BobP = Here is some info about the "one digit" that you might find useful. This is in regards to "repeat digits" from previous game played.
When you are coming off a 3 different digit game and there is a 1 digit repeat, you reduce the 220 possible combinations to 84 possible combinations. This equates to 28 possible 3 digit combinations per each of the 3 digits that previously played.
If you can eliminate 2 of those digits then you have only 28 possible 3-digit combinations to work with that include Doubles and Pures. In this scenario, out of the 28 possibles 15 of the 28 are 3 different digit combinations leaving 13 for Doubles & Pures.
If you are working off a No Repeat game, the 220 possibles are reduced to 80 probables. Out of this 80, 50 of the possibles deal with Doubles & Pures while 30 are 3 different digits.
I found here in Texas Midday that with the 50/30 split with Doubles & Pures being the larger group that Doubles happen more often coming off a No Repeat then a Repeat situation by a 57 to 45 ratio in the database I have that goes back to 2-3-2003. There are periods of time when each is running about 50/50 but there seem to be longer periods of time when the case of a Double appearing on a Non-repeat instance happens more often then coming on a Repeat situation.
When I checked the South Carolina Combined Database, things were different. Seems a Double happening with a repeat digit from previous game (whether the single digit or the one that doubles) happens more often then a Double coming off a "non-repeat" circumstance.
In fact, Doubles in SC (database goes back to 10-1-2003) they were about 50/50 starting out until Doubles with a repeat digit took a slight lead to be 69 to 62 in the final analysis. For this reason I wonder if things might be different if only Individual databases were analysised and it would also be on a "state-by-state" basis as to whether the higher ratio would pertain to Repeat situations or No Repeat situations.
Something else, if you are coming off a Doubles game and have a 1 digit repeat, the 220 combinations are reduced to 72 possibles. Of these 50 will be 3 different digits and 22 will be Doubles/Pures. I know we sometimes see "back to back" doubles and you would think the probability of that happening would be less of a chance with the higher ratio of 3 different digits probable then Doubles.
Anyway....thought the above would be a little "food for thought" <G>