GOUT has no cue marks.

IP might be 2000ft reels ... I'll likely never see an IP to know for sure.
But these are some we can see now:
- domestic Pan/Scan CAV laserdisc
- Japanese SC laserdisc (first widescreen transfer)
- GOUT

Now for this example, let us consider the first changeover in ESB.
First reel ends with Artoo standing outside the base, scanning for Luke.
Threepio tries to reassure themselves "He'll be quite alright."
Second reel begins with Luke collapsing in the snow (returning from wampa cave.)

Pan/Scan laserdisc is missing 4 frames of Artoo + 3 frames of Luke
3+4 = 7 frames missing compared to GOUT

JSC laserdisc is missing 5 frames of Artoo + 2 of Luke = 7 fewer than GOUT

[GOUT might be missing some frames too, compared to IP, but we'll never know it.]

So these are the raw data I've collected.
They're typical of what I've found across the board for all three films.

What does it mean?
Well I'd say . . . IP can exist on 2000ft reels if-and-only-if
a new IP was pepared (from 1000ft negative reels) for each laserdisc release.

zombie, how does that match your earlier research? I've searched in vain for your post(s) about when new IPs were printed. IIRC you had written on that topic before.
Were new IPs made for each of those laserdisc releases?

By the way I've done some interpolating, and some extrapolating based on Moth3r's cue dots, and I've decided "Thank the maker" reel begins at GOUT offset 27,820 (give-or-take four-or-five frames ;)