I think this is one case where I would seriously consider a gps, perhaps connected to a duino to simulate a speed signal just for "simplicities sake". On a car it is a non-issue (unless you spin the tires constantly), but a boat is of course another matter.
There might be some other methods, i.e. I assume boat speed is largely a function of engine rpm plus some fudge for weight/acceleration. If you have a duino monitoring rpm, and you map out the slippage at different rpms and the slippage under various acceleration points, you could get a pretty accurate distance perhaps, and wish you spent the $60 on a gps module
