It is very possible that we will never have an absolute agreement on the way the Mesoamerican calendar relates to our own; too much has been lost, too little is known. As previously mentioned, two of the major authorities on-site in the years immediately after the Conquest, Sahagun and Duran, placed the beginning of the Aztec year on February 2 and March 1 respectively. Since there is some evidence that the major cities in Mesoamerica at the time of the Conquest may not have all been using the same day-count, it isn't even out of the question that both were correct, since one was stationed in Tenochtitlan and the other in Texcoco.
Alfonso Caso has proposed a correlation anchored on a statement that the day of the fall of Tenochtitlan, that is, the day Cuahtemoc surrendered to Cortez, which was, according to the chroniclers August 13, 1521, was said to be a day 1-Coatl. This day count, according to several authors is still in use in some highland Maya communities in Mexico.
We know--as well as we know anything about this--that the year 1519, the year Cortez arrived in Mexico, was a year 1-Acatl. Much is made of this in Indian myth; that year, occuring once every fifty-two years, was a year of change. If 1519 was 1-Acatl, then 1520 must necessarily be a year 2-Tecpatl and 1521 the year 3-Calli. On a year in which 1-Coatl occurs on August 13, the day 3-Calli, the year start, occurs on May 2.
The problem with this is that if the New Year occurs on May 2, then Cortez did not arrive in a year 1-Acatl, but rather in a year 13-Tochtli, since his landing was in April. Secondly, all the ceremonies are thrown off in odd ways; the rainy season in the Valley of Mexico is from the end of May until October, and the first five or so ceremonies of the year are devoted to asking for that rainy season to arrive, pretty absurd if it already has. Using the Sahagun or Duran correlations these ceremonies are logically placed toward the end of the dry season.
There are two reasons why the Caso correlation may be in error; first, during their "residence" in Tenochtitlan, the Spaniards by their own admission lost track of the days ("we celebrated Mass on what we believed to be a Sunday"), and it isn't out of the question that the hardships of the seige may have caused the Mexica to do likewise. As far as the Maya groups who use it, it must be remembered that virtually all the Maya groups (except the Lacandon) were disrupted by Spanish rule; the day-count was quite probably lost and regained, maybe several times, and perhaps from some academic source. We do not know what if any day-count the Lacandon use.
Leap-years present a different problem; we know the Mesoamericans observed them, we do not know how or when. Von Zandwijik proposes doubling the last day of each Tecpatl year (1 of 4 possible, all years are either Acatl, Tecpatl, Calli, or Tochtli), and my program does the same, though it does not name the day, it merely calls it a "Leap Day;" it could be called (N)Ehecatl (the last day of any Tecpatl year being (N)Ehecatl, where N=1-13) or (N+1)Calli, in either case a duplicated day. (the Von Zandwijik correlation uses Feb. 18 as the year-start, which is as reasonable a guess as any other).