p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 2, 2014 11:10:36 GMT -6
I was asked in the last few days about the following example and wanted to share it with you as to how the PR can be skewed during the season.
The question to me was as to how Zachary is rated ahead of Denham Springs in Power Ratings? Denham Springs PR is 6.705 and Zachary is 7.021.
Now I use these 2 teams because I already did the work in helping to answer the question. I do get asked similar questions regularly so it really could be any set of 2 teams. The facts at the time... Denham Springs: 1) was undefeated in district, 2) had more PR wins (than Zachary), 3) had a better record (than Zachary), and 4) beat Zachary head-to-head
The short answer: Zachary beat Grace King on November 24. Since King had 16 PR Wins (as of a few days ago), that earned Zachary a 21.000 for that game. However, if I were to exclude that one win over King, Zachary's PR would only be 5.750. So they get lots of credit for beating a win-filled team.
This is an example of a game that will smooth out some over time. Right now that 21.000 is being averaged over 12 games. In the absence of another big win, by the end of the season, that 21 will be smoothed over 20-22 games (cutting it's impact almost in half). If they pull a 5.750 in each of the next 10 games, their PR would drop to 6.443 (behind Denham Springs). Or if they win another similar game, it will help keep their PR up.
Granted, Grace King will likely win some more games, but so will opponents of both of these squads will so I didn't try to factor all of that in. Simply trying to show how the averages can be skewed during the early & middle part of the season (and yes, sometimes still at the end of the season for numerous reasons).
|
|
|
Post by methuselah on Jan 2, 2014 11:59:58 GMT -6
It looks like Zachary has opted for a very challenging schedule. Catholic, Dutchtown, Baton Rouge, Grace King, E.D. White, St. Mikes, etc. And they got a result gainst King. Sounds reasonable to me.
- A bit off topic but Saturday's Zachary v St.Mikes, and Monday's Denham Springs v. Lutcher games should both be good ones.
|
|
|
Post by methuselah on Jan 7, 2014 7:32:44 GMT -6
. . . The question to me was as to how Zachary is rated ahead of Denham Springs in Power Ratings? Denham Springs PR is 6.705 and Zachary is 7.021. Now I use these 2 teams because I already did the work in helping to answer the question. I do get asked similar questions regularly so it really could be any set of 2 teams. The facts at the time... Denham Springs: 1) was undefeated in district, 2) had more PR wins (than Zachary), 3) had a better record (than Zachary), and 4) beat Zachary head-to-head The short answer: Zachary beat Grace King on November 24. Since King had 16 PR Wins (as of a few days ago), that earned Zachary a 21.000 for that game. However, if I were to exclude that one win over King, Zachary's PR would only be 5.750. So they get lots of credit for beating a win-filled team. This is an example of a game that will smooth out some over time. Right now that 21.000 is being averaged over 12 games. In the absence of another big win, by the end of the season, that 21 will be smoothed over 20-22 games (cutting it's impact almost in half). If they pull a 5.750 in each of the next 10 games, their PR would drop to 6.443 (behind Denham Springs). Or if they win another similar game, it will help keep their PR up. Granted, Grace King will likely win some more games, but so will opponents of both of these squads will so I didn't try to factor all of that in. Simply trying to show how the averages can be skewed during the early & middle part of the season (and yes, sometimes still at the end of the season for numerous reasons). This is an example of how things change and smooth out over the course of a season. Denham Springs will get the power points for our (Lutcher's) wins by virtue of beating us last night. Of course it won't reflect the extent of the victory but I would think it would make an impact on the power rankings in any event. I haven't followed Zachary's recent record. I know they played St. Mike's but not sure of the result so that might impact their power points. For any voters who haven't had a look at them yet: Denham Springs is a very good team. They have excellent team speed at multiple positions and have solid spacing, movement and anticipation off the ball. They play a fairly direct game but in a smart sense, looking for optimum chances when they send the ball. I think they have a good number of club players from the BR area clubs (judging by stickers on cars in the parking lot Their style works very well on their field which is pretty narrow. Not a venue you want to play too heavy of a passing game at as we learned. I'd give 'em a look if you can before the end of the year.
