|
Post by How on Apr 4, 2011 9:10:35 GMT -4
This ranking system is really confusing to me.
How is a Columbus team who has a loss and is 13-1 ranked higher than McCarthy who is 14-0 and has not dropped a set and clearly has faced tougher competition (teams ranked higher)?
How is a team with 6 losses (Oviedo), ranked so high (number 12)? I am not hating on them, they are a solid team but they have 6 losses! How is a team with 5 losses (Chaminade) ranked number 18?
|
|
|
Post by Me too on Apr 4, 2011 9:18:41 GMT -4
I am confused as well. What is put into the rankings? Does strength of schedule effect the rankings?
|
|
|
Post by yes on Apr 4, 2011 9:59:09 GMT -4
it is supposed to, but how can you tell me that columbus has a stronger strength of schedule than mccarthy?
it is also supposed to go by sets, not only games won and lossed, yet mccarthy hasnt lost a set all season.
i am not from mccarthy, just using them as an example. i do respect columbus but just dont see how they are ranked number 1
|
|
|
Post by Point system on Apr 4, 2011 12:08:59 GMT -4
it is supposed to, but how can you tell me that columbus has a stronger strength of schedule than mccarthy? it is also supposed to go by sets, not only games won and lossed, yet mccarthy hasnt lost a set all season. i am not from mccarthy, just using them as an example. i do respect columbus but just dont see how they are ranked number 1 From my understanding, Its a point system, like how strong your RPI is..BLAH BLAH..take the BCS for example and there point systems, Columbus has played alot more teams that are ranked higher then AMHS so therefore the wins they have mean more, so Varela defeated by Palmetto with just one loss is pretty high, loss in 5 sets, Oviedo defeated Palmetto so that helps there number, Ransom winning there pool defeating Coral Park boost tere number, same with Spanish river, Bishop more if that makes sense, AMHS deserves to be the outright number one team in the state, have yet to lose, but also at the end of they year you will see it shape up the way its suppose to look...Columbus has one lost to BM, Spanish River is high up there because there losses are to teams higher then them, Depends also, if you are in a district thats not strong and u run the table and those are the only games you play then you will go up pretty high, but once you lose to a good team down you go...also if you win a game in 5 its not worth as much as you win a game in three..despite columbus having one lost they have beaten teams ranked high in three..and there OH left early from the tournament so they could easily be undefeated as well....Southwest will climb back up as well, they only have one loss to spanish river its confusing bbut theres sort of a logic to it...I might be totally wrong on how its done but thats my understanding!
|
|
|
Post by wrong on Apr 4, 2011 12:52:01 GMT -4
it is supposed to, but how can you tell me that columbus has a stronger strength of schedule than mccarthy? it is also supposed to go by sets, not only games won and lossed, yet mccarthy hasnt lost a set all season. i am not from mccarthy, just using them as an example. i do respect columbus but just dont see how they are ranked number 1 From my understanding, Its a point system, like how strong your RPI is..BLAH BLAH..take the BCS for example and there point systems, Columbus has played alot more teams that are ranked higher then AMHS so therefore the wins they have mean more, so Varela defeated by Palmetto with just one loss is pretty high, loss in 5 sets, Oviedo defeated Palmetto so that helps there number, Ransom winning there pool defeating Coral Park boost tere number, same with Spanish river, Bishop more if that makes sense, AMHS deserves to be the outright number one team in the state, have yet to lose, but also at the end of they year you will see it shape up the way its suppose to look...Columbus has one lost to BM, Spanish River is high up there because there losses are to teams higher then them, Depends also, if you are in a district thats not strong and u run the table and those are the only games you play then you will go up pretty high, but once you lose to a good team down you go...also if you win a game in 5 its not worth as much as you win a game in three..despite columbus having one lost they have beaten teams ranked high in three..and there OH left early from the tournament so they could easily be undefeated as well....Southwest will climb back up as well, they only have one loss to spanish river its confusing bbut theres sort of a logic to it...I might be totally wrong on how its done but thats my understanding! yes your logic makes no sense and IS totally wrong. 