[EM] Question regarding Copeland variant

[Kristofer Munsterhjelm]

Consider this "iterated Copeland" variant:

For the zeroth iteration, each candidate has a score of one.

For the nth iteration, set each candidate's score equal to twice the 
(n-1)th iteration scores of the candidates it beats, plus the scores of 
the candidates it ties.

The winner for the n-th iteration is whoever has the greatest score.

(end of definition)

When this method is iterated just once, the results are according to the 
Copeland variant where candidates score two points for each candidate 
beaten, and one point for each they tie.

When the method is iterated twice, does it always elect from the 
uncovered (Landau) set? If so, does that hold no matter how many times 
it's iterated as long as it's more than once?