So do you think he actually did the math...or the trial and error method...or a combination of the two.
As long as it's "I think" and not "I know" ....
I think it came from snowmobiles, where people would hit an unfrozen creek at speed (maybe by accident the first time) and realize that if you kept your speed up, the sled would keep on going on top of the water, like a flat pebble skipping along, and wouldn't sink until you slowed down.
Then someone said "I'll bet I can do that on a bike", and rode into a shallow lake and the bike would ride on the water for a LITTLE while and then sink, so they added some skis until they had the right size, modified the knobbies with hill-climb paddles, and off they went.
Probably all empirical and experimentation, and dredging up a few bikes at first. I've seen similar things happen!