Here’s a link to the actual document:
Some from Toyota claimed that Dr. Gilbert’s test doesn’t match the real world. Looks to me that his analysis and simulation match the real world very well.
The scenario is very plausible. Basically, there are two signal wires from the pedal to the computer, and at least one power wire to the pedal.
Dr Gilbert found that if the two signal wires short together, the computer doesn’t notice. The car may not behave quite right, but the driver likely wouldn’t notice.
The wires don’t have to be perfectly shorted: some resistance in between them gives the same result.
Now, while those wires are partially connected together, if somehow the power lead to the pedal briefly touches one of them, the computer snaps the throttle wide open. That means, a problem with the wiring harness could cause the motor to run at full throttle.
The computer does not set any error codes when this happens.
Shazam: A wiring harness problem can send the car into runaway with no trouble code set. And, since wiring problems are often intermittent, it can “fix itself” for some unknown period of time.
Dr. Gilbert seems to have very good technical credentials. They are summarized in the PDF linked above.