|
Post by flat3 on Nov 22, 2006 17:03:07 GMT -6
I was calculating Computer Rankings during the last season, but it had problems due to subjective "start values." Without every game result being posted, we have to depend on some kind of initial "ranking" to calculate anything on computer. With efforts by coachray40, putmeincoach2, and other people, most of early scores have been posted. If this continues, we can calculate completely objective "Power Ratings" just like we do for football. I have been working on the formula a little. While I can't guarantee if I will calculate Power Ratings for the entire season, I might want to try for a while if we can agree on the formula. I would like to get some inputs from the members before spending any more time. It's a little complicated, but please bear with me. Here is the Power Ranking formula for FOOTBALL (from www.lhsaa.org/sports/football/footballby-laws.pdf ). 1. You get 10 points for a win. 2. You add the number of wins by your opponent. 3. You add 2 points for every class you play up. 4. Rankings change every week reflecting the results of all opponents in the past. 5. Your power ranking is the average score per game. An example: If a 3A school beats a 5A school with a record of 3-4, you get 10 (win) + 3 (# of win by the opponent) + 4 (playing up) = 17 points. If this 5A school wins two more games during the season, the score goes up to 19 points (two more wins by the opponent). My initial plan for soccer is as follows. 1. You get 10 points for a win. 2. You get 5 points for a tie. 3. You add the winning proportion x 10 by your opponent. For winning proportion, ties count one half win. You can't use the number of wins in soccer since every school plays a different number of games. 4. You add 2 points for every class you play up. 5. You add 2 points for a shutout. 6. You add 2 points for a 3+ goal differential. 7. Ratings change every time one of your opponents plays a game. An example: University tied St. Amant last night. St. Amant's record is 2-0-1 (2.5 wins for 3 games, or .833 winning proportion). University is DIII and St. Amant is DI. The Cubs earned 5 points for a tie, 8.33 points for St. Amant record (quality point), and 4 points (2 points per class) for playing up. Their total for the game was 17.33. If St. Amant wins the next game, their record becomes 3-0-1, winning proportion is .875 (3.5/4), and UHS' score goes up to 17.75. Another example: St. Amant beat Episcopal 3-0 earlier. EHS' record is 2-1 (.667 proportion, or 6.67 quality points). The Gators earned 10 points for a win, 6.67 points for EHS' record, 2 points for a shutout, and 2 points for 3 goal differential. Their total for the game was 20.67. If Episcopal loses the next game, their record becomes 2-2 (.50 proportion, or 5.00 quality points), and St. Amant's score goes down to 19. I tentatively set up an Excel work sheet to reflect most results in the BTR metro area. I can report preliminary results after the Bayou Invitational Tournament. My questions are as follows. 1. Do we need to give 2 points for playing up every class? This gives advantages to DIII and DII schools, particularly if we compare them against DI schools. This adjustment, however, is a good one if you want to make the system similar to football (so that we may propose LHSAA to adopt the system for a playoff purpose). 2. Do we need to give 2 points for a shutout? I personally think there is a large difference between 2-0 win and 4-2 win. We give some points for shutouts in many tournaments. 3. Do we need to give 2 points for a 3+ goal differential? I think there is a large difference between 5-1 win and 2-1 win. We give some points for 3+ goals differentials in many tournaments. We should think as follows. Agains the same opponent, loss is 0 point, tie is 5 points, and win is 10 points. How many more points, if any, would you like to give for a shutout and a blowout? Is a 3-0 (or 8-0) win worth 14 points? Your inputs will be much appreciated.
|
|
|
Post by PutMeInCoach2 on Nov 22, 2006 17:33:44 GMT -6
I don't think playing up in a Division should recieve points. Here is why. A football team has an advanatage because they have more of a pull to pull from. However they also need more players to play a game. I think a school of 300, is just as equal to a school of 1500. It will depend on the feeder recreation/primier that feeds that program either way. For example. St. Louis would be expected to beat a Morgan City. Therefore it makes no since to award them with points.
