Nuts, Mac, did you consider timing these things is probably a waste of time because different routines sort/group differently and it all depends on the pattern they are working with? One might be faster than another until the data is scrambled a different way, then it would be slower. You cannot with any ease make a test condition that evaluates all of these possibilities so that the test is fair. RANDOMIZE TIMER can only approximate a winning algorithm based on an average of hundreds of runs. You would have to make a database of all those randomizations and then run them through each entries code and the lowest average time would be the winner.
WHOGAS? - If you win by 1MS, you don't even get a crummy QBasic.con T-shirt.
Let's just all agree right now that Fruit-Groups is the best and put this puppy to bed - LOL. (Just sealed my own doom by stating that!)