I've had the game intermittently working a few times only to be fooled into thinking it's fixed, but last night I replaced the pre-driver transistor array U1 (its a CA3081 I believe?) with a new one (socketed it of course) and so far so good. Played about 10-20 games and haven't seen a problem pop up.
Unfortunately replacing a 16 pin DIP IC isn't everyone's cup of tea so I don't know how much this will cheer you up. On something like a Sys 11 it's way easier to troubleshoot and replace a pre-driver.
Quoted from CUJO:I am gonna check the trace to the J5 connector first to make sure its not a bad solder joint or broken trace issue.
Update: Looks like the center pin on Q13 makes it fine to J5-12 Pin. Could it just be a bad connector at Position 12?
Good job on troubleshooting so far identifying that the coil and playfield circuitry appears to be adequate. It sounds to me like you certainly have a board problem which on the easy end can simply be some solder reflowing, or on the hard end could require some IC's to be replaced. Wouldn't hurt to repin any connectors involved with an issue like this (or all of them if the parts are on hand) but at a minimum I would reflow the headers and replace Q13 to be 100% sure nothing internal is an issue. Check the other two legs of the transistor to make sure they are making connections to the end of their circuits as well.
One tactic I use for older boards (especially with Williams awesome 40 pin connector) is physically (carefully!) flexing and holding the driver board flexed in different ways during a coil/switch/light test to see if functionality comes back. Unfortunately if you don't get functionality back it really doesn't mean anything, but if it does come back it indicates a poor connection somewhere, either at a solder joint, a trace, or even inside an IC.
For Meteor I pressed in the center of the driver board right on U1 actually and while flexed in that state, the coil would always work. When relaxed it was rare if ever. Since I'd already replaced the driver twice, reflowed and checked every connection related to the circuit, I concluded it could in fact be the internal components in the old IC chips acting up (especially since my E target was starting to fail too, also running thru U1). So I replaced U1 and all seems well right now. If it craps out again I shall report back.
For right now, if I were you , I'd try pressing reasonably firmly on U3 during a coil test, but I take no responsibility for you breaking stuff. If you don't want to do that, I'd be sure all related circuits are fine, and replace U3.