They didn't prove that all motorcycle-like things buildable from that car with those tools are useless. These were the possible outcomes from the Mythbusters test:
1. They build it and it works. That would prove that building a useful MLTFTC (motorcycle-like thing from that car) was possible, but would not prove that Leray had done so.
2. They are unable to find a way to build an MLTFTC. This does not prove it cannot be done, but they are smart and inventive builders and had accounts and photos of Leray's alleged MLTFTC to go on, so if they could not figure it out it would cast very strong doubts on Leray's story.
3. The result they actually got. They are able to build an MLTFTC, but it doesn't work well enough to ride.
That is the least satisfying possible result. It tells us that MLTFTCs are possible, but doesn't tell us if useful ones are possible. That theirs was not useful fails to rule out that others might be useful because two-wheeled vehicles are quite sensitive to geometry. It has not hard to have two such vehicles that look very similar, but one is rideable and one is not due to a few degrees difference in part of the geometry.
Why not build a tricycle instead? Much more stable, not much more complicated, and the parts shouldn't be an issue since you'll have four wheels and two axles lying around already.
AFAIK the premise of the myth is that the first two tires broke. If only one tire broke I think you can plausibly drive the car without any modifications by shifting weight away from the flat tire.
1. They build it and it works. That would prove that building a useful MLTFTC (motorcycle-like thing from that car) was possible, but would not prove that Leray had done so.
2. They are unable to find a way to build an MLTFTC. This does not prove it cannot be done, but they are smart and inventive builders and had accounts and photos of Leray's alleged MLTFTC to go on, so if they could not figure it out it would cast very strong doubts on Leray's story.
3. The result they actually got. They are able to build an MLTFTC, but it doesn't work well enough to ride.
That is the least satisfying possible result. It tells us that MLTFTCs are possible, but doesn't tell us if useful ones are possible. That theirs was not useful fails to rule out that others might be useful because two-wheeled vehicles are quite sensitive to geometry. It has not hard to have two such vehicles that look very similar, but one is rideable and one is not due to a few degrees difference in part of the geometry.