A century before Callippus, Meton invented the cycle of 19 years that counted 6,940 days, which exceeds 235 lunations by almost a third of a day, and 19 tropical years by four tenths of a day. It implicitly gave the solar year a length of 6940/19 = 365 + 1/4 + 1/76 days = 365 d 6 h 18 min 56 s. But Callippus knew that the length of the year was more closely 365 + 1/4 day (= 365d 6h 00m 00s), so he multiplied the 19-year cycle by 4 to reach an integer number of days, and then dropped 1 day from the last 19-year cycle. Thus he constructed a cycle of 76 years that contains 940 lunations and 27,759 days, and has been called the Callippic after him.
Although the cycle's error has been calculated as one full day in 553 years, or 4.95 parts per million., in actuality 27,759 days in 76 years has a mean year of exactly 365 + 1/4 days, which relative to the mean northward equinoctial year is about 11 minutes too long per year, in other words the cycle drifts another day late per 130 + 10/11 years, which is considerably worse than the drift of the unrounded Metonic cycle. If the Callippic cycle is considered as closer to its unrounded length of 27,758 + 3/4 days (based on 940 lunations) then its accuracy is essentially the same as the unrounded Metonic cycle (within a few seconds per year). If it is taken as 940 lunations less one day then the Callippic mean year will be shortened by 1/76 day (18 minutes 57 seconds), making it grossly too short, and it will also grossly drift ahead with respect to the mean lunar cycle at the rate of 1/940 of a day (1 minute 31 seconds) per lunar month. If the cycle length is truncated to 27,758 days then the mean year is 365 days 5 hours 41 minutes 3 seconds, or almost 8 minutes too short per year, and it will drift ahead of the mean lunar cycle by about (3/4)/940 day (1 minute 9 seconds) per lunar month. Altogether, the purported accuracy of this cycle is not impressive, but it is of historical interest.
The first year of the first Callippic cycle began at the summer solstice of 330 BC (28 June in the proleptic Julian calendar), and was subsequently used by later astronomers. In Ptolemy's Almagest, for example, he cites (Almagest VII 3, H25) observations by Timocharis in the 47th year of the first Callippic cycle (283 BC), when on the eighth of Anthesterion, the Pleiades were occulted by the Moon.
The Callippic calendar originally used the names of months from the Attic calendar, although later astronomers, such as Hipparchus, preferred other calendars, including the Egyptian calendar. Also Hipparchus invented his own Hipparchic calendar cycle as an improvement upon the Callippic cycle. Ptolemy's Almagest provided some conversions between the Callippic and Egyptian calendars, such as that Anthesterion 8, 47th year of the first Callippic period was equivalent to day 29 in the month of Athyr, in year 465 of Nabonassar. However, the original, complete form of the Callippic calendar is no longer known.