I did a speed-read through the posts. One request was to summarize. I'd like to do that, but I feel there are several conflicting statements, coming from who I consider reliable sources.
I do understand that the torsion value changes when the engine warms up. But the cam timing is directly affected mechanically and mechanically moving the cam sprocket does make remarkable differences in fuel economy. Changes as small as .5 can make 4mpg difference in fuel economy .
Normally, the torsion value seems to be 0.0 to -1.5 in most cars, but there are exceptions. We have seen fuel economy's in the 50's with cars set anywhere from +4.5 to -5.7. Something must account for the variation.
There are a large number of variables that can affect the timing. For example, one recently removed BRM crankshaft sprocket could be moved about 1 degree in either direction. Certainly, if all variables fell one direction or the other, torsion value will be skewed in that direction. I.E., if you moved the crank sprocket in a positive direction, it would require a negative torsion value to compensate and still get the same results.
I don't think using an idle-speed run to determine liter per hour (Block 15) is a useful number. If anything, running at a designated rpm at temperature through a flat course would be a better measure. I do think that measurement could be useful.
Although checking the setting before tear down MIGHT be useless, it would be somewhat valuable if the engine were performing reasonably well. Particularly if the measurement received is considerably out-of-line with what would be considered 'normal'. If the usual 0.0 or -1.5 variables don't work and the engine is running poorly, don't forget where the starting point was. My feeling would be that if an engine is running good at -5.7, you change cams or timing belt, etc., reset the torsion value to 0.0 and lose performance, start gravitating your torsion negative toward -5.7 until you find the 'sweet spot'.
The good thing... it's not that hard to fiddle with. Also, if you set the belt on the cam sprocket correctly and the 0.0 is somewhere in the middle of the adjustment, you can't reach a point that you would get interference contact with valve and pistons. It is somewhat 'idiot proof' to that point.