Patience--most respondents here require sleep (I left you at 11:30 PM last night) and work full-time. I have only now been able to return to this.
Please note that you should have laid out the entire problem from the beginning, as I would have approached it differently. As I understand it, starting from the beginning, you have data like the following:
User ID Score1 Score2 Wins TBact TBpts diff
john 1 1 1 2 3 2 1
john 2 1 0 1 5 2 3
john 3 4 8 12 38 30 8
tester 1 5 12 17 2 2 0
tester 2 6 6 12 38 38 0
jason 1 3 3 6 16 12 4
jason 2 5 4 9 38 42 -4
Please confirm the assumptions I am making before I take this further:
1) {@Wins} = Score1 + Score2
2) You want the row with the highest ID number, NOT necessarily the highest {@Wins} per user. In other words, for tester, you want ID #2 even though {@Wins} is higher for ID#1.
3) You then want the user with the highest {@Wins} in the selected rows, and if there is a tie, break it with the lowest {@diff}. Your final report will show this one user.
4) I am assuming that TBactual and TBpoints are database fields, not report summaries or formulas.
I have a solution in mind, but would like you to verify the above. If I am wrong on any single point, please be sure to provide enough detail to understand. In particular, if any of these fields are formulas, please supply the exact formula.
-LB