1. Palmetto loosing to Oviedo hurts Varela, does not help them. It makes palmetto look like a weaker team, therefore varela a weaker team. varela would need palmetto to do well so the one loss is to a good team. 2. McCarthy has beaten the number 6 team (DP) who was previously undefeated before lossing to McCarthy. McCarthy has beaten the number 7 team (River) twice, whom is undefeated otherwise. Columbus lost to BM who already has a loss and is ranked number 10. Columbus' best win was against the number 17 ranked team (Berkely Prep) who already had a loss prior to playing Columbus. Which means Columbus has a weaker strength of schedule. 3. Yes columbus was missing their OH but the rankings do not take this into consideration. There is no formula that says if a player is missing the loss doesnt count, they lost. Just like palmetto lossed "without their starting setter". Yes i agree it might actually matter, but it does not matter in the ranking system. You can take it into consideration when you vote
|
|
|
Post by rankings on Apr 4, 2011 17:25:38 GMT -4
Over the years Dave has explained this chess system a bunch of times. Don't get your panties in a wad about the rankings. The system is an algorithm used in ranking chess players, these things are not an exact science so they sometimes will give you a wacky result. At the end of the day the I think that the rankings usually prove to be true.
If I remember correctly, the quantity of games won or loss are important. Wins outside a tournament count more than wins inside of a tournament because a non-tournament win gives you three games won as opposed to only two games won in a tourney. AMHS and Columbus have played the same number of matches but since McCarthy has played 2 tournaments, in which a win is best out of three they have actually won fewer sets than Columbus even with a loss by Columbus. A team like Oviedo has six losses but ten wins so far so that is a lot of positive games won. Once other teams have as many games, as the season progresses, you may see Oviedo go down even if they keep winning as other teams reach the same number of wins. Palmetto (10 wins) and Chaminade (11 wins) both loss to Oviedo which might be why they are ranked a little lower.
It is math. Even if my knowledge of the formula is not exact, understand that this is how algorithms work, there is a logic behind them and they will continue following it. I am sure Dave can better explain, over the years the rankings are generally close at the end of the season. This is why there is a coaches and fan polls.
|
|
|
Post by imnumber4 on Apr 4, 2011 17:44:00 GMT -4
the rankings dont have anything to do with who goes to states or any implications of the playoffs so why does it matter? Dave has it set up so people can see how everyone is doing and they are usually fairly accurate but I believe if you look at it its more of a general view of the top teams, most everyone looking at the rankings knows how each team is and how they stack up. If you care about rankings so much wait for the end of the season and rank your own top 8 from states.
|
|
|
Post by David Crowder on Apr 6, 2011 3:33:19 GMT -4
Here we go sports fans and math geeks,
The ranking system is based off the ELO system which is used to rank chess players internationally. Grand Masters and stuff like that.
System was developed by a Hungarian mathematician named ELO. He was a chess player, hence the reason he developed the system.
The ELO system since is the basis for all other mathematical and computerized ranking systems in existence today. RPI, Sagrin etc etc etc for ALL sports.
Each set is worth 32 points. At the beginning of the season each team antes 16 points to put 32 points into the pot. Winner of the set gets 32 points. Next set played the special mathematical formula developed by ELO comes into play. The formula determines how many points each team antes next. So the winner of the first set will probably ante like 17 or 18 points the loser will ante the difference like 15 or 14 points.
The special formula is based off of some statistical distribution which at 3:30 am I can't remember and won't google either.
So after many many matches and sets are played every team is closer(er) to their true ranking, so when they meet and play a really good team playing a really bad team will ante something like ...GOOD 28 points Bad 4 points...so as you see a GOOD team beating a bad team really does not help the GOOD team or really hurt the BAD team, but a loss by the GOOD team does hurt the GOOD team and help the BAD team.
To improve your ELO ranking it helps to play GOOD teams, play as many SETS as possible, and WIN. The more you play the more points you can earn. Winning a tourney will build your points up very quickly because you are beating several GOOD teams.
