|
Post by pOkLE on Jan 29, 2015 6:34:46 GMT -6
CoachRay, I ran a 3-2-1 scenario using last year's complete data for Coach Inman. Send me your email & I'll shoot you a copy of the summary. So what's the verdict then? Is 3-2-1 better, worse, same?
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2015 6:49:14 GMT -6
CoachRay, I ran a 3-2-1 scenario using last year's complete data for Coach Inman. Send me your email & I'll shoot you a copy of the summary. So what's the verdict then? Is 3-2-1 better, worse, same? I'll make a generalization. I have run so many scenarios that I've reached the following conclusion. In any scenario, about 90-95% (give or take) of the teams are essentially the same. The other 5-10% are where the debates come in. And so for me, the issue is what do you want to debate? Is result more important? Is strength of opponent more opponent? Each scenario, each year, etc. that 5-10% changes (and so do people's opinions based on when they put the names to it). I've got meetings most of the day, but I have a few minutes before I have to leave. I'll try to post the summaries here & you all can weigh in. I'll get as many as I can. What follows will be the 3/2/1 with 100/75/50 when compared with the 5/2.5/0 with 100/75/50.
|
|
|
Post by rora on Jan 29, 2015 6:50:01 GMT -6
A question for those with expertise in the rulebook.
A school in my district scheduled another, smaller local school for a match toward the beginning of the season. The other school then postponed the match and has subsequently postponed a couple attempted reschedules. My question is what is the recourse the school in my district can/should take. Will the match be ultimately canceled or will the other team be forced to play or forfeit? (Sorry, I cannot check LHSAA site for the originally scheduled date given that it is apparently partially down now. However, I do know that the match was in fact scheduled on the LHSAA site.)
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2015 6:54:52 GMT -6
Division I - actually this may have more variation than the 5-10%... Team | 5/2.5/0 | Rank POS | 3/2/1 | Rank POS | Change | St. Paul's | 18 | 1 | 16.14 | 1 | 0 | Brother Martin | 15.97 | 2 | 14.05 | 4 | -2 | Acadiana | 15.94 | 3 | 14.55 | 2 | 1 | Catholic BR | 15.8 | 4 | 14.15 | 3 | 1 | Captain Shreve | 14.75 | 5 | 12.91 | 7 | -2 | Lafayette | 14.45 | 6 | 13.03 | 6 | 0 | Jesuit | 14.14 | 7 | 12.64 | 10 | -3 | St. Amant | 14.02 | 8 | 12.8 | 8 | 0 | Baton Rouge | 13.95 | 9 | 13.22 | 5 | 4 | Grace King | 13.75 | 10 | 12.23 | 12 | -2 | Sulphur | 13.68 | 11 | 12.4 | 11 | 0 | Northshore | 13.67 | 12 | 12.65 | 9 | 3 | Mandeville | 13.41 | 13 | 12.19 | 13 | 0 | Destrehan | 12.45 | 14 | 11.05 | 16 | -2 | Alexandria | 12.38 | 15 | 11.88 | 14 | 1 | Dutchtown | 11.84 | 16 | 11.34 | 15 | 1 | Fontainebleau | 11.82 | 17 | 11.04 | 17 | 0 | West Monroe | 11.68 | 18 | 10.68 | 18 | 0 | Airline | 11.62 | 19 | 10.6 | 21 | -2 | Caddo Magnet | 11.52 | 20 | 10.49 | 22 | -2 | Denham Springs | 11.47 | 21 | 10.34 | 23 | -2 | Central Lafourche | 10.78 | 22 | 9.88 | 28 | -6 | East Ascension | 10.76 | 23 | 10.61 | 20 | 3 | New Iberia SH | 10.73 | 24 | 9.93 | 26 | -2 | Zachary | 10.69 | 25 | 9.95 | 25 | 0 | McKinley | 10.53 | 26 | 10.63 | 19 | 7 | Bonnabel | 10.53 | 26 | 9.23 | 33 | -7 | Comeaux | 10.43 | 28 | 9.7 | 29 | -1 | Ruston | 10.41 | 29 | 9.45 | 32 | -3 | Hahnville | 10.39 | 30 | 9.63 | 30 | 0 | Slidell | 10.39 | 30 | 9.89 | 27 | 3 | Barbe | 10.3 | 32 | 10.2 | 24 | 8 | Hammond | 9.86 | 33 | 9.55 | 31 | 2 | Covington | 9.42 | 34 | 8.97 | 35 | -1 | Pineville | 9.4 | 35 | 8.96 | 36 | -1 | C.E. Byrd | 9.15 | 36 | 8.99 | 34 | 2 | Haughton | 8.86 | 37 | 8.69 | 38 | -1 | Woodlawn | 8.7 | 38 | 8.52 | 40 | -2 | Holy Cross | 8.63 | 39 | 8.58 | 39 | 0 | Parkway | 8.57 | 40 | 8.44 | 41 | -1 | Thibodaux | 8.52 | 41 | 8.38 | 42 | -1 | Rummel | 8.47 | 42 | 8.8 | 37 | 5 | Chalmette | 8.13 | 43 | 7.73 | 45 | -2 | Ouachita | 7.91 | 44 | 7.65 | 48 | -4 | Central | 7.86 | 45 | 7.66 | 47 | -2 | John Ehret | 7.77 | 46 | 7.27 | 52 | -6 | East St. John | 7.53 | 47 | 7.6 | 49 | -2 | West Jefferson | 7.24 | 48 | 7.36 | 51 | -3 | Broadmoor | 7.23 | 49 | 6.6 | 58 | -9 | Terrebonne | 7.05 | 50 | 7.16 | 54 | -4 | Ponchatoula | 7.02 | 51 | 7.69 | 46 | 5 | Carencro | 6.94 | 52 | 7.76 | 44 | 8 | Helen Cox | 6.91 | 53 | 7.03 | 55 | -2 | Sam Houston | 6.79 | 54 | 7.79 | 43 | 11 | Natchitoches Central | 6.71 | 55 | 7.24 | 53 | 2 | South Lafourche | 6.64 | 56 | 6.79 | 57 | -1 | Shaw | 6.55 | 57 | 7.37 | 50 | 7 | Southwood | 6.5 | 58 | 6.86 | 56 | 2 | Live Oak | 6.16 | 59 | 6.56 | 59 | 0 | Walker | 6 | 60 | 6.54 | 60 | 0 | Higgins | 5.98 | 61 | 6.48 | 61 | 0 | H.L. Bourgeois | 5.23 | 62 | 6.23 | 62 | 0 |
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2015 7:06:06 GMT -6
Division II - very little movement at top (and yes Bolton was still #1) Team | 5/2.5/0 | Rank POS | 3/2/1 | Rank POS | Change | Bolton | 16.14 | 1 | 14.49 | 1 | 0 | Beau Chene | 14.27 | 2 | 13.15 | 2 | 0 | Lakeshore | 13.83 | 3 | 12.65 | 3 | 0 | St. Thomas More | 12.96 | 4 | 12.32 | 4 | 0 | Vandebilt | 12.69 | 5 | 11.69 | 5 | 0 | Neville | 12.51 | 6 | 11.12 | 7 | -1 | St. Michael's | 12.08 | 7 | 11.12 | 7 | 0 | Teurlings | 11.56 | 8 | 11.36 | 6 | 2 | Benton | 11 | 9 | 9.78 | 9 | 0 | E.D. White | 10.61 | 10 | 9.63 | 10 | 0 | Loyola | 10.6 | 11 | 9.53 | 11 | 0 | North DeSoto | 10.44 | 12 | 9.19 | 13 | -1 | East Jefferson | 10.43 | 13 | 9.39 | 12 | 1 | Lutcher | 10.38 | 14 | 9.19 | 13 | 1 | Ben Franklin | 10.16 | 15 | 9.11 | 15 | 0 | Leesville | 9.45 | 16 | 8.5 | 18 | -2 | Morgan City | 9.41 | 17 | 8.84 | 17 | 0 | DeRidder | 9.22 | 18 | 8.99 | 16 | 2 | Salmen | 8.27 | 19 | 7.58 | 20 | -1 | Plaquemine | 8.07 | 20 | 7.19 | 24 | -4 | Pearl River | 7.9 | 21 | 7.51 | 21 | 0 | Belle Chasse | 7.71 | 22 | 7.51 | 21 | 1 | Opelousas | 7.46 | 23 | 7.93 | 19 | 4 | Westgate | 7.28 | 24 | 6.92 | 27 | -3 | South Terrebonne | 7.08 | 25 | 7.27 | 23 | 2 | Riverdale | 6.71 | 26 | 6.21 | 33 | -7 | Belaire | 6.62 | 27 | 6.81 | 28 | -1 | West Ouachita | 6.56 | 28 | 6.69 | 31 | -3 | Northwood | 6.39 | 29 | 6.72 | 30 | -1 | Tara | 6.32 | 30 | 7.05 | 25 | 5 | Ellender | 6.18 | 31 | 7.01 | 26 | 5 | De La Salle | 6.08 | 32 | 6.73 | 29 | 3 | Minden | 5.88 | 33 | 6.45 | 32 | 1 | Tioga | 5.71 | 34 | 5.86 | 38 | -4 | Bastrop | 5.38 | 35 | 6.16 | 35 | 0 | Peabody | 5.21 | 36 | 6.21 | 33 | 3 | Landry-Walker | 5.14 | 37 | 5.97 | 36 | 1 | Franklinton | 5.07 | 38 | 5.9 | 37 | 1 | Wossman | 0 | 39 | 0 | 39 | 0 |
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2015 7:07:55 GMT -6
Division III - again very little movement at top Team | 5/2.5/0 | Rank POS | 3/2/1 | Rank POS | Change | Northlake Christian | 15.52 | 1 | 14 | 1 | 0 | St. Louis | 15.39 | 2 | 13.89 | 2 | 0 | University | 13.3 | 3 | 11.83 | 3 | 0 | Menard | 12.91 | 4 | 11.8 | 4 | 0 | Newman | 12.36 | 5 | 11.39 | 5 | 0 | Westminster | 12.12 | 6 | 10.43 | 8 | -2 | Episcopal BR | 11.78 | 7 | 10.6 | 7 | 0 | Dunham | 11.56 | 8 | 10.64 | 6 | 2 | Lusher Charter | 11.49 | 9 | 10.24 | 9 | 0 | Catholic NI | 10.99 | 10 | 10.14 | 10 | 0 | Country Day | 10.05 | 11 | 9.08 | 12 | -1 | St. Charles | 9.82 | 12 | 9.04 | 13 | -1 | Episcopal Acadiana | 9.81 | 13 | 9.38 | 11 | 2 | North Vermillion | 8.59 | 14 | 7.42 | 22 | -8 | Opelousas Catholic | 8.49 | 15 | 8.15 | 15 | 0 | Haynes | 8.37 | 16 | 8.09 | 17 | -1 | Runnels | 8.36 | 17 | 7.16 | 25 | -8 | Ouachita Christian | 8.24 | 18 | 7.74 | 19 | -1 | Rapides | 8.09 | 19 | 7.51 | 20 | -1 | St. Thomas Aquinas | 7.98 | 20 | 8.16 | 14 | 6 | Christian Life | 7.87 | 21 | 7.48 | 21 | 0 | St. Martins | 7.75 | 22 | 7.79 | 18 | 4 | Parkview | 7.68 | 23 | 7.27 | 24 | -1 | Lafayette Christian | 7.62 | 24 | 6.95 | 27 | -3 | Calvary | 7.55 | 25 | 6.93 | 29 | -4 | St. Frederick | 7.49 | 26 | 8.14 | 16 | 10 | Pope John Paul | 7.36 | 27 | 6.73 | 30 | -3 | Abbeville | 7.06 | 28 | 6.94 | 28 | 0 | Evangel | 7.04 | 29 | 6.54 | 33 | -4 | Ascension Episcopal | 6.68 | 30 | 7.04 | 26 | 4 | Glenmora | 6.5 | 31 | 6.39 | 35 | -4 | Riverside | 6.44 | 32 | 6.73 | 30 | 2 | David Thibodaux | 6.38 | 33 | 7.38 | 23 | 10 | Grace Christian | 6.09 | 34 | 6.47 | 34 | 0 | Erath | 6.08 | 35 | 6.62 | 32 | 3 | Bossier | 5.99 | 36 | 6.1 | 37 | -1 | West Feliciana | 5.87 | 37 | 6.02 | 38 | -1 | Catholic PC | 5.82 | 38 | 5.99 | 39 | -1 | Hannan | 5.73 | 39 | 6.33 | 36 | 3 | Thomas Jefferson | 5.11 | 40 | 5.61 | 40 | 0 | Buckeye | 5.05 | 41 | 5.5 | 41 | 0 | Houma Christian | 4.92 | 42 | 4.72 | 44 | -2 | North Caddo | 4.69 | 43 | 4.69 | 45 | -2 | Lutheran | 4.44 | 44 | 5.44 | 42 | 2 | Redemptorist | 4.39 | 45 | 5.29 | 43 | 2 | Patrick Taylor | 3.5 | 46 | 4.5 | 46 | 0 | Red River | 2.88 | 47 | 3.88 | 47 | 0 |
