Observations

Stats I got.

I count the average number of comparisons now. There's a challenge, implementing linear interpolation for strings, for instance. Doable, I guess. Thanks a lot for the idea of the integral of Cantor Set. It helped a lot to improve the algorithm.

dataset	size	boosted search	standard binary
Sine	10000	9.00	15.36
Cantor Set	1025	16.00	17.86
Convex/concave	218983	4.52	6.45
Concave	109990	4.96	5.00
Linear	50000	3.00	15.36
Convex	108993	3.60	5.00
Concave	10002	3.71	4.10