If a deck of 52 cards is cut exactly in half (top half and bottom half), then the top half can be partitioned into T segments (keeping the order of the cards) and, similarly, the bottom half can be partitioned into B segments. If T=B then there are two ways to interleave these partitions of cards -- dropping segments of cards alternately from T and B, beginning with T (the first way) or beginning with B (the second way). If T=B+1, then there is only one way to interleave the segments -- one must begin with T. If T+1=B, one must begin with B. Any other relationship between T and B make them impossible to interleave. We can count the number of partitions of 26 in all possible arrangements, they are: #partitions of 26; #of ways to make partitions of this number respecting order 1 1 2 25 3 300 4 2300 5 12650 6 53130 7 177100 8 480700 9 1081575 10 2042975 11 3268760 12 4457400 13 5200300 14 5200300 15 4457400 16 3268760 17 2042975 18 1081575 19 480700 20 177100 21 53130 22 12650 23 2300 24 300 25 25 26 1 We can now count all possible T,B pairs where T and B are at most one apart. This gives us a grand total of 374,369,872,911,804 possible riffle shuffles of 52 cards. Considering that 52! is the enormous 80658175170943878571660636856403766975289505440883277824000000000000, I can't see how to show that there is some number, N, of riffle shuffles such that one can always produce any particular permutation of the 52! by applying at most N riffle shuffles to the identity permutation.