my conclusions to certifying a track that is a road race

If the track portion is just a small part of the full length of a road race I think the calibrated bicycle method is accurate enough. But I wonder about my ability to stay 20 or 30cm from the inside lane line around the turns of a track. Add to that to the inaccuracy of calibrating on pavement and then measuring on what in most cases will be a rubberized track surface.
Why not require a measurement by a calibrated measuring wheel? This method has been shown to be accurate by both Pete and myself.
I am also wondering if we should also stipulate that if the measurement comes out to close to 400m or 440y, that it should just be certified as one of those standard distances. Do we really want to certify a track at 400.3m? Especially if that measurement required a bike rider to maintain a constant circular path for 180 degrees?

If a course is to be multiple loops on a track surface, so be it. It is not a certification of the track, but a road race on an oval.

Calibrate on the surface, not pavement.

Forget about the requirements for tracks - curb or no-curb. If there is a curb, it is treated as a curb on a street. If there is no curb, measure as you would if you were restricted to a road shoulder - "Cannot cross painted line on curves". Disregard that it is a track. We are measuring a course on an oval.

I think the biggest point for discussion is, do we include SCPF? The bike will wobble. If you can tape it, and the curves are uniform, you can use the tape-and-math method. If not uniform, use a bike or calibrated measuring wheel. But, clearly state if SCPF is used, or is not used, on the map and certificate.

Consider this situation as very similar to a loop for racewalking. The turnaround is a larger arc, but you are following a relatively straight line between the turnaround arcs. We use SCPF for racewalk courses. Should we when measuring on a striped oval? Would we if we measured on a go-kart track that was configured on an oval, just like a track?

How about leaving track on the certificate but with a qualifier (fine print?) clarifying that this is not a full certification for the track but only a measurement suitable for distance races not involving use of lanes?

This might allow for distance events like 10 miles, 5K, or 10K on the track, one hour runs, etc., to get an assurance that the track is accurate, without infringing on the official track certification procedures.

I think something like Bob Baumel's method is appropriate for this, but I would always advocate for a calibrated bike or wheel measurement as a check on this. (Don't try to measure 20 or 30 cm from the inside line, just run the wheel or bike on the inside line, then do the math to find the official measured path.) The adjustment is + 1.26 m for 20 cm offset, and + 1.88 m for 30 cm offset.

My 2 cents.