Well, I finally jumped into playing the Texas Pick 3 games. After a careful analysis of the odds and of the Sum It Up option, I decided that it was the Sum It Up where the hits were to be made so I played exact numbers with minimum $0.50 wagers but selected numbers to cover a spread of sums adjusting the bet amount to maximize the expected return at close to a $0.50 on the dollar expectation. It would seem that playing boxed simply drops the expectations down to $0.48 on the dollar from $0.50 so it's better to purchase more three number combinations to cover the Sum It Up spread which tended to range in expectation from $0.3780 on the dollar return on the 10/17 sums to $0.50 on the dollar returns on the 3/24 and 0/27 sums. The other sums approached $0.50 return but often only if you wager $1 or $5 instead of just the base $0.50 on the Sum It Up. If you used their published odds which are rounded off, some of the expectations exceeded $0.50 but if you corrected the odds then only 3/24 and 0/27 sums are at the $0.50 mark, all others only approach it. I don't understand why the pay table shows the payouts for $2, $3, and $4 wagers when the playslips only allow for $0.50, $1 and $5 wagers on the Sum It Up.
My first foray costed $48.50 and won $40, a $0.82 on the dollar return which is somewhat more than expected and given that $11 were all or nothing bets on the outrigger 000, and 999 bets, I actually made money on the covering the spread section of the wager. All the wins were on the Sum It Up rather than a hit on the Pick 3.
I guess I didn't do badly and it does seem like you can get some consistency out of the Pick 3's but it's still little more than waiting for the big one to strike which as long as you wager on a 000 or 999 with a $5 Sum It Up wager, can be as high as $2,750 or $3,000 depending on if you have a $0.50 base wager or a $1 base wager. Still at 1 in 1,000 odds which means if you played just 000 and 999 for each draw both with the $5 sum it up wager, it would probably take 2.39 years to be 95% certain of hitting that $2,750. The 50% mark which would likely be the median time to hit that prize would be 0.55 years and would cost $3,808.50 to hit that $2,750 which as lotteries go is pretty good but playing just 000 and 999 wouldn't get you much else while waiting for the big one to hit. I suspect that the way to play the game is to cover the spread of the most likely Sums but not all the Sums and then the outrigger bets on 000 and 999. 333 and 666 seems a bit unusual as the payouts are slightly better than most of the other sums at the $1 wager mark but really not by much if you take the actual odds into consideration instead of the published rounded off odds.
Interestingly, in the 5,115 night draws in their history file, there are 7 0-0-0 hits and no 9-9-9 hits. At 1:1000 odds for either, you would've expected it to be more balanced over the years. The day draws are more balanced at 2 hits a piece over 2,454 draws which is right in line with what one would expect.