Reference to my message n° 1516

and:

http://www.projectpluto.com/accuracy.htm#planetary

"The accuracy of the Moon is a separate problem. For this, Guide can
use a truncated version of the ELP-2000 theory of Michelle Chapront-
Touze and Jean Chapront, of the BDL (Bureau des Longitudes) in
France. The accuracy looks to be in the neighborhood of a hundred
meters or so, but I'm still investigating that question. (Such
accuracy may appear to be "more than anyone would ever need", but in
truth, greater accuracy wouldn't hurt when doing lunar graze work.
One can never be too rich, too thin, or have too much accuracy in
one's astronomy software.)"


The moon position ephemeride in Guide has to be updated.

Guide uses an old theory (the reference is the old JPL DE200 data
file)valid over only the interval 1750 to 2169.

Please refer also:

http://ssd.jpl.nasa.gov/iau-comm4/de403iom/

For the moon, it is now quite apparent that DE118 (and therefore,
DE200) contains a significant error in the lunar
longitude for the years prior to 1750. The integration of DE118 was
performed in a number of separate
computer runs, where it was necessary to restart the integration
process each time from where the previous run
had ended. The backward integrations covered the time-spans, 1969 to
1900, 1900--1850, 1850--1800, 1800--1750,
1750--1664, and 1664--1600; the forward pieces covered 1969--1990,
1990--2050, 2050--2169, and 2169 to 2200.
Back when these runs were made, the restart process was not
automatic: there was a major error, affecting all
planets, committed at the restart in 2169 which has long been known;
also, there was a minor error made in the
restart at 1750 which has only recently come to light. This latter
error affects mainly the lunar longitude,
causing a run-off which reaches nearly 40 arcseconds by the year
1600. As a result, DE118 and DE200 are
strictly .


Regards