"blank split" functionality breaks "best split"

Adding a "blank split" to place it into a run in the right order still counts towards the best split, so you might end up with a bunch of (for example) 0.0-second best times across the board.

times entered for "blank splits" shouldn't be factored into Best Split, in my opinion.
