Bracket Projection Comparison - Through the Years
Through the years, I've learned plenty about projecting brackets from choosing indicators to weighting the quality of wins. However, experience doesn't necessarily breed results. Ultimately, the experience of projecting brackets prevents aspiring experts from making embarrassing mistakes. I was able to dig up my final projected brackets from 2005 and 2006 and I calculated the main indicators. What I noticed was somewhat disturbing.
2007
TEAMS SELECTED: 63/65 (32/34 at-large)
PREDICTED WITHIN ONE LINE: 49/65
EXACT SEEDS PREDICTED: 27/65
Undoubtedly, 2007 marked the season in which I dedicated the most attention to bracket projections in about three or four years. This meant more games watched, more access to outside views, and more time to examine the resumes and project the brackets.
2006
TEAMS SELECTED: 62/65 (31/34 at-large)
PREDICTED WITHIN ONE LINE: 50/65
EXACT SEEDS PREDICTED: 27/65
2006 was a different story despite the nearly identical stat lines. At the time, I was finishing up my graduate studies and had little time to dedicate to the brackets outside of Sunday nights. I didn't even begin thoroughly examining teams outside of North Carolina and George Mason (the teams I follow most) until early February. Additionally, I was out of town during the final weekend of the season and had a provisional bracket with built-in scenarios for the games on Selection Sunday.
2005
TEAMS SELECTED: 63/65 (32/34 at-large)
PREDICTED WITHIN ONE LINE: 55/65
EXACT SEEDS PREDICTED: 31/65
Outside of the teams selected, the stat line for 2005 is flat-out gawdy. The reason why this is, ... I do not know. In fact, I was waist-deep in my graduate program and working in my spare time, so bracket projections were sloppy and poorly analyzed in comparison with the work put forth in 2007. My lack of time to view games and study the RPI amongst other indicators led me to believe Miami-Ohio and Buffalo of the Mid-American Conference were the 64th and 65th teams in the field. How little I knew.
As it stands, my "best" results were in the year that I did the least analysis, watched the fewest games, and did not have a steady set of indicators. I say "best", because best in this case does not resemble my best work. It just means that my findings for 2005 were more in line with the Tournament Selection Committee.
In closing, this signifies two things. First, it reveals that the subjective element plays a larger role than perceived or desired. Second, this meshes with the committee's desire to have its bid numbers work in concordance with conference RPI rankings. In the end, the result is that the best 65 teams will not participate in the NCAA Tournament. Nevertheless, it's the best three weeks of basketball throughout the entire year, so we'll put this on the backburner by the time that the Elite 8 rolls around.
Through the years, I've learned plenty about projecting brackets from choosing indicators to weighting the quality of wins. However, experience doesn't necessarily breed results. Ultimately, the experience of projecting brackets prevents aspiring experts from making embarrassing mistakes. I was able to dig up my final projected brackets from 2005 and 2006 and I calculated the main indicators. What I noticed was somewhat disturbing.
2007
TEAMS SELECTED: 63/65 (32/34 at-large)
PREDICTED WITHIN ONE LINE: 49/65
EXACT SEEDS PREDICTED: 27/65
Undoubtedly, 2007 marked the season in which I dedicated the most attention to bracket projections in about three or four years. This meant more games watched, more access to outside views, and more time to examine the resumes and project the brackets.
2006
TEAMS SELECTED: 62/65 (31/34 at-large)
PREDICTED WITHIN ONE LINE: 50/65
EXACT SEEDS PREDICTED: 27/65
2006 was a different story despite the nearly identical stat lines. At the time, I was finishing up my graduate studies and had little time to dedicate to the brackets outside of Sunday nights. I didn't even begin thoroughly examining teams outside of North Carolina and George Mason (the teams I follow most) until early February. Additionally, I was out of town during the final weekend of the season and had a provisional bracket with built-in scenarios for the games on Selection Sunday.
2005
TEAMS SELECTED: 63/65 (32/34 at-large)
PREDICTED WITHIN ONE LINE: 55/65
EXACT SEEDS PREDICTED: 31/65
Outside of the teams selected, the stat line for 2005 is flat-out gawdy. The reason why this is, ... I do not know. In fact, I was waist-deep in my graduate program and working in my spare time, so bracket projections were sloppy and poorly analyzed in comparison with the work put forth in 2007. My lack of time to view games and study the RPI amongst other indicators led me to believe Miami-Ohio and Buffalo of the Mid-American Conference were the 64th and 65th teams in the field. How little I knew.
As it stands, my "best" results were in the year that I did the least analysis, watched the fewest games, and did not have a steady set of indicators. I say "best", because best in this case does not resemble my best work. It just means that my findings for 2005 were more in line with the Tournament Selection Committee.
In closing, this signifies two things. First, it reveals that the subjective element plays a larger role than perceived or desired. Second, this meshes with the committee's desire to have its bid numbers work in concordance with conference RPI rankings. In the end, the result is that the best 65 teams will not participate in the NCAA Tournament. Nevertheless, it's the best three weeks of basketball throughout the entire year, so we'll put this on the backburner by the time that the Elite 8 rolls around.
Posts
No comments:
Post a Comment