It is interesting how the year is still in some ways intimately tied to religion. The objective of defining a "year" is an attempt to allow a standardized the calendar to account for the seasons, a consequence of the tilt of the Earth and the orbit around the Sun. Unfortunately, a tropical year is 365.2421896698 days as of the year 2000, and this number changes slightly over time. Consequently, if a year was 365 days, then every year the calendar would fall behind by approximately five hours, 48 minutes and 45 seconds relative to the actual relationship of the Earth to the Sun.
Thus, Julius Caesar introduced a calendar with leap years: every fourth year would include an extra day, so that over the the period of four years, the calendar would now be ahead, but only by 44 minutes and 59 seconds. This translates to the calendar advancing ahead of the actual relationship between the Earth and the Sun by 11 minutes and 15 seconds per year. Thus, this calendar becomes out of sync by one day every 128 years.
While this is not much each year, over the course of centuries and millennia, this does add up, and so Pope Gregory introduced a revision to the calendar, whereby each century loses a leap year, unless that century is a multiple of 400, in which case it keeps its leap year. Thus, every four hundred years has 303 regular years and 97 leap years. This therefore averages to 365.2425 days per year, and thus our calendar falls out of sync 27 seconds per year. Thus, this calendar becomes out of sync by one day every 3222 years and four-and-a-half months. Henschel suggested another revision by which every multiple of 4000 years is no longer a leap year, but this is just yet another tweek.
Humorously enough, some sects of Christianity chose not to adopt this "Roman Catholic" calendar due to its relationship with the church at Rome, and continued to use Julius Caesar's calendar.
However, this is still not ideal, as the two numbers still differ, and thus, I wondered if there was an ideal number of years between leap years that should be left out. As it turns out, this is actually quite easy to calculate: 365.2421896698 - 365.25 = -0.0078103302 and the reciprocal of this number is approximately -128.03556 (the number we saw above). Consequently, we should leave out a leap year every 128 years. It is actually quite convenient that 128 is a multiple of four; there is no natural reason this had to occur.
Consequently, the average year under these conditions would be (97 × 365 + 31 × 366)/128 = 365.2421875 days long, and our calendar would fall out of sync one second every five years and four months. Such a calendar would not fall out of sync by even a day until 460872 years have passed.
These days, there are numerous web sites that propose such a cycle of leap years. When I first made this calculation almost two decades ago, there was only one other web site that made such a proposal; I wish I could find that site. I suspect there are physicists or mathematicians out there who have proposed this much earlier, but never-the-less, I find this approach much more interesting. Humorously, if we started our 128-year cycle with 1 CE, then no adjustment would need to be made to our current calendar, only the next non-leap-year would be 2048 CE and not 2100 CE. If we can deal with the Y2K problem, I suspect we can deal with this issue, as well. Also, as this is purely based on mathematics and physical observation, there should be no reason for any sect of Christianity to not adopt this new calendar.