Ok, just tested that theory, didn't work out. Looked good at first when with only four past draws in the history there was a straight all odd A chart hit, but that was the last straight win for the remaining 38 days I tested using the A chart. The B chart did worse, took 14 days to get it's first straight hit and it wasn't an extreme.
Looking at the situation, we know there are 120 boxed singles, 38 of which are extreme as all OOO Odd, EEE Even, LLL Low, HHH High.
When we include the extremes, the wheel doesn't return an additional 38 lines, it adds another 10 lines so we have basically a 1 in 4 shot at being correct the 1 in 3 times an extreme singles draws shows up when singles are drawn.
The original default setting does so well because it's throwing up almost a third of the non-extreme singles, so it can be expected to win a third of the time a normal single is drawn which is much more often then a double, because while there are almost as many doubles (90) as singles (120) there are only 270 possible doubles straight/boxed compared to 720 singles straight/box pool for the winner to come from.
Supposedly every Pick-3 number has an even chance of being drawn, but because six boxed combinations are possible from singles compared to only three boxed combinations from doubles when a program is picking box (that can be played straight/box as well) not only are the singles better represented and weighed to appear more often in box form in drawings, many believe if black marbles are twice as many in the drum as white, they will be drawn more often then white if the drawn marble is replaced for the next draw. Not to mention the software is looking at how the numbers are doing positionally and loading the wheel based on previous performance. BobP