I think it is just the usual order on the rational numbers, but since he uses only some of the rationals, it does amount to using ordinals.
But, I think your solution is off by at least a factor of omega.
But, I think your solution is off by at least a factor of omega.
You probably even could get an embedded omega squared with
1 - (1/2)^n - (1/3)^m with m>n
So 1/2 = omega
3/4 = 2 * omega
7/8 = 3 * omega
and 1 is omega * omega
He might have omega squared in his version, but I didn't read it that carefully.
Now play the version where the secret ordinal is an uncountable ordinal or a strongly inaccessible cardinal.