The problem here is that 0.25 is actually an overestimate by about 3 in every 400 years. That’s why we don’t have leap years on every hundredth year, but we do have them again every 400. (And, of course, you can get even more precise than that.)