I do think points should be given for a shutout and goal differential up to 3.
My one question would be a penalty kick loss, should they recieve 2.5 points, and the opposing team recieve 10, it was just a thought. I really have no clue...
|
|
|
Post by flat3 on Nov 22, 2006 17:50:16 GMT -6
I don't think playing up in a Division should recieve points. Here is why. A football team has an advanatage because they have more of a pull to pull from. However they also need more players to play a game. I think a school of 300, is just as equal to a school of 1500. It will depend on the feeder recreation/primier that feeds that program either way. For example. St. Louis would be expected to beat a Morgan City. Therefore it makes no since to award them with points. OK, your point is well taken. I included this primarily due to a consistency with football. I do think points should be given for a shutout and goal differential up to 3. Good. I assume you think 0 point for a loss, 5 points for a tie, 10 points for a win, and 2 points each for a shutout and a blowout are reasonable. My one question would be a penalty kick loss, should they recieve 2.5 points, and the opposing team recieve 10, it was just a thought. I really have no clue... A good question. I didn't think of that. I would say 7.5 for winner and 2.5 for loser, or 7 and 3.
|
|
|
Post by soccerguy2628 on Nov 22, 2006 20:05:36 GMT -6
i think you all have way too much time on your hands...
|
|
|
Post by lakeview on Nov 22, 2006 20:25:00 GMT -6
I'd go with your initial plan and see how it works this season. Thanks for taking the time to do this. Kudos also to the sucky player number 30 and coachray for all the time keeping up with scores/standings.
|
|
|
Post by flat3 on Nov 22, 2006 21:27:44 GMT -6
i think you all have way too much time on your hands... So what?
|
|
guru
All-District
Posts: 111
|
Post by guru on Nov 22, 2006 22:30:46 GMT -6
I don't think playing up in a Division should recieve points. Here is why. A football team has an advanatage because they have more of a pull to pull from. However they also need more players to play a game. I think a school of 300, is just as equal to a school of 1500. It will depend on the feeder recreation/primier that feeds that program either way. For example. St. Louis would be expected to beat a Morgan City. Therefore it makes no since to award them with points.... Put... I disagree here. In flat 3's system you will gain points for playing a team in a higher division but, if that team is weak and has few wins the net effect may actually hurt you in the long run. Yes D3 schools can get a higher power rating than D1 schools but remember, Flat3's system is for ranking a team for the playoffs (division specific) not for ranking all teams in all divisions against each other. If St. Louis can beat a top ranked D-1 team they deserve extra points (compared to other D-3 Schools) for beating a quality D-1 school. Feeder systems are as much an issue in football as they are in soccer. There are many large 5A schools whose football programs are terrible. The simple fact of the matter is that the larger the school the greater the number of athletes that program has to choose from. Therefore there should be some reward for playing up against quality opponents.
|
|
|
Post by flat3 on Nov 22, 2006 23:31:39 GMT -6
Flat3's system is for ranking a team for the playoffs (division specific) not for ranking all teams in all divisions against each other. I was thinking about the same thing, but I was not sure about this. I might want to kill two birds with one stone. In other words, I would like to use this rating system not only for division specific comparisons, but also for across all divisions. The comparison between University and St. Amant is an interesting exercise. They tied with each other, and they have identical record (2-0-1). According to any objective rating, they should be rated equally, since we can't consider each school's other opponents (Parkview and McKinley for University and Episcopal and Tara for St. Amant). St. Amant gained 13.33 and University gained 17.33 (4 more points for playing up). This is INCORRECT if we compare University (DIII) and St. Amant (DI) against each other. If we use the power rating only within each division as guru says, points for playing up make sense. Since I would like to use this rating across divisions, I am leaning toward getting rid of playing-up advantages. I thank guru for noting this important point.
|
|
|
Post by wrightphillips2 on Nov 23, 2006 10:01:39 GMT -6
if u get points for playing up. that screws a D1 team from not getting max points because they can enver play "UP"
|
|
|
Post by soccerboy07 on Nov 23, 2006 10:54:04 GMT -6
I don not think puts should be awarded for playing up. Points awarded for beating a team of a good record should be enough to tell if the team is good. Points being awarded up would only give lower division schools more points for possibly beating worse a higher division team.
