Ah, but it doesn't have to predict the best cycle! I am fairly sure that even a rough approximation is better than nothing. And, depending on the distribution of optimal sleep-cycle times, their guesses could actually be pretty good for most people.
To be useful, something just has to be better than other options as long as you factor any costs in as well. Since this site does not have any significant costs, it just has to be even little better than nothing to be useful.
Finally, even if the actual calculations aren't useful, it did get me to think about working out better sleep patterns, which is likely to be healthy in and of itself. (Who am I kidding, I'm probably going to forget... but it's a nice thought :P.)
To be useful, something just has to be better than other options as long as you factor any costs in as well. Since this site does not have any significant costs, it just has to be even little better than nothing to be useful.
Finally, even if the actual calculations aren't useful, it did get me to think about working out better sleep patterns, which is likely to be healthy in and of itself. (Who am I kidding, I'm probably going to forget... but it's a nice thought :P.)