NCAA Division I Methodology
Calculating the Ratings
Each runner's rating is calculated based on how he or she performs in each race compared to every other runner in the race. Because it is based on a comparison to other runners, the absolute time that is run is irrelevant.
The rating of each runner is set to the value that will cause the smallest variance between the results of all the races during the season and the expected results based on the ratings.
The team ratings are the scoring of the team's top five runners in a large sample meet with varying levels of competition.
Tournament Simulation
The tournament simulation is run just like the actual tournament, but uses each runner ratings to predict the outcomes. Each region championship is simulated, then the teams who advance are selected, and finally the championship race is simulated. The process is then repeated a set number of times with each runner's performance randomly adjusted each trial to account for the natural variation in race performance.
Without knowing the order the teams will finish in regional races, it's hard to predict which teams will advance through an automatic or at-large bid. One can predict who will advance based on a specific order of teams, but if the order of even one region championship is different, it can change not only which teams advance from that region, but also which at-large teams from other regions as well. Because the tournament simulation on this site does not rely on one predicted set of regional results, but instead the likelihood of each outcome, it also produces the predicted likelihood of each team advancing based on all possible outcomes.
The tournament simulation also provides interesting statistics on the likely placing of each team in the region championship, how likely each team is to make it to the NCAA Championship race, and how each team is likely to place in the NCAA championship race.
The selection process for the championship round used in this simulation is the same as the NCAA uses to advance teams and is outlined below:
1. The top two teams at each regional championship meet automatically qualify.
2. Thirteen at-large teams are then chosen. The top two unselected teams from each regional are considered for each selection
3. The next at-large berth is awarded to the team with the highest win total against teams already selected. There are tie breaker procedures first based on head to head competition, then common opponents.
4. Teams may only accumulate these wins in races where both teams are competing with at least four of the seven runners who compete in the regional qualifying meet.
5. If the unselected team with the most wins is the second unselected team from a region, both teams will advance if the higher placing team would not eventually adavance on its own. Otherwise, neither team is selected at this time. Both teams may only advance if the push process has not been used in this region earlier and there are two or more at-large positions available.
6. The process is complete when 31 teams have been selected (18 automatic and 13 at-large).
Complications
The greatest difficulty in trying to predict future races is that it's impossible to know which runners have run as hard as they can in previous races. This is particularly true for the elite runners, who often can win races without giving it their all. Accordingly, these ratings assume each runner has given maximum effort in each race.
Another challenge is knowing which runners are injured and the extent of those injuries. Primarily, a runner will be included in the tournament simulation unless it is commonly known that he or she is out for the season with an injury. The effect of injured runners can have a large effect on the tournament statistics, especially if there are injuries to multiple runners on the same team that aren't taken into account.
