"I might be convinced to take a hard look at the ECM. "
Not since the OBD revolution (which is actually has little to do with).
That happened right around the advent of the move from through-hole to surface mount components, and the rapid influx of dirt-cheap custom chips and common components. Rarely do ECU's fail since, oh, 1995 or so.
Only one I've ever had fail on me in any way was my '91 shadow "racing" car, and that had suffered electrical mishaps in the past, one unknown short when my then-fiancee borrowed it in an emergency, right after the weight-stripping that removed the dashboard. I hadn't insulated the ignition switch yet , which was a modular single component, and her nice metal keychain made a circuit from ground at the tumbler to an still-unknown circuit with enough amperage behind it to "try to kill her" in her words, and she's technical and not prone to exaggeration. She said it blasted sparks all over the interior of the car and it burnt off a corner of the offending key. A few months after that I had to drive it on a regular basis after totalling my regular car, and ended up field-replacing some of the fusible links as the car was simply dying at random, no codes, but I caught it once when I had a voltmeter handy and figured it out. Then the FPR relay quit, which was on one of the circuits the bad link was on, and a couple more months after that the car simply just up and quit. THAT was the ECU. I pulled it apart and got another off ebay for $50; Chrysler's potting of that era was pretty much transparent, and thanks to the turbo dodge groups' more technical members, someone had an engineering schematic and the one torched resistor I found was on the Injector active circuit with the FRP relay activation hung off it on a filter.
After that the car ran great, but I had little use for it anymore so sold it to a guy who (supposedly) made a circle track racer of it. Wait, wow, what a tangent.
The most common causes of "old" ECU death were corrosion and bad grounds. Corrosion is obvious, but bad grounds can cause elevated current in wierd places and even modern ECUs may have trouble with that. Sucks to be an ECU when a rusty ground causes the alternator to try and energize itself through the return for the throttle position sensor or whatever else has a connection and a better ground.
That, and a handful of TFI modules for my '84 Tbird were the only "bad electronics packages" issues I've had yet
'
