Someone commented about the RPI. It is helpful, but it has some big flaws.
The RPI is a computer program. It isn't magic. (I wrote and managed the first RPI for college and TB softball back in the mid-1990s). The RPI was developed to rank international chess players who play in tournaments. So, it was developed to rank individuals (not teams) who play tournaments (not leagues) over a period of years (not months). The RPI is very good for golf and tennis, but not as good for team sports.
The program ranks teams based upon who they beat or lost to. So, if team A beats team B and team B beats team C, then team A is better than team C. A little grouping is done based upon standard deviations, and out pops a ranked list of teams.
INDIVIDUAL vs. TEAMS
The RPI doesn't take into account the players on the field. This is huge in softball--probably more so than any other sport.
Arizona State with Dallas Escobedo on the mound is one team. Arizona State with Popescue on the mound is not nearly as good.
At the beginning of the season, the coach pitches the #2 and #3 pitchers more. As a result, you get some crazy results. Utah beat Arizona State at the beginning of the season. Arizona State didn't pitch Escobedo and played some other non-starters, and Arizona State lost.
The RPI doesn't care...it just says, "Arizona State beat Utah", and lowers the RPI numbers. That loss to Utah is always factored into the rankings.
Alabama plays a bunch of second tier teams at the beginning of the season. They crushes these teams, and the RPI doesn't care.
TOURNAMENT
The second problem with the RPI is that it is based, fundamentally, on a tournament model rather than a league model. What do I mean? In TB (or chess), everyone plays in a weekly tournament. Every tournament ends with the best two teams playing. So, every time there is a competition in TB, golf, tennis or chess, the best two are always identified, and there is always a winner who is the best over a large field of competition.
In other words, Tiger Woods and Phil Mikelson play against each other all the time, so over the course of a year, you can get a pretty good idea as to who is better. Therefore, the RPI database for TB, tennis, golf or chess contains lots of games between top competitors. In that situation, the RPI is very accurate.
In college softball, that doesn't ahppen. There is tournament play at the beginning, and then it becomes league play. When league play starts, basically the RPI of the teams in the league are insulated. The RPI of the league isn't going to change much.
When a lesser team (like Tennessee or LSU) beats a better team (like Alabama), the RPI moves the lesser teams up more than it moves the better team down.
The RPI of Texas is a prime example of the RPI problems. In the last two weeks, Texas lost to Oklahoma, Texas A&M, and Missouri. No reasonable person would think that Texas is the number 1 team in the USA. But, the RPI disagrees.
The RPI is a computer program. It isn't magic. (I wrote and managed the first RPI for college and TB softball back in the mid-1990s). The RPI was developed to rank international chess players who play in tournaments. So, it was developed to rank individuals (not teams) who play tournaments (not leagues) over a period of years (not months). The RPI is very good for golf and tennis, but not as good for team sports.
The program ranks teams based upon who they beat or lost to. So, if team A beats team B and team B beats team C, then team A is better than team C. A little grouping is done based upon standard deviations, and out pops a ranked list of teams.
INDIVIDUAL vs. TEAMS
The RPI doesn't take into account the players on the field. This is huge in softball--probably more so than any other sport.
Arizona State with Dallas Escobedo on the mound is one team. Arizona State with Popescue on the mound is not nearly as good.
At the beginning of the season, the coach pitches the #2 and #3 pitchers more. As a result, you get some crazy results. Utah beat Arizona State at the beginning of the season. Arizona State didn't pitch Escobedo and played some other non-starters, and Arizona State lost.
The RPI doesn't care...it just says, "Arizona State beat Utah", and lowers the RPI numbers. That loss to Utah is always factored into the rankings.
Alabama plays a bunch of second tier teams at the beginning of the season. They crushes these teams, and the RPI doesn't care.
TOURNAMENT
The second problem with the RPI is that it is based, fundamentally, on a tournament model rather than a league model. What do I mean? In TB (or chess), everyone plays in a weekly tournament. Every tournament ends with the best two teams playing. So, every time there is a competition in TB, golf, tennis or chess, the best two are always identified, and there is always a winner who is the best over a large field of competition.
In other words, Tiger Woods and Phil Mikelson play against each other all the time, so over the course of a year, you can get a pretty good idea as to who is better. Therefore, the RPI database for TB, tennis, golf or chess contains lots of games between top competitors. In that situation, the RPI is very accurate.
In college softball, that doesn't ahppen. There is tournament play at the beginning, and then it becomes league play. When league play starts, basically the RPI of the teams in the league are insulated. The RPI of the league isn't going to change much.
When a lesser team (like Tennessee or LSU) beats a better team (like Alabama), the RPI moves the lesser teams up more than it moves the better team down.
The RPI of Texas is a prime example of the RPI problems. In the last two weeks, Texas lost to Oklahoma, Texas A&M, and Missouri. No reasonable person would think that Texas is the number 1 team in the USA. But, the RPI disagrees.
Last edited:
