Wow, I never realized this guy was such a show boater. His criteria is so vague you can put yourself on either side of it an most of his "keys". He just wants to hedge his bets and make his sterling record, which was broken in 2000 look somewhat clean.
The model may have problems but 2000 was not one of them. When he developed the model in the 1980's he clearly said this was only to predict the popular vote winner since 1860. He used 1876 and 1888 as examples where his model would have called the popular vote winner, not the EC winner.
There is of course the fact that in both of these elections there was massive voter suppression of Blacks in the South, so dubious whether the Democrats would have won the popular vote under fair conditions in either one, especially in 1888.