Re: [find_orb] Re: (6995) Minoyama/asteroid perturbers

Bill Gray Feb 13, 2013

Hi Alessandro,

I'm finding (after rejecting a few outliers) that with asteroids plus all planets,
I get a root-mean-square residual of 0.557". Shut off Venus and try again, and it
climbs to 0.568".

In running this sort of comparison, it's important to make sure that the same
observations are included/excluded with both runs. Otherwise, you can get situations
where the "wrong" setup (with, say, the effects of Venus not included) seems to
show lower residuals, but only because a few observations got so bad that they
were thrown out.

I _have_ seen situations where the residuals are smaller with one or more perturber
omitted, even though the observations used are the same. It's not a common situation,
but it does happen. But this doesn't mean one should exclude perturbers. As I
mentioned in a previous e-mail, you can be certain that the actual universe leaves
all perturbers on, all the time. The only reason to leave perturbers off is for
better speed. (And even that reason is becoming less significant, with faster
computers.)

-- Bill