Hi Ben,

A parallax such as 4 +/- 14 milliarcseconds doesn't tell you
all that much. It does tell you that it wouldn't be surprising
if the star had an actual parallax of 18 mas (one sigma off),
meaning a distance of about 55 parsecs, giving an estimate of
how close it might be; but it doesn't provide any constraint
as to how far away the star might be.

I don't recall exactly how the Tycho mapper did its stuff.
There were a lot of complications beyond what you might have
expected from your garden-variety CCD imaging setup (and this
was all being done with mid-1980s technology). Some odd things
crop up when you're mapping the entire sky down to milliarcsecond
precision... perhaps someone can fill in details here?

I do know that several people have commented that at the
faint end, the Tycho instrument was pushing the limits of what
it could do at _all_. So the parallax and photometry sigmas
all became quite inflated, basically meaning: "We barely
detected this star in the first place, much less got good
data for it. So regard all of these numbers as being very
approximate indeed."

-- Bill