|
|
|
Post by Punkaro on Nov 23, 2006 11:08:52 GMT -6
i like it
|
|
guru
All-District
Posts: 111
|
Post by guru on Nov 23, 2006 11:50:17 GMT -6
Flat3's system is for ranking a team for the playoffs (division specific) not for ranking all teams in all divisions against each other. I was thinking about the same thing, but I was not sure about this. I might want to kill two birds with one stone. In other words, I would like to use this rating system not only for division specific comparisons, but also for across all divisions. The comparison between University and St. Amant is an interesting exercise. They tied with each other, and they have identical record (2-0-1). According to any objective rating, they should be rated equally, since we can't consider each school's other opponents (Parkview and McKinley for University and Episcopal and Tara for St. Amant). St. Amant gained 13.33 and University gained 17.33 (4 more points for playing up). This is INCORRECT if we compare University (DIII) and St. Amant (DI) against each other. If we use the power rating only within each division as guru says, points for playing up make sense. Since I would like to use this rating across divisions, I am leaning toward getting rid of playing-up advantages. I thank guru for noting this important point. Flat3 If you want to compare across divisions eliminate the points given for playing up. I still would like to see you keep the bonus points for rating teams within their own division. I think the overriding purpose here is to give the lhsaa a new system for placing teams in brackets for the playoffs. If you are up to it I think running 2 different sets of ratings, (one for all teams with no playing up points and one with playing up points for division bracket placement) is the way to go. Either way I appreciate your work and am glad to see someone interested in taking a scientific approach to solving our playoff bracket problems.
|
|
|
Post by numeroquattro on Nov 23, 2006 15:02:59 GMT -6
if you use this system what will establish a good team from a bad team for the wins against good team and wins against bad teams?
|
|
|
Post by soccerstud722 on Nov 23, 2006 21:00:42 GMT -6
I don't think playing up in a Division should recieve points. Here is why. A football team has an advanatage because they have more of a pull to pull from. However they also need more players to play a game. I think a school of 300, is just as equal to a school of 1500. It will depend on the feeder recreation/primier that feeds that program either way. For example. St. Louis would be expected to beat a Morgan City. Therefore it makes no since to award them with points.... Put... I disagree here. In flat 3's system you will gain points for playing a team in a higher division but, if that team is weak and has few wins the net effect may actually hurt you in the long run. Yes D3 schools can get a higher power rating than D1 schools but remember, Flat3's system is for ranking a team for the playoffs (division specific) not for ranking all teams in all divisions against each other. If St. Louis can beat a top ranked D-1 team they deserve extra points (compared to other D-3 Schools) for beating a quality D-1 school. Feeder systems are as much an issue in football as they are in soccer. There are many large 5A schools whose football programs are terrible. The simple fact of the matter is that the larger the school the greater the number of athletes that program has to choose from. Therefore there should be some reward for playing up against quality opponents. I disagree....I think that The difference between the top D1 and D3 schools are very small. Therefore if say Teulings Beats Acadiana, they Get a siginifacntly more amount of points as opposed to if Acadiana beet Teurlings, when in the long run I think Teurlings would be favorite. Take out the Another reason is that Div II is moer weak as a whole than the other2 divisions, so we would be awarding points to D 3 teams for beating less quality D 2 teams who mop up in their respective division. Makes no sense. Keep the Divisions out of it and I think you will have a good overall, non-objective ranking system.
