A couple of days ago I download PP3 and said that I would run a test using PA Data. According to Todd, the creator of PP3, the more historical data you enter the better the software gets at predicting winning numbers (or numbers that are close). He also said that in his tests with CA Data, on average he got at least a boxed winner every 27 to 31 days.
Well for my test I entered all PA Daily Number drawings from the very first (03/01/1977) until 07/20/2001. I then ran PP3 to get the predicted number and compared that to the actual drawing. I then added the actual number to the database and moved onto the next day. I did this for 365 drawings between 07/20/2001 and 07/19/2002.
PP3 correctly picked 3 numbers only 2 times. Once boxed, and once straight. It correctly picked 2 numbers 50 times. 1 number was correctly picked 161 times. And zero numbers came up 152 times. If I had played the number that PP3 gave me for every drawing as a 50/50 ticket, I would have been UP $5 at the end of my year long test. However, PP3 costs $70 ($50 for Lotter Post users until the end of the month). So overall I would have lost $45 to $65.
After seeing these test results, I was curious to see how a random number generator would fair. So I went to **don't post links**.com and ran the generator 365 times and compared them to the same PA numbers I used for the PP3 test. The random number generator correctly pick 3 numbers ONCE (Boxed). It got 2 right, 38 times. 1 right, 192 times. 0 right, 134 times. All total my loss for the year would have been $325.
If anyone wants to see the numbers PP3 picked compared to the real numbers, I can post them here later on.
All in all, PP3 is a nice program, but it doesn't seem to hit that frequently for PA players.