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 7, 2014 9:23:57 GMT -6
. . . The question to me was as to how Zachary is rated ahead of Denham Springs in Power Ratings? Denham Springs PR is 6.705 and Zachary is 7.021. Now I use these 2 teams because I already did the work in helping to answer the question. I do get asked similar questions regularly so it really could be any set of 2 teams. The facts at the time... Denham Springs: 1) was undefeated in district, 2) had more PR wins (than Zachary), 3) had a better record (than Zachary), and 4) beat Zachary head-to-head The short answer: Zachary beat Grace King on November 24. Since King had 16 PR Wins (as of a few days ago), that earned Zachary a 21.000 for that game. However, if I were to exclude that one win over King, Zachary's PR would only be 5.750. So they get lots of credit for beating a win-filled team. This is an example of a game that will smooth out some over time. Right now that 21.000 is being averaged over 12 games. In the absence of another big win, by the end of the season, that 21 will be smoothed over 20-22 games (cutting it's impact almost in half). If they pull a 5.750 in each of the next 10 games, their PR would drop to 6.443 (behind Denham Springs). Or if they win another similar game, it will help keep their PR up. Granted, Grace King will likely win some more games, but so will opponents of both of these squads will so I didn't try to factor all of that in. Simply trying to show how the averages can be skewed during the early & middle part of the season (and yes, sometimes still at the end of the season for numerous reasons). This is an example of how things change and smooth out over the course of a season. Denham Springs will get the power points for our (Lutcher's) wins by virtue of beating us last night. Of course it won't reflect the extent of the victory but I would think it would make an impact on the power rankings in any event. I haven't followed Zachary's recent record. I know they played St. Mike's but not sure of the result so that might impact their power points. For any voters who haven't had a look at them yet: Denham Springs is a very good team. They have excellent team speed at multiple positions and have solid spacing, movement and anticipation off the ball. They play a fairly direct game but in a smart sense, looking for optimum chances when they send the ball. I think they have a good number of club players from the BR area clubs (judging by stickers on cars in the parking lot Their style works very well on their field which is pretty narrow. Not a venue you want to play too heavy of a passing game at as we learned. I'd give 'em a look if you can before the end of the year. Great point, methuselah. You are right. 1-0 or 7-0 would have had the same impact, but beating Lutcher with their 14 PR Wins (13 wins & 2 ties) has a huge impact on Power Ratings. Prior to the game, Denham Springs PR was 7.000 and they were in 60th place overall (34th in Division I, tied with Haughton). By beating Lutcher (and ignoring any other games that were played), their PR jumps to 8.000 and they move into 40th place overall (23rd in Division I, tied with Airline).
|
|
|
Post by methuselah on Jan 15, 2014 12:55:40 GMT -6
Ok, I know this is kind of getting down to minutiae, but I'm curious so I'll ask:
Team A wins a game against Team B.
Team B later plays a game against a district foe (let' call them Team C) that is tied at the end of play, goes to PKs and Team A wins on PKs.
Does Team A get 1/2 a power point, a whole power point (or none) for that game Team B won against their district foe ,Team C, in PKs after play was tied at the end of regulation/OT?
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 15, 2014 13:26:10 GMT -6
Ok, I know this is kind of getting down to minutiae, but I'm curious so I'll ask: Team A wins a game against Team B. Team B later plays a game against a district foe (let' call them Team C) that is tied at the end of play, goes to PKs and Team A wins on PKs. Does Team A get 1/2 a power point, a whole power point (or none) for that game Team B won against their district foe ,Team C, in PKs after play was tied at the end of regulation/OT? Keep in mind that there are 2 proposals on the table (as I understand it). 1 to pass Power Ratings (PR). 1 to allow District games to end in ties. Our calculations this year are based on what if BOTH of those pass. So that's the framework I'll use for answering the question. A team's PR wins = Wins in regulation plus 0.5 win for any game tied at end of regulation. Since Team A beat Team B, they will get 100% of Team B's "PR wins" (for the whole season) as part of their own PR calculation. It doesn't matter at what point in the season either of the games are played. Because Team B was tied at the end of regulation in their game against Team C, each team gets credit for 0.5 of a PR win. At that point, Team A earns the additional 0.5 PR win in their calculation. Keep in mind that the 0.5 points is for the game that they beat Team B. It is important to remember that in determining a PR rating, all games are averaged together. So in the example, the 0.5 points is averaged across all games played by Team A, which is why there's a lot of movement early in the season, but less severe jumps late in the season. Does that help?
|
|
|
Post by methuselah on Jan 15, 2014 14:10:58 GMT -6
Yep. That answers what I was wondering. Thanks!
|
|
|
Post by kevin on Jan 15, 2014 18:05:08 GMT -6
Another question on power ratings: LHSAA rule 6.11.4 states: "Sub-varsity teams are not allowed to participate in a varsity tournament in a sport that uses a power rating system. (Baseball, Basketball, Softball,and Volleyball)." I don't know how much that affects the boys' side, but several of the girls' tournaments this season had JV teams playing against varsity teams (sometimes by design, sometimes by last minute dropouts). Many of those JV teams (and I'm sure this is true for the boys as well) can beat lots and lots of varsity teams. If nothing else, the JV teams are a good way to fill out a bracket. And for the top JV teams, there may be very few other JV teams that can compete with them. Tournaments are a good chance for them to go up against some good teams, JV or otherwise.