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2015 7:09:38 GMT -6
Those last 3 tables reflect all of 2013-14 season comparing whether we use 5 W, 2.5 L, and 0 T or change it to 3 W, 2 L, and 1 T. Both calculations still include 100%/75%/50% of opponents' wins.
Gotta roll...
|
|
|
Post by pOkLE on Jan 29, 2015 7:31:50 GMT -6
But it's not the variation that matters in evaluating, right? I guess it takes more than raw numbers to evaluate whether the system more accurately awards points based on result (w, l, or t).....in other words, I'm waiting on reading the number gurus' analyses.
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2015 14:49:20 GMT -6
But it's not the variation that matters in evaluating, right? I guess it takes more than raw numbers to evaluate whether the system more accurately awards points based on result (w, l, or t)...in other words, I'm waiting on reading the number gurus' analyses. You're right (and I don't tell you that often). Variation doesn't matter. My point in showing it that way is to show the similarities between the 2 models. Let's use Division II. 9 of the top 11 wouldn't have changed their seeding under either model. I look at the big drops to see how I feel about those (Riverdale would drop 7 spots & Opelousas would go up 4 spots). The question then becomes which is "right" - where did they deserve to be? My conclusion is that none of them are perfect, but they all get somewhat similar results. In this case, the 3/2/1 tends to reward strength of schedule a little more than the 5/2.5/0 model does. Is that right? I don't know. That's why I show the comparisons. And those of you that have seen more of these teams can debate it. But in the end, the top 10 isn't changing much.
|
|
|
Post by methuselah on Jan 29, 2015 15:19:43 GMT -6
Ok, maybe this next point should be filed under navel-gazing because it's on the edges. But here's something I noticed dealing with power points and ties between opponents both of whom a team has played and beaten and it got me curious. It goes to the "a tie is a tie" rule, even if PKs are used for district standings. It's something that popped into my head by seeing something in the "scores" thread about a tie which was broken (for district standings only, but not for record) by PKs.
Team A plays team B and wins. Team A plays team C and wins.
Team B&C play and tie.
Does team A get 1/2 a point for team B's tie and 1/2 a point for team C's tie? Which would add up to the same 1 point they would get if either team had "won".
Or does team A get 75% for team C's tie and 75% for team C's tie? Which would add up to 1.5 points and thus be more than if either team had won.
