Centering stars without a reticle eyepiece

I (think I) have discovered a very accurate star centring technique
that works without the need for a reticle eyepiece. Google turned up
nothing, but it's difficult to find search terms that reduce the
irrelevant hits, so I may have missed a reference in the pile. It's
inconceivable to me that nobody else has found this, so I'm sure I've
just re-invented a known method, but I'd like to hear other opinions
on it (questions nearer the end).

I got my LX200 a couple of months ago, and subsequently blew my
accessories budget on 4 Naglers (way, way more than I intended to
spend, but so worth it). As a result, I haven't yet got a reticle
eyepiece and just "eyeball-centre" guide stars during alignment. I
know pointing accuracy depends on consistency of pointing at guide
stars and the easiest way to be consistent is to centre them.
Unfortunately, this difficult to do with wide FOV lenses, especially
if you're not sure you can see the actual edge of the field all round.

To date, my solution has been to 2x Barlow a 6mm super plossl (846x
!?!?!)and defocus the image a bit so that the "disc" of the defocused
star fills 1/3rd the FOV. This makes it quite easy to centre the star,
but there is still an element of estimation and hence some error. And
there's a slight inconvenience in having to faff with 1 1/4" gear when
the night's observing will be a 2" EP.

When I was aligning the scope the other night (happened to be on
Arcturus) I noticed something surprising. As I moved the scope so that
the star image moved down the FOV – a very faint "ghost star" moved up
the field of view in exactly the opposite direction by exactly the
same amount. The ghost is visible if the star is defocused or not. The
key thing though – an imaginary line between the "ghost" and the real
star image crosses the exact centre of the FOV and is completely
independent of pupil position over the eyepiece!!! Move the scope
until the ghost and the guide star overlap and bingo – EXACT centring
has been achieved.

I tried this with 9, 12, 17 and 31mm Naglers and a range of Super
Plossls and it worked with all of them (harder to see in the 31mm
Nagler). I noticed a marked improvement in pointing accuracy after
using this method too. Brighter stars work better – but it seems to
work down to mag 2 or so).

Note: I think I noticed it this time because I was doing a
re-alignment after knocking the scope. Unlike other times when I set
up (in twilight), my eyes were more dark adjusted - that "ghost star"
really is very faint.

Questions:
What causes the ghost star – is it internal reflection in the EPs or
the SCT? (Cue SCT-bashers to highlight this "appalling optical defect
inherent in the SCT design" :-)

Should I worry about "fake" stars in rich fields? I will investigate
when skies clear again, but given it only seems to work with very
bright stars, I doubt it's an issue (and I'm sure there would be a
thousand-message topic on saa or MAPUG if it was).

Does this work with other lens/telescope designs too?

Do you think it's accurate? I do, and I'm not going to bother buying a
reticle eyepiece until I move onto the CCD thing next year (easily
absorbed into the cost of guide scope, wedge, CCD kit and observatory
to put it all in :-)

Cheers
Beats