So has there been any thought to trying to exempt soccer from this rule?
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2014 14:34:48 GMT -6
It looks like much of the LHSAA convention has been cancelled; however, the Business Meeting is set for Friday and voting should take place on this proposal (among others).
It reads as though by-laws would change, but a coach assured me that it would be for next year.
As a side note, I had a discussion about Belaire today & how they are rated 21st in Power Ratings & would make Playoffs if the system passed. The short answer is that they are benefitting from having St. Michael's (19 PR Wins) and Lutcher (19 PR Wins) and getting to play both of them twice.
If anyone is at the convention, please let us know how this turns out....
|
|
|
Post by methuselah on Jan 29, 2014 22:41:27 GMT -6
I think Belaire is not a typical situation. From what I've heard, their player numbers were decimated (I'm thinking mostly grades) which greatly changed them from what they were when they started out the season. I think they started something like 1 win 2 losses and 2 ties. But that changed after their numbers fell.
|
|
warrior16
Data Expert
Michael Stein - Volunteer Assitant
Posts: 2,169
|
Post by warrior16 on Jan 29, 2014 22:59:29 GMT -6
It looks like much of the LHSAA convention has been cancelled; however, the Business Meeting is set for Friday and voting should take place on this proposal (among others). It reads as though by-laws would change, but a coach assured me that it would be for next year. As a side note, I had a discussion about Belaire today & how they are rated 21st in Power Ratings & would make Playoffs if the system passed. The short answer is that they are benefitting from having St. Michael's (19 PR Wins) and Lutcher (19 PR Wins) and getting to play both of them twice. If anyone is at the convention, please let us know how this turns out.... It would be unusual for Belaire to make the playoffs despite having only one win (should the system pass). Yes, having good teams like Lutcher and St. Michael's in district definitely helps their power rating. 19/2 = 9.5 9.5 x 4 = 38 38 points (and this could increase if Lutcher and St. Michael pick up more wins) certainly isn't a bad result for four losses. It almost seems kind of unfair when losing to Lutcher or St. Michael's by 10 goals will get a team more PR points than beating a four-win team by 10 goals. That example is a bit hyperbolic, but the point still stands. In fact, Belaire actually gets about three more points for losing to us then we do by beating them. Perhaps a slight tweak in the formula, such as 6 points for a win, 3 for a draw, and 0 for a loss would balance things out a bit better. Strength of schedule should definitely be a crucial part of the formula, but perhaps it carries a little too much weight in this current formula at the expense of results. Maybe even 7 points for a win, 3.5 points for a draw, and 0 for a loss would be better. I plan on creating and testing several different power ratings systems when I have time over the summer to see what produces the most realistic results. Obviously the men who designed our current power ratings system put in countless hours creating and testing this system, and their effort is appreciated. But that doesn't mean we shouldn't look to improve the formula if it produces some odd results. I look forward to seeing how the power ratings work out in each division by the end of the regular season. Some of the strange things, like Jesuit being so low, can be explained by them playing an inordinate amount of Mississippi teams, thus making each loss to a Louisiana team have much more weight than it would to most teams.