I seem to remember reading that a tie is 1/2 a win but also seem to remember something about 75% and I'm not sure what the context for each of those is.
|
|
|
Post by rlb2024 on Jan 29, 2015 15:37:14 GMT -6
Ok, maybe this next point should be filed under navel-gazing because it's on the edges. But here's something I noticed dealing with power points and ties between opponents both of whom a team has played and beaten and it got me curious. It goes to the "a tie is a tie" rule, even if PKs are used for district standings. It's something that popped into my head by seeing something in the "scores" thread about a tie which was broken (for district standings only, but not for record) by PKs. Team A plays team B and wins. Team A plays team C and wins. Team B&C play and tie. Does team A get 1/2 a point for team B's tie and 1/2 a point for team C's tie? Which would add up to the same 1 point they would get if either team had "won". Or does team A get 75% for team C's tie and 75% for team C's tie? Which would add up to 1.5 points and thus be more than if either team had won. I seem to remember reading that a tie is 1/2 a win but also seem to remember something about 75% and I'm not sure what the context for each of those is. Team A would get 1/2 point for each tie -- same as if one team had won and the other had lost. It's pretty much a wash for district play. Where it matters is for a non-district opponent (say, Team Z) who plays Team B but not Team C; whether the game is a win or a tie would then have an effect on Team Z's PR.
|
|
|
Post by chelsea007 on Jan 29, 2015 19:15:45 GMT -6
I would assume team A would get a half of a point from each (100% of an opponents win total) for that particular draw plus 3 points for each for the two wins. I believe strength of schulde should play a role. 3,2,1 seems to do a better job with that imo than 5,2.5,0.
|
|
|
Post by futboldad on Jan 31, 2015 12:37:25 GMT -6
I still find myself confused at to the 15 game issue. If I read correctly, if you are in the top 32 for power rating you will still make the playoffs if you have not played 15 games but you will be seeded behind teams that have played 15. If you don't have 15 games, you should not have an easier path to playoffs. This could be solved by only using the highest 15 game point totals of a team and dividing by 15. If you played less than 15 games then your number is STILL divided by 15. All of the top 32 makes playoffs unless a district champ/co champ was ranked lower than 32 (very rare) and would knock 32 out. Where the 15 games requirement comes in, is for seeding. If yiu didn't play 15, you go behind all others in the 32 team field already decided and qualified by power ratings and your SEED is lowered to behind all of those other teams who met the required 15 game minimum. This is how it's written and enforced and was reconfirmed to us at a meeting in January (By Gary Duhe and Lacey McDirmid) as I said earlier in thread BCCoach, Just to make sure I understand the new seeding for playoffs, it seems to me that all that is happening is that the power rankings are replacing the coach's polls for seeding. The top eight spots will still get filled by district winners--by power ranking order--despite a district runner-up in a very strong district--such as EJ--having a better power ranking than a winner in some other district. I would bet that the coach's poll does a better job in predicting the outcomes of palyoff games than the power rankings. Power rankings may seem more objective, but they just do not included as much information as coaches have about other teams and about key members of other teams, the other coaches themselves.
|
|
|
Post by futboldad on Jan 31, 2015 16:51:12 GMT -6
But it's not the variation that matters in evaluating, right? I guess it takes more than raw numbers to evaluate whether the system more accurately awards points based on result (w, l, or t).in other words, I'm waiting on reading the number gurus' analyses. You're right (and I don't tell you that often). Variation doesn't matter. My point in showing it that way is to show the similarities between the 2 models. Let's use Division II. 9 of the top 11 wouldn't have changed their seeding under either model. I look at the big drops to see how I feel about those (Riverdale would drop 7 spots & Opelousas would go up 4 spots). The question then becomes which is "right" - where did they deserve to be? My conclusion is that none of them are perfect, but they all get somewhat similar results. In this case, the 3/2/1 tends to reward strength of schedule a little more than the 5/2.5/0 model does. Is that right? I don't know. That's why I show the comparisons. And those of you that have seen more of these teams can debate it. But in the end, the top 10 isn't changing much. The real question about what type of power ranking should be used (or if something else, such as coaches' poll) for seeding is rally a matter of what do we want the seeding to do. IMO, tournament or playoff seeding should reflect or rather predict tournament ranking outcomes. Which method does this. In statistical forecasting, there is a way to measure the power of a measure in forecasting some event. See en.wikipedia.org/wiki/Root-mean-square_deviation.
