Elimination 101
Many people ask how I determine that entries are eliminated when there are still several games or rounds remaining. There are 2n possible outcomes in the tournament, where n is the number of remaining games. (There are 9,223,370,000,000,000,000 possible outcomes at the beginning of the tournament. Even after the first round - more than half of the games - is complete, there are still 2,147,483,648 remaining combinations.) Thus, it's impossible for me to check every possible outcome until the number of remaining games gets pretty small. So, I simplify the process by classifying eliminations in two categories: Simple and Similar Entries.
Simple Elimination is the more obvious. Within a computation method, I report each entry’s actual accumulated points and total possible points. The total possible points assume that all of an entry’s remaining picks will win as projected in the entry. If Entry A’s total possible points are less than another entry’s actual accumulated points for games already completed, Entry A is Simply Eliminated.
Similar Entry elimination is more complex. To help explain the process, assume the following entries are competing:
| Entry | Actual / Possible | Champion | Final Four |
| Entry A | 27 / 150 | Illinois | Illinois, Duke, North Carolina, Wake Forest |
| Entry B | 20 / 155 | Oakland | Oakland, Montana, Air Force, Milwaukee |
| Entry C | 28 / 153 | Illinois | Illinois, Duke, North Carolina, Wake Forest |
If Entry A is not Simply Eliminated, I test to see if it can win the pool. To do this, I construe a scenario comparing the remaining games for Entry A with the remaining games for the entry with the highest possible points (Entry B). This scenario will choose all of Entry A’s remaining teams to win and go against the remaining picks of the leader. Then, I score all entries assuming this construed scenario. If Entry A would win under this scenario, it obviously is not eliminated.
If Entry A is not the winner under the construed scenario, I check to see which entry is the winner. If it is Entry B (the original leader), Entry A is eliminated. Quite often, however, the construed scenario results in a third entry (Entry C) as the leader. If this happens, another construed scenario is developed, now picking games against both Entry B and Entry C. This process is repeated until a scenario results in Entry A’s victory (not eliminated) or has a competing entry win twice (eliminated).
In the above example Entry C leads Entry A in terms of actual points garnered from the early rounds, and his remaining entries are quite similar. (They both have the same Final Four.) Thus, for Entry A to win, the games that must go in his favor would also advance Entry C. Since Entry C has a higher starting score, the similarity of their picks prevents Entry A from winning.
It's possible that I can project somebody as eliminated when they're actually not. The inverse is not possible.