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Jan 29, 2014 23:15:38 GMT -6
It looks like much of the LHSAA convention has been cancelled; however, the Business Meeting is set for Friday and voting should take place on this proposal (among others). It reads as though by-laws would change, but a coach assured me that it would be for next year. As a side note, I had a discussion about Belaire today & how they are rated 21st in Power Ratings & would make Playoffs if the system passed. The short answer is that they are benefitting from having St. Michael's (19 PR Wins) and Lutcher (19 PR Wins) and getting to play both of them twice. If anyone is at the convention, please let us know how this turns out.... It would be unusual for Belaire to make the playoffs despite having only one win (should the system pass). Yes, having good teams like Lutcher and St. Michael's in district definitely helps their power rating. 19/2 = 9.5 9.5 x 4 = 38 38 points (and this could increase if Lutcher and St. Michael pick up more wins) certainly isn't a bad result for four losses. It almost seems kind of unfair when losing to Lutcher or St. Michael's by 10 goals will get a team more PR points than beating a four-win team by 10 goals. That example is a bit hyperbolic, but the point still stands. Perhaps a slight tweak in the formula, such as 6 points for a win, 3 for a draw, and 0 for a loss would balance things out a bit better. Strength of schedule should definitely be a crucial part of the formula, but perhaps it carries a little too much weight in this current formula at the expense of results. Maybe even 7 points for a win, 3.5 points for a draw, and 0 for a loss would be better. I plan on creating and testing several different power ratings systems when I have time over the summer to see what produces the most realistic results. Obviously the men who designed our current power ratings system put in countless hours creating and testing this system, and their effort is appreciated. But that doesn't mean we shouldn't look to improve the formula if it produces some odd results. I look forward to seeing how the power ratings work out in each division by the end of the regular season. Some of the strange things, like Jesuit being so low, can be explained by them playing an inordinate amount of Mississippi teams, thus making each loss to a Louisiana team have much more weight than it would to most teams. Last year's proposal (that was not passed) allowed 10 points for a win, 5 for a tie, 0 for a loss with same %'s. The big complaint then was that it awarded too many points for winning against lesser opponents. That system encouraged taking anyone because a 10 point win was equal to a loss to a team with 20 wins. This year's proposal has the breakpoint lower & has created some of the challenges that have been raised. I did run about 7 different scenarios last year to do what you're suggesting as a summer project. I would say that 90-95% of the comparisons were identical. But there were always about 5-10% of outliers (including the coaches seedings at times). None of them were perfect (unfortunately).
|
|
warrior16
Data Expert
Michael Stein - Volunteer Assitant
Posts: 2,169
|
Post by warrior16 on Jan 29, 2014 23:29:24 GMT -6
It would be unusual for Belaire to make the playoffs despite having only one win (should the system pass). Yes, having good teams like Lutcher and St. Michael's in district definitely helps their power rating. 19/2 = 9.5 9.5 x 4 = 38 38 points (and this could increase if Lutcher and St. Michael pick up more wins) certainly isn't a bad result for four losses. It almost seems kind of unfair when losing to Lutcher or St. Michael's by 10 goals will get a team more PR points than beating a four-win team by 10 goals. That example is a bit hyperbolic, but the point still stands. Perhaps a slight tweak in the formula, such as 6 points for a win, 3 for a draw, and 0 for a loss would balance things out a bit better. Strength of schedule should definitely be a crucial part of the formula, but perhaps it carries a little too much weight in this current formula at the expense of results. Maybe even 7 points for a win, 3.5 points for a draw, and 0 for a loss would be better. I plan on creating and testing several different power ratings systems when I have time over the summer to see what produces the most realistic results. Obviously the men who designed our current power ratings system put in countless hours creating and testing this system, and their effort is appreciated. But that doesn't mean we shouldn't look to improve the formula if it produces some odd results. I look forward to seeing how the power ratings work out in each division by the end of the regular season. Some of the strange things, like Jesuit being so low, can be explained by them playing an inordinate amount of Mississippi teams, thus making each loss to a Louisiana team have much more weight than it would to most teams. Last year's proposal (that was not passed) allowed 10 points for a win, 5 for a tie, 0 for a loss with same %'s. The big complaint then was that it awarded too many points for winning against lesser opponents. That system encouraged taking anyone because a 10 point win was equal to a loss to a team with 20 wins. This year's proposal has the breakpoint lower & has created some of the challenges that have been raised. I did run about 7 different scenarios last year to do what you're suggesting as a summer project. I would say that 90-95% of the comparisons were identical. But there were always about 5-10% of outliers (including the coaches seedings at times). None of them were perfect (unfortunately). I can see how 10-5-0 was considered a bit extreme, and it would definitely punish teams that play really hard schedules and lose a fair share of games, while letting teams that play cakewalk schedules to coast to the top. And it would create a scheduling nightmare for coaches of really good teams trying to find non-district opponents, since most teams would be afraid to play them and plummet in the power ratings. Something like 5-2.5-0 or 6-3-0 would definitely produce a more realistic and fair playoff seeding. No system is perfect, but I think if we really work hard on the formula we can get pretty darn close. Perhaps the formula doesn't need to be adjusted at all. It will certainly be easier for you guys to manage the power rankings next year with all scores having to be reported to the LHSAA directly. Hopefully that will mean less wasted time for you guys trying to hunt down unreported scores.