|
|
nolapelota
All-District
No longer "booming,,,"
Posts: 203
|
Post by nolapelota on Jan 31, 2015 17:04:56 GMT -6
All of the top 32 makes playoffs unless a district champ/co champ was ranked lower than 32 (very rare) and would knock 32 out. Where the 15 games requirement comes in, is for seeding. If yiu didn't play 15, you go behind all others in the 32 team field already decided and qualified by power ratings and your SEED is lowered to behind all of those other teams who met the required 15 game minimum. This is how it's written and enforced and was reconfirmed to us at a meeting in January (By Gary Duhe and Lacey McDirmid) as I said earlier in thread .The top eight spots will still get filled by district winners--by power ranking order--despite a district runner-up in a very strong district--such as EJ--having a better power ranking than a winner in some other district.. Whoa whoa whoa.. NO NO NO, the district winners do NOT get the first eight ranking. They are simply guaranteed into the playoffs, but their seeding is determined by the PR. No, a district winner is not seeded in the top 8.
OMG... D-III has a two-team district and several that only have one decent team. It also has two or three districts that the 4th place team would win other districts. The insanity of the district composition is one reason the district system is not respected.
Make 8-9 team districts, balance each district with strong and weaker programs, require home-home, suddenly district means something, for scheduling, for Power Ranking, and for evening out the regional advantage that is so obvious because of soccer strength distribution.
|
|
|
Post by futbolfiend on Jan 31, 2015 17:15:08 GMT -6
When calculating pr ..?
Team a beats team b
Team b record is 10--1-4
Team a gets 5 pts for win over team b + 100% of their win total which is ?
10 or 12 (each tie counts .5 win)..?
Sent from my SM-G900P using proboards
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 31, 2015 17:47:46 GMT -6
When calculating pr ..? Team a beats team b Team b record is 10--1-4 Team a gets 5 pts for win over team b + 100% of their win total which is ? 10 or 12 (each tie counts .5 win)..? Sent from my SM-G900P using proboards 12 so 17 points total
|
|
|
Post by futboldad on Jan 31, 2015 18:05:16 GMT -6
.The top eight spots will still get filled by district winners--by power ranking order--despite a district runner-up in a very strong district--such as EJ--having a better power ranking than a winner in some other district.. Whoa whoa whoa.. NO NO NO, the district winners do NOT get the first eight ranking. They are simply guaranteed into the playoffs, but their seeding is determined by the PR. No, a district winner is not seeded in the top 8.
OMG. D-III has a two-team district and several that only have one decent team. It also has two or three districts that the 4th place team would win other districts. The insanity of the district composition is one reason the district system is not respected.
Make 8-9 team districts, balance each district with strong and weaker programs, require home-home, suddenly district means something, for scheduling, for Power Ranking, and for evening out the regional advantage that is so obvious because of soccer strength distribution.
Thanks for clearing that up for me, nolapelota. I am really glad to hear it. I had done a search on "seeding" and "power ranking" and this thread was all I found. Seeding, IMO, should be about better teams, giving those at the bottom some chance, but also having records and strength of schedule mean something. The seeding should reflect some prediction of the way the tournament will turn out. BTW, I am so glad that next year we will be going to 4 somewhat equal-sized divisions with 24 teams making the playoffs in each division. In D-II this year we will have 32 teams in the playoffs but only 39 teams in the division. I am pretty sure that there will be teams making the playoffs who have not won any district games. We may as well just have a division open tournament.
|
|