I have been giving it a whirl too, and it looks like the problem is that is counts the number that it is trying to match.
For example: Previous number is 883 and current number is 230, so you get the 2 digit code 82, when it searches for 82 as a match it not onlys counts all previous 82 instances - it's also counting that very same instance. Not only am I finding this to disturb the count, but it is also resulting in the "false" correct picks. Again, if it didn't count the 82 code that it is actually trying to match than you get the actually tabulated pick versus the current predisposition towards what we already know is correct.
If I have lost anyone or someone doesn't agree, let me know. But if your finding the same thing... ...let's see what we can do to fix it.
BTW, hypersoniq... ...this is simply a piece of art you have made here. Very well thought out.
Also instead of having the picks go from 0 to 9 or 9 to 0, I added an extra column (where the black out column was) to count the number of high numbers that matched and turn the background red. So, if the 82 code had 3 "0's" to follow, 3 "4's" to follow, and 3 "8's" to follow. The cell turns red and tells me that there are 3 matches for that numbered sequence.
Again, hope this makes sense... ...Let me know, Pogo