Sorry, that's my fault for typing while half asleep...
You don't have to cut the original wire. Just find a way to 'skip over'/jumper/bypass it, using a second wire connected to the same connection points. i.e. One end of a piece of test wire goes to the sensor, the other end goes to where the sensor wire enters the ECU. You'll just need to find a reliable way to make that connection, particularly at the ECU - because it's an awkward position to work in. (As I recall, it was difficult to even get a multimeter probe in there.)
Once you have your reliable connection at both ends, reset the computer, drive around for a while, see if the code returns. (You need to KNOW you have a reliable connection, not just held by someone by hand for example, creating uncertainty... "Damn, I don't know if it's throwing a code now because the connection let go for a second, or because the fault is something else!?)
The idea being, with a known good connection - if the code stops appearing - you now know there's a break, or short circuit, along the length of the sensor's wire.
You'll find a way to connect the test wire at the ECU that best works for you. For example, one way might be to make a sideways nick in the insulation with a razor blade, then spread the copper strands apart slightly with a thin jewelers screwdriver, and push a pin through the middle of the copper strands and out the other side and close it up again against the pin - so the pin is sticking out the same distance on both sides.
Then strip a good inch or two of insulation off the test wire, and twist that around the pin and original wire in a kind of figure eight if you get what I mean - to make a good connection. And electrical-tape it up nice and tight, so it can't disconnect or touch anything else it shouldn't. Because you might choose to do the test for a week of driving to be sure.
(You could try soldering the test wire to the pin first before pushing it through. But I've tried that a couple of times solder refused to stick to the pin for me.)
Or maybe remove that right-angle plastic cover off the end of the ECU plug - the one that's held together with two screws. You might then see a way to reliably jam a bent paperclip, etc. in. Again, whatever you find works best for you. These are just ideas.. It all depends what you have to hand.
With the MAF, so you're saying you can definitely see the spider-web-thin nichrome wire is unbroken? (See the picture bgpzfm142 posted here: https://ozvolvo.org/discussion/7351/no-start#latest.)
Re: Your water pump question... One type I've seen has several right-angle 'blades', and the other has two discs a few mm apart. I removed the first type of water pump from a 240 once that had all those blades rusted away. So it was spinning great but barely pumping any water. (That's one way a water pump can appear to be working but not be.) But as you said, if the temp gauge is ok, then it's ok.
An electrical fault is possibly your easiest thing to eliminate as a cause. But do check the MAF wire... I have no way of checking all the connections inside the ECU/EZK, but I keep thinking back to my no-start problem... The O2 sensor light was on, I replaced it, the light was still on. It was because the MAF wire was open circuit. i.e. Sometimes two or more things work together - producing a fault indicating A is busted, when it's actually B or C.
