'04 Girls FBR Rankings - 05-01-13

BW, how long before we get the final rankings for the season? Do you wait until after all the tournament fever has passed?

soccerdad1969 wrote:BW, how long before we get the final rankings for the season? Do you wait until after all the tournament fever has passed?

The rankings are never really "final". I keep a running database of the previous 16 weeks of results. We've already entered spring tourney season, yet some leagues are still playing their final 1-2 weeks of games. In 2 more weeks, "pre-season" tourney results from the end of Jan/beginning of Feb will start to drop off, and I will keep adding tourney results from May/June tournaments.

Typically, things are a bit slow for the U9's going into U10 from about Mid-June through the beginning of August, so the closest thing to a "final" spring ranking will be around that Mid-June time frame.

I noticed several of the teams have the same number of points. Is there a tie breaker scenario used to determine which team ranks higher? I realize this is an inexact science but things like this interest me.

10-13 will shake itself out this weekend.

I think 2-4 sort themselves out too

Primetime?

yep

Keepersdad wrote:I noticed several of the teams have the same number of points. Is there a tie breaker scenario used to determine which team ranks higher? I realize this is an inexact science but things like this interest me.

The points are actually calculated out to the second decimal place. For whatever reason, the page that I use in FBR to cut and past the rankings to post on the forum, rounds off the point value displayed to the closest whole # value (i.e. 45.20 and 44.65 would both be displayed as 45, but the team with 45.20 will have the higher rank). So... if there is a difference in actual rank of two teams that have the same point value in my posting, the higher ranked team actually has a higher point total when you carry it out to the 2nd decimal place. If the actual rank of 2 teams is the same in my posting, then those teams are actually tied in points all the way out to the 2nd decimal.

Ouch! I wouldn't want to be NTX Strikers in this one.

Thanks BWG. I thought maybe it was alphabetical or something.

bwgophers wrote:
canaryman wrote:
BW, where does Tier 2 start and stop?



"The ranking tiers do not line up exactly with the rankings. However, the general breakdown of the tiers for the '04's is currently:

Tier 1: RASE and DT South (only unbeaten, un-tied teams)
Tier 2: Remainder of top 10, except TSNR
Tier 3: TSNR + majority of teams between 11-30
Tier 4: Majority of teams between 31-50
Tier 5: Majority of teams between 51-85
Tier 6: Majority of teams between 86-95


Out of curiosity, please explain your specific logic on TSNR?"

The above was from last weeks thread. I'm still curious.

bwgophers wrote:

Keepersdad wrote:I noticed several of the teams have the same number of points. Is there a tie breaker scenario used to determine which team ranks higher? I realize this is an inexact science but things like this interest me.

The points are actually calculated out to the second decimal place. For whatever reason, the page that I use in FBR to cut and past the rankings to post on the forum, rounds off the point value displayed to the closest whole # value (i.e. 45.20 and 44.65 would both be displayed as 45, but the team with 45.20 will have the higher rank). So... if there is a difference in actual rank of two teams that have the same point value in my posting, the higher ranked team actually has a higher point total when you carry it out to the 2nd decimal place. If the actual rank of 2 teams is the same in my posting, then those teams are actually tied in points all the way out to the 2nd decimal.

Thanks for the response. Just out of further curiosity, how much time in a given week do you spend gathering all of the data and inputting it into your program?

Keepersdad wrote:

bwgophers wrote:

Keepersdad wrote:I noticed several of the teams have the same number of points. Is there a tie breaker scenario used to determine which team ranks higher? I realize this is an inexact science but things like this interest me.

The points are actually calculated out to the second decimal place. For whatever reason, the page that I use in FBR to cut and past the rankings to post on the forum, rounds off the point value displayed to the closest whole # value (i.e. 45.20 and 44.65 would both be displayed as 45, but the team with 45.20 will have the higher rank). So... if there is a difference in actual rank of two teams that have the same point value in my posting, the higher ranked team actually has a higher point total when you carry it out to the 2nd decimal place. If the actual rank of 2 teams is the same in my posting, then those teams are actually tied in points all the way out to the 2nd decimal.

Thanks for the response. Just out of further curiosity, how much time in a given week do you spend gathering all of the data and inputting it into your program?

The gathering and inputting of data is the easy part. Explaining it to the rest of us OTOH..

Good point. Maybe I should stop asking dumb questions.

Keepersdad wrote:Good point. Maybe I should stop asking dumb questions.

That's the whole purpose of this site

KnKsDad wrote:bwgophers wrote:
canaryman wrote:
BW, where does Tier 2 start and stop?



"The ranking tiers do not line up exactly with the rankings. However, the general breakdown of the tiers for the '04's is currently:

Tier 1: RASE and DT South (only unbeaten, un-tied teams)
Tier 2: Remainder of top 10, except TSNR
Tier 3: TSNR + majority of teams between 11-30
Tier 4: Majority of teams between 31-50
Tier 5: Majority of teams between 51-85
Tier 6: Majority of teams between 86-95


Out of curiosity, please explain your specific logic on TSNR?"

The above was from last weeks thread. I'm still curious.

Prior to 4/23, TSNR's record against teams in Tier 2 was 0-2 (0%). They did not belong in Tier 2 based on that fact.

After 4/23, TSNR's record against teams in Tier 2 was 1-2 (33%). At this point, I look at their record against Tier 3 teams, which was 4-1 (80%). Becomes a bit of a judgement call at this point. In these cases, I look at what assigning the team to each Tier does to the overall record of 1 Tier vs. another. In this case, I decided that the overall better case scenario was for TSNR to be in Tier 3.