|
|
|
Post by flat3 on Nov 26, 2006 19:52:25 GMT -6
I have been updating the scores for the power rating purpose. I have most results except many in Rebels Cup (please supply the results to coachray40, if you know any unreported results). I will show the first ratings as soon as this week's Top 10 is released. I don't want to influence anyone's votes this week. I made a couple of changes since I first presented the formula. Here is the revised formula. 1. Wins earn 10 points, and ties get 5 points. Shootout wins earn 7 points and shootout losses count 3 points. 2. There is no "play-up advantage," since I would like to compare teams across divisions. People's opinions seem to favor no advantage, also. 3. Whether win, lose, or tie, you will earn the quality point (10 x opponent's win proportion). 4. Winners get 1 point for shutouts, 1 point for mild blowouts (2 goal differentials), and another point for blowouts (3+ goal differentials). 5. I decided to subtract the same points for shutout and/or blowout losses. After all, losing to Jesuit 1-2 and 0-7 should make a difference in rankings. The maximum points per game is 10 (win) + 1 (shutout) + 2 (blowout) + 9 (if the opponent has .90 winning proportion) = 22. The minimum per game is 0 (loss) - 1 (shutout) - 2 (blowout) + 1 (if the opponent has .10 winning proportion) = -2. If you lose to an undefeated team 1-2, you earn 0 (loss) + 10 (10 x 1.0 winning proportion) = 10 points. If you beat a winless team by 2-1, you earn 10 (win) + 0 (.00 winning proportion) = 10 points. For this week, the largest per game point ("best win") was earned by Newman over Covington (3-0, earning 21 points). The smallest point ("worst loss") was suffered by Holy Cross (lost 0-3 to Central Lafourche, earning -.5 point).
|
|
|
Post by PutMeInCoach2 on Nov 26, 2006 19:58:52 GMT -6
all votes are in and have been tallied... I am just waiting for the release time... So you can release when you want, or wait til 9pm like we do...
|
|
|
Post by flat3 on Nov 26, 2006 20:23:50 GMT -6
Here is the result for the first Power Rating scores. Those teams with three or fewer games didn't get rankings, since they are not reliable. If you know discrepancies (most likely unreported results), please PM to myself AND coachray40. The team records are based on coachray40's result thread. We need to have every score reported to make these rankings valid. For the following results, games vs JV or scrimmage results are not included. The records of out-of-state teams are to my best knowledge. St. Louis' score would be 17.3 if we don't count their three losses against three Texas teams (they still won't be ranked since they had only two games vs LA teams). Any comments will be appreciated. No trash talk please. They are rated according to the system provided on top of Page 2 of this thread.
