing engine failure on a properly-maintained piston engine for most of the common models of engine. Is the data out there?
Short (pilot's) Answer
"Reliable enough for you to fly behind one, flaky enough that you should always expect them to fail at the least convenient moment."
Short (Mechanic's) Answer
"More reliable than the pilot!"
You can't really know the failure rate, but you can know the accident rate attributable to engine failures.
The NTSB looked at this back in 1972 and Embry-Riddle still has the report available. It's very interesting reading, and goes into great detail on the specific causes of engine failure.
Totaling up everything they classified as "engine failure" gave a failure-related-accident rate of just under 18 per 100,000 hours (i.e. "Assuming proper maintenance you could theoretically fly 5,000 hours in piston GA aircraft before you'd have an engine-failure-related accident.")
There may be a more current NTSB study on this, I honestly didn't look too hard as I expect the numbers to be broadly similar today.
The 24th Nall report shows 61 "powerplant" accidents (all engine types, and including the propeller and associated systems) - the number is not broken down per 100,000 hours but doing a little extrapolation from their other figures it works out to about 0.4 accidents per 100,000 flight hours (or "Assuming proper maintenance you could theoretically fly 200,000 hours in piston GA aircraft before you'd have an engine-failure-related accident.")
The mechanic's answer of "More reliable than the pilot!" comes into play when you look at the pilot-related accident numbers: In the 1972 NTSB report I cited engine failures made up 13% of the total accident figure. In the 24th Nall report all mechanical accidents were 14% of non-commercial fixed-wing accidents (17% of commercial accidents) - that includes a number of other systems (e.g. landing gear) which are unrelated to the engine.
That data may not be directly available but you may be able to piece it together. This article has some of the only numbers I have ever seen (although they may not be 100% accurate). Here are some random facts on engine failures that have no sources quoted so take it for what you will. There is some interesting info on why the engines fail here and here. Your best bet for piecing the information together is to look over the actual NTSB reports. For this you are going to want to pull them and have a way to comprehensively sort and filter them (which is really a computer science question not an aviation question). Since the plane is listed you should be able to extrapolate the engine from the report that way. You can also find some discussion on the matter here as well.
The issue with building a chart like this is there are many many variables and generally speaking (at least from the NTSB reports I have read) engine failures are often linked to improper maintenance or improper use more than they are related to an engine component simply breaking. In many cases it only takes one failure for an AD to be issued and the problem to be solved. Engines like the more common Lycomings and Continentals you find in GA planes today have been around for some time and proven to be reliable and simple. Mechanics know them through and through and there are years of validation to support TBO numbers and things like that.
This thread seems to provide real numbers pulled from NTSB reports (at least thats what the guy claims). Its an interesting read if the numbers are correct. The fact remains that engine failures are rare (in the grand scheme of things) but do occur and should always be prepared for.