As of 4/30, Piper has been relegated to Tier 3, based on the loss to TSNR on 4/20 and their loss this past weekend to SRFC. With that change, TSNR's record against Tier 2 is now 0-1 (0%) and against Tier 3 is 5-2 (71%). Turns out my decision to keep TSNR in Tier 3 after 4/23 was the correct one in this case.

Keepersdad wrote:

bwgophers wrote:

Keepersdad wrote:I noticed several of the teams have the same number of points. Is there a tie breaker scenario used to determine which team ranks higher? I realize this is an inexact science but things like this interest me.

The points are actually calculated out to the second decimal place. For whatever reason, the page that I use in FBR to cut and past the rankings to post on the forum, rounds off the point value displayed to the closest whole # value (i.e. 45.20 and 44.65 would both be displayed as 45, but the team with 45.20 will have the higher rank). So... if there is a difference in actual rank of two teams that have the same point value in my posting, the higher ranked team actually has a higher point total when you carry it out to the 2nd decimal place. If the actual rank of 2 teams is the same in my posting, then those teams are actually tied in points all the way out to the 2nd decimal.

Thanks for the response. Just out of further curiosity, how much time in a given week do you spend gathering all of the data and inputting it into your program?

Too much...

BW - I have a question....

Since "tournament teams" can be vastly different from "league teams", is there any consideration given to that scenario? Is there a seperate ranking table for a "frozen" roster?

We have all seen a lower-level big club team get better than expected results in a tournament simply because the top-level team(s) in that club happen to be off that particular weekend - resulting in guest players that will never play for that team in a frozen roster situation such as a league or qualifying tournament...so I think that tends to skew the rankings a bit when tournaments are included...but it would be a real bugger to try to account for that..

10sDad wrote:BW - I have a question....

Since "tournament teams" can be vastly different from "league teams", is there any consideration given to that scenario? Is there a seperate ranking table for a "frozen" roster?

We have all seen a lower-level big club team get better than expected results in a tournament simply because the top-level team(s) in that club happen to be off that particular weekend - resulting in guest players that will never play for that team in a frozen roster situation such as a league or qualifying tournament...so I think that tends to skew the rankings a bit when tournaments are included...but it would be a real bugger to try to account for that..

Nope. Guest playing and roster movement is a fact of life in NTX academy. It's there in league play just as is in tourney play. No way to account for it in a mathematical formula like FBR, and I won't even try. My experience is that over the long haul, it pretty much balances itself out one way or another.

bwgophers wrote:
Prior to 4/23, TSNR's record against teams in Tier 2 was 0-2 (0%). They did not belong in Tier 2 based on that fact.

After 4/23, TSNR's record against teams in Tier 2 was 1-2 (33%). At this point, I look at their record against Tier 3 teams, which was 4-1 (80%). Becomes a bit of a judgement call at this point. In these cases, I look at what assigning the team to each Tier does to the overall record of 1 Tier vs. another. In this case, I decided that the overall better case scenario was for TSNR to be in Tier 3.

As of 4/30, Piper has been relegated to Tier 3, based on the loss to TSNR on 4/20 and their loss this past weekend to SRFC. With that change, TSNR's record against Tier 2 is now 0-1 (0%) and against Tier 3 is 5-2 (71%). Turns out my decision to keep TSNR in Tier 3 after 4/23 was the correct one in this case.

KnK, what 2 teams outside the top 10 has Spirit North lost to this year?

bwgophers wrote:

10sDad wrote:BW - I have a question....

Since "tournament teams" can be vastly different from "league teams", is there any consideration given to that scenario? Is there a seperate ranking table for a "frozen" roster?

We have all seen a lower-level big club team get better than expected results in a tournament simply because the top-level team(s) in that club happen to be off that particular weekend - resulting in guest players that will never play for that team in a frozen roster situation such as a league or qualifying tournament...so I think that tends to skew the rankings a bit when tournaments are included...but it would be a real bugger to try to account for that..

Nope. Guest playing and roster movement is a fact of life in NTX academy. It's there in league play just as is in tourney play. No way to account for it in a mathematical formula like FBR, and I won't even try. My experience is that over the long haul, it pretty much balances itself out one way or another.

So I guess the safe assumption would be that a lower ranked "big-club" team is inherently over-ranked when tournaments are taken into account. Therefore, applying this logic means:

- if you are playing a particular "shirt" in a TOURNAMENT, be ready for them to be better than their ranking suggests...especially if their higher ranked team(s) are off that weekend.

and....

- if you are playing a particular "shirt" in a LEAGUE or next year's QT, don't assume that if they are significantly higher ranked than your team, that its an enormous, gigantic upset if you beat them.

True?

or if you are the "higher" ranked team and you lose, maybe you are not as good as you thought. Bringing in a player or 2 completely changing the results means you were closer than you thought. They are still your age group so you need to beat whats in front of you

10sDad wrote:

bwgophers wrote:

10sDad wrote:BW - I have a question....

Since "tournament teams" can be vastly different from "league teams", is there any consideration given to that scenario? Is there a seperate ranking table for a "frozen" roster?

We have all seen a lower-level big club team get better than expected results in a tournament simply because the top-level team(s) in that club happen to be off that particular weekend - resulting in guest players that will never play for that team in a frozen roster situation such as a league or qualifying tournament...so I think that tends to skew the rankings a bit when tournaments are included...but it would be a real bugger to try to account for that..

Nope. Guest playing and roster movement is a fact of life in NTX academy. It's there in league play just as is in tourney play. No way to account for it in a mathematical formula like FBR, and I won't even try. My experience is that over the long haul, it pretty much balances itself out one way or another.

So I guess the safe assumption would be that a lower ranked "big-club" team is inherently over-ranked when tournaments are taken into account. Therefore, applying this logic means:

- if you are playing a particular "shirt" in a TOURNAMENT, be ready for them to be better than their ranking suggests...especially if