Listed here are: Rank, Power Rating, School (# in parentheses are Divisions), Records. * 19.67 Caddo 1 0 0 1 16.91 Jesuit 6 0 0 2 16.63 Newman (3) 4 0 2 3 15.24 Teurlings (3) 7 0 1 4 15.23 Bro. Martin 5 0 1 * 14.83 Neville (2) 2 0 0 * 14.42 Byrd 2 0 0 5 14.27 St. Amant 3 0 3 6 14.08 Grace King 3 0 1 7 13.74 Fountainebleau 4 1 0 8 13.73 Ponchatoula 5 1 0 * 13.50 Slidell 1 0 0 9 13.37 Dutchtown 4 0 2 * 13.33 Loyola (2) 3 0 0 10 13.14 Woodlawn 5 1 1 ** 12.99 Pineville (2) 1 1 0 11 12.93 Lafayette 2 1 1 12 12.81 Rummel 5 1 1 13 12.63 Hammond 3 1 0 ** 12.62 NISH 2 0 1 14 12.53 Carencro 3 1 0 15 12.49 Shreve 3 2 1 16 12.38 Tara (2) 3 1 0 17 12.33 East St. John 3 1 0 18 12.20 Menard (3) 5 1 1 ** 12.13 Chalmette (2) 1 1 0 19 12.05 St. Thomas More (2) 4 2 1 ** 12.00 Airline 2 1 0 20 11.94 Acadiana 2 2 0 21 11.70 Catholic 2 1 3 22 11.49 University (3) 3 1 2 23 11.46 Barbe 2 1 1 24 11.45 DeRidder (2) 6 1 1 ** 11.33 Parkway (2) 2 1 0 25 11.27 Covington 4 1 0 ** 11.20 ASH (2) 2 1 0 26 11.00 Episcopal (3) 3 2 2 ** 10.86 Westgate 0 1 0 27 10.72 St. Louis (3) 2 3 0 28 10.37 Baton Rouge 2 2 1 29 10.28 Terrebonne 1 1 3 30 10.27 Mandeville 2 1 2 31 10.04 Comeaux 2 2 1 32 9.83 Vandebilt Cath. (2) 2 3 1 33 9.81 St. Paul's 2 4 0 34 9.55 Lee (2) 2 2 0 ** 9.50 Haughton (2) 1 1 0 35 9.10 E. Ascension (2) 2 3 0 ** 8.95 Parkview (3) 1 2 0 ** 8.83 West Monroe 1 1 0 ** 8.67 Salmen (2) 0 1 1 36 8.63 Northshore 2 2 1 37 8.50 Central 2 2 0 ** 8.42 Natch. Central 1 1 0 38 8.34 Beau Chene (2) 2 2 0 39 8.13 St. Thomas Aquinas (3) 1 3 0 ** 8.13 St. Frederick (3) 0 1 0 40 8.07 Opelousas (2) 1 2 1 41 8.07 St. Michael (2) 2 3 1 42 8.06 Destrehan 2 3 1 43 7.90 Ben Franklin (2) 1 3 0 44 7.78 Sulphur 2 3 1 45 7.38 St. Martin's (3) 1 3 0 ** 7.17 East Jefferson (2) 1 1 1 ** 7.13 Bolton (3) 0 1 0 46 7.03 Hahnville 1 4 1 ** 7.00 Bossier (2) 0 2 0 47 6.92 C. Lafourche 1 4 1 ** 6.75 West Jefferson 0 1 0 ** 6.75 Pope John Paul (3) 0 1 1 ** 6.33 Northwood (2) 0 2 0 48 6.21 E. D. White (3) 1 3 0 ** 6.13 Independence (3) 0 1 0 ** 6.08 Westminster (3) 1 2 0 ** 6.06 Catholic PC (3) 1 2 0 ** 5.88 Shaw 0 2 0 ** 5.67 Opelousas Cath. (3) 0 1 0 ** 5.33 Abbeville (2) 0 1 0 ** 5.13 Tioga (2) 0 1 0 ** 5.10 St. Charles (3) 0 2 0 ** 5.08 HLB 0 2 0 ** 5.00 DeLaSalle (2) 0 1 0 ** 4.92 Ellendar 0 2 0 ** 4.86 Episcopal, Acadiana (3) 0 3 0 ** 4.86 S. Lafourche 0 1 0 49 4.67 W. Feliciana (3) 0 4 0 ** 4.56 Calvary (3) 0 3 0 ** 4.50 Huntington 0 2 0 50 4.17 Northlake Ch. (3) 0 4 0 ** 4.08 McKinley (2) 0 3 0 ** 4.05 Leesville (2) 0 2 0 51 4.00 Christian Life (3) 0 5 0 ** 3.79 Notre Dame (3) 0 3 0 ** 3.67 Southwood 0 1 0 ** 3.61 Zachary (2) 0 2 0 ** 3.00 Broadmoor (2) 0 1 0 ** 2.94 Redemptorist (3) 0 3 0 ** 2.86 Northside (2) 0 3 0 ** 2.08 Sam Houston (2) 0 2 0 ** 1.67 Holy Cross 0 3 0
|
|
|
Post by coachray40 on Nov 26, 2006 20:26:10 GMT -6
So cool.
Interesting numbers, especially with Grace King traveling to St Amant on Thursday.
I cant believe any team is lower than Zachary right now--thank heaven 2 of their 4 losses came to JV teams or they would be rated 2 pages further down.
|
|
|
Post by wrightphillips2 on Nov 26, 2006 20:46:46 GMT -6
aare those all the teams in the state?
|
|
|
Post by PutMeInCoach2 on Nov 26, 2006 20:48:37 GMT -6
aare those all the teams in the state? no
|
|