Bluefan Posted June 23, 2006 Share Posted June 23, 2006 Okay, here is what this is: I took each corps last score from the 2005 season and subtracted it by their first score. I divided this number by the number of days that corps was on tour (not the number of days tour lasted). That gives the number of points each corps improved per day. Then I took each corps most recent score from this season and extrapolated to finals night using last years omprovement rate. Obviously there are many things wrong with this method, but this is just for fun. I will be examining more scores from throughout the season from the past few years and toying around with regression patterns and randomizing factors such as staff changes to adjust, but that will come later. So here are the predictions as of right now. 15. Pioneer - 67.35 14. Magic - 74.45 13. Mandarins - 75.1 12. Pacific Crest - 75.4 11. Southwind - 80.2 10. Colts - 87.15 9. Santa Clara Vanguard 88.728 8. Glassmen - 89.1 7. Blue Knights - 90.45 6. Madison - 92.15 5. Blue Devils - 94.3 4. Phantom Regiment - 98.1 3. Cavaliers - 98.15 2. Bluecoats - 98.7 1. Blue Stars - 99.6 *Blue Stars prediction obviously quite skewed from D2 scores last year used in calculation. Any ideas how to adjust? I was thinking about looking at scoring trends from previous corps that have made the switch and coming up with some coefficient, simple but effective. Quote Link to comment Share on other sites More sharing options...
Bluefan Posted June 23, 2006 Author Share Posted June 23, 2006 (edited) Title typo fixed by Krista, many thanks Edited June 23, 2006 by Bluefan Quote Link to comment Share on other sites More sharing options...
dciguy01 Posted June 23, 2006 Share Posted June 23, 2006 (edited) interesting how your equations always have the Bluecoats on top* Cool equation though, I like :) Edited June 23, 2006 by dciguy01 Quote Link to comment Share on other sites More sharing options...
dbc03 Posted June 23, 2006 Share Posted June 23, 2006 Okay, here is what this is:I took each corps last score from the 2005 season and subtracted it by their first score. I divided this number by the number of days that corps was on tour (not the number of days tour lasted). That gives the number of points each corps improved per day. Then I took each corps most recent score from this season and extrapolated to finals night using last years omprovement rate. Obviously there are many things wrong with this method, but this is just for fun. I will be examining more scores from throughout the season from the past few years and toying around with regression patterns and randomizing factors such as staff changes to adjust, but that will come later. So here are the predictions as of right now. 15. Pioneer - 67.35 14. Magic - 74.45 13. Mandarins - 75.1 12. Pacific Crest - 75.4 11. Southwind - 80.2 10. Colts - 87.15 9. Santa Clara Vanguard 88.728 8. Glassmen - 89.1 7. Blue Knights - 90.45 6. Madison - 92.15 5. Blue Devils - 94.3 4. Phantom Regiment - 98.1 3. Cavaliers - 98.15 2. Bluecoats - 98.7 1. Blue Stars - 99.6 *Blue Stars prediction obviously quite skewed from D2 scores last year used in calculation. Any ideas how to adjust? I was thinking about looking at scoring trends from previous corps that have made the switch and coming up with some coefficient, simple but effective. Looks about right to me Quote Link to comment Share on other sites More sharing options...
Krista Posted June 23, 2006 Share Posted June 23, 2006 Oh, and I'm aware I didn't spell mathematical right, how emmbarrasing fixed. :) Quote Link to comment Share on other sites More sharing options...
BaritoneUnit Posted June 23, 2006 Share Posted June 23, 2006 Bluefan, you are a stats master! Too bad you'll have to do it again tomorrow after the Cadets and Crown have their first scores. :P Quote Link to comment Share on other sites More sharing options...
jblamb1401 Posted June 23, 2006 Share Posted June 23, 2006 Just a thought: Wouldn't it be better to find the average improvement so far for 2006 since the squads for the majority are completly different. Quote Link to comment Share on other sites More sharing options...
tubacurt Posted June 23, 2006 Share Posted June 23, 2006 Okay, here is what this is:I took each corps last score from the 2005 season and subtracted it by their first score. I divided this number by the number of days that corps was on tour (not the number of days tour lasted). That gives the number of points each corps improved per day. Then I took each corps most recent score from this season and extrapolated to finals night using last years omprovement rate. Obviously there are many things wrong with this method, but this is just for fun. I will be examining more scores from throughout the season from the past few years and toying around with regression patterns and randomizing factors such as staff changes to adjust, but that will come later. So here are the predictions as of right now. 15. Pioneer - 67.35 14. Magic - 74.45 13. Mandarins - 75.1 12. Pacific Crest - 75.4 11. Southwind - 80.2 10. Colts - 87.15 9. Santa Clara Vanguard 88.728 8. Glassmen - 89.1 7. Blue Knights - 90.45 6. Madison - 92.15 5. Blue Devils - 94.3 4. Phantom Regiment - 98.1 3. Cavaliers - 98.15 2. Bluecoats - 98.7 1. Blue Stars - 99.6 *Blue Stars prediction obviously quite skewed from D2 scores last year used in calculation. Any ideas how to adjust? I was thinking about looking at scoring trends from previous corps that have made the switch and coming up with some coefficient, simple but effective. oh math, how hilarious thee are. I found this pretty funny, obviously the thing wrong with it is that it doesn't take into effect corps that start real good and thus have less room to grow, but you already knew that. I like the 7 point gap between 10th and 11th, and the 4 point gap between 4th and 5th. Quote Link to comment Share on other sites More sharing options...
Bluefan Posted June 23, 2006 Author Share Posted June 23, 2006 Just a thought:Wouldn't it be better to find the average improvement so far for 2006 since the squads for the majority are completly different. Since corps always improve faster early in the season, that would put many corps well over 100 (in fact, the Bluecoats would win in that experiment too with a score of 110.5). Now from there you could scale it, but I'd prefer to take into account the mid-season lull. I will eventually be coming up with formulas that take this seasons progress thus far into consideration (after all, this season is the one that counts). Using historical scores is simply a way to provide a model for points later on in the summer. Quote Link to comment Share on other sites More sharing options...
jblamb1401 Posted June 23, 2006 Share Posted June 23, 2006 Pretty cool... kinda like the "grinder" in the blue devils' forum Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.