I can’t claim to be particularly versed in the debates about Newcomb’s paradox, so I might be wrong here, but it seems to me like you got Joyce’s argument precisely backwards. His entire point seems to be that Rachel and Irene are in fact not facing the same options.
Irene has the options
One-box and most likely leave with $1.000.000, but possibly leave empty empty-handed
Two-box and most likely leave with $1.001.000, but possibly leave with $1.000.
Rachel has the options
One-box and most likely leave empty-handed, but possibly leave with $1.000.000
Two-box and most likely leave with $1.000, but possibly leave with $1.001.000.
From Rachels perspective, the two statements “Irene’s options are enviable”, and “Irene should have chosen option ii” don’t seem to contradict each other. They seem like the logical equivalent of envying the hand of your poker opponent, while simultaneously insisting that you played your inferior hand better (even though your opponent did in fact end up winning).
The larger point Kuhn is getting at is that early adoptees of new theories rarely do so “because of any scientific merit”, but rather because the new theory appeals to them on a personal or aesthetic level. Novel theories, like those of Copernicus and De Broglie, often do a worse job of explaining the evidence than their established rivals, and early adoptees are essentially making a leap of faith in defiance of the bulk of the evidence. Only through the work of these early, ‘irrational’ adoptees are the rational reasons for adopting the new theory uncovered.