1) No repeats after Reset
In this case, lets say you have values A, B, C, D with the List set to Random.  The List samples B, C, A, D for the first trial, when the List is reset for the next set of trials, exp authors typically would not want "D" to be the next values chosen since it was the last one chosen in the first trial and would cause repeated values.  This is most easily achieved by the List ensuring that the same level/row is not the first after the reshuffle.  If you had A, A, B, B, C, C, D, D, you could have two D values in a row because the D in row/level 7 is not the same as the D in row/level 8.  Specifically, in this case the same row would not be selected again after the shuffle and no comparison of values of attributes/columns would be used.

2) No repeates based on condition
In this case, you don't want a specific condition based on attribute value pairs to not occur 2x or x-back in a row.  You would specify a listing of attribute pairs to make the determination.  This is more difficult since it is either difficult or impossible to specify nested attribute values as part of the condition comparisons.

Note that in these cases this shuffling occurs only after the List is reset and not during a reset to fulfill the condition.