So with McCarthy vs Columbus. As stated earlier, McCarthy has won 10 tournament matches playing only 2 SETS, Columbus has won the same number of matches but has done so by winning 3 of 5 matches.
I do believe everyone would agree that winning 3 of 5 is more difficult than 2 of 3 and thus should be rewarded as such.
For McCarthy to regain the ELO #1 ranking they need to win 2 matches this week or have Columbus drop a Set or 2 this week.
Bottom line, the more you play, the more you win, the better your ELO power ranking.
Dave
|
|
|
Post by David Crowder on Apr 6, 2011 3:41:43 GMT -4
Also, I would like to add that the 10 or so coaches that vote in the coaches poll each week has picked McCarthy as #1 every week of the season. #2 has flucuated betweeen several teams.
The poll may not have every team pegged exactly where they should be but it is very very close. And it does do a good job and letting everyone in the state know who is doing good.
A good example is Varela. They are off to a really good start. There had been no discussion about this team until this week. Are they a top 10 team? Well this week there were, next week may be different.
|
|
|
Post by David Crowder on Apr 6, 2011 3:42:51 GMT -4
Also to improve your team's score in the ELO system please report any missing scores. No score by a match played is 0 points for your team.
|
|
|
Post by ELO on Apr 8, 2011 9:16:35 GMT -4
The major difference with the ELO used for chess and the one being used here for volleyball, is that the chess ELO doesn't have a season and their rankings go on forever. For volleyball each season starts with all teams having 1500 points, which means all teams are statistically even at the beginning of the season and we all know that's not true. Playing a "weaker" team early will earn a "stronger" team more points at the beginning of the season then at the end of the season after the "weaker" team has lost numerous times. That's why there are some strange rankings early in the season and the top of the rankings sorts out at the end of the season when "stronger" teams start playing "stronger" teams in district finals, regionals and states. Only way around that is to handicap the beginning point totals for all the teams, but that would be a subjective process and we know how badly subjective opinion is taken on this website. So, it won't solve anything.
|
|
|
Post by David Crowder on Apr 9, 2011 22:15:36 GMT -4
ELO, yes you are correct. At the beginining of the season ALL TEAMS are EQUAL. Everyone is 0 and 0. No wins no losses. No reputations to live up to or hurdles to jump.
Math Geek stuff here guys. Statistical theory says that you need 20 samples to begin to approximate the NORMAL DISTRIBUTION. So it pretty much takes all season before the Rankings actually become stable. As the season does progress though everyone's ELO rating does approach their true rating, so I have thought about running the ranking algorithm twice to see if the second run improves the rankings. So far I have not seen any indication that it does improve the ratings significantly.
The ranking system does do a GOOD JOB of clumping the teams though, meaning all the teams that are about equal are clumped together with the clumps being accurate but the ranking within the clump being off some.
Again, I think it works very well and where else have you ever seen a ranking for all the teams in a sport?
One interesting side note, is that one thing I have learned or confirmed is that it is much harder to WIN THEM ALL (an undefeated season) than to lose them all. It seems every year with have several 0 FER teams and of course there CAN ONLY BE ONE undefeated team.
Dave
|
|
|
Post by whats going on on Apr 29, 2011 8:55:12 GMT -4
what is going on with the rankings?!?!?
how are teams like oviedo, sunset and park vista suddenly jumping so high?
and teams like palmetto, DP and lake mary dropping all of a sudden?
|
|
|
Post by cflballer on Apr 29, 2011 9:03:53 GMT -4
yea really odd that a team with records of 16-8, 18-5 and 18-7 can be ahead of teams with 20-3 records. And don't give us late in the year strength of schedule argument. it doesn't work especially when the 20-3 teams have already beat the other teams.
|
|
|
Post by dumb on Apr 29, 2011 12:46:01 GMT -4
Who cares about the rankings?! Any given team can beat any other team on any given day. So just let the rankings be and play your game.
|
|