Just completed an analysis for the SC Combined database I have here for the Doubles/Repeat/No Repeat happenings.
In the SC Combined database I have that goes back to 10-1-2003, there doesnt seem to be a strong balance leaning one way or the other. This might be different according to Individual database standings, however, by the Combined database play there was a 69-62 ration with more Double games coming from when there is a Repeat digit of previous game.
If anything, there is a stronger possibiliy of a doubles game playing coming from a repeat digit of prior game then a doubles playing when there is no repeat digit from prior game. The "Repeat" situation seem to stay in the lead while they did start out pretty evenly matched in time the "Repeat" count edged ahead and stayed ahead to the final conclusion.
This factor might be something that is more State based then across the board based for Pick 3 tracking. And....again...might be more applicable on an Individual database tracking then a Combined Database tracking.
At least I know in Texas, Doubles are playing off more Non-repeat situations then Repeat in the Midday game which is the one I play.