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Feb 1, 2014 5:21:02 GMT -6
I was fielding questions much of the day yesterday and need some help to clarify or confirm what the decisions on Friday mean. I actually went to agenda items 37 & 38 from the convention to review it as proposed (maybe some of it was amended prior to passing). Anyone with specific facts is invited to help make it clear or correct anything that needs correction. Trying to capsulize what Friday's decisions mean to most of us.
1) Power Ratings passed for NEXT season. This year they continue to be strictly for informational purposes, but the point system we've been using is the one that passed.
2) District games CAN end in a TIE next year (which is how we've reflected them in the current year power rating calc).
3) There is NOT any BONUS/PENALTY for playing teams outside of your DIVISION.
4) Scores will be REPORTED to LHSAA within a certain timeframe? Need clarity on that - we've talked about it, but I don't actually find it in print.
5) All 3 Divisions will have 32 teams in the 2014-15 playoffs (including Division II with its 38 teams currently). The 3 equal division clause isn't set to take effect until 2015-16. Or has something over-ridden that??
6) Next year ONLY District WINNERS will advance to playoffs regardless of their Power Rating. The remaining 21-23 spots will be filled based on Power Rating.
7) Teams that play FEWER than 15 games can make playoffs, but ONLY after all other teams are included and seeded.
8) SEEDING will be based purely on Power Ratings for all teams, including District Winners. Coaches seeding has been eliminated AFTER this year.
|
|
|
Post by Chalmetteowl on Feb 1, 2014 7:21:24 GMT -6
i'd love to see the district standings be like the rest of the world does soccer, 3 points for a win, 1 for a draw, and goal difference used up to +8 for one team in a game
i do think 15 games as a minimum is too much for some programs... i'd like to see that number lowered to 10 or 12
|
|
|
Post by kevin on Feb 1, 2014 7:25:19 GMT -6
i do think 15 games as a minimum is too much for some programs... i'd like to see that number lowered to 10 or 12 With tournaments there is zero excuse for not playing 15 games. Even now the teams that aren't playing 15 games are the teams that have no money or don't care about soccer.
|
|
p_malinich
Data Expert
www.elevenlions.com
Posts: 4,201
|
Post by p_malinich on Feb 3, 2014 7:45:47 GMT -6
5) All 3 Divisions will have 32 teams in the 2014-15 playoffs (including Division II with its 38 teams currently). The 3 equal division clause isn't set to take effect until 2015-16. Or has something over-ridden that?? It's weird to quote yourself, but can anyone clarify if my statement/question is true. The Power Ratings proposal on the agenda indicates 32 from each District (for next year). The equal division clause in handbook doesn't kick in until 2015-16. Is that right or has one of those been modified in some way?
|
|
|
Post by raiderfan on Feb 3, 2014 8:16:26 GMT -6
With tournaments there is zero excuse for not playing 15 games. Even now the teams that aren't playing 15 games are the teams that have no money or don't care about soccer. Well, I'm not sure having money should be a requirement to compete. It's easier for the "rich" schools to travel to tournaments and such, and that does give them more experience and help them become better teams. But I think everyone should have a chance to compete in the playoffs, if they have a good season. It's also easier for the larger schools. If you have 20+ kids on the soccer roster, it's not a big deal to schedule more games, even if a few of the kids might have a conflict with the date. If you have 14 or 15 kids on the roster, you need to have just about everyone show up to every game, especially if you end up with an injury or two. I don't think 15 games is too unreasonable, but it will be a stretch for some smaller schools.
|
|
|
Post by Chalmetteowl on Feb 3, 2014 9:56:50 GMT -6
"even if a few of the kids might have a conflict with the date. If you have 14 or 15 kids on the roster, you need to have just about everyone show up to every game,"
funny how this isn't a problem in other sports...
|
|
|
Post by raiderfan on Feb 3, 2014 10:21:56 GMT -6
"even if a few of the kids might have a conflict with the date. If you have 14 or 15 kids on the roster, you need to have just about everyone show up to every game," funny how this isn't a problem in other sports... Who says it's not a problem in other sports? Some of the smaller schools, for example, can barely field enough players for a football team (American football, that is), even though it's one of the most popular sports. But if they do, they're not expected to travel to tournaments and other extra games - generally just one game per week for the season, and often with an open week thrown in. Basketball is not quite as challenging for a small school, as you only need 5 people on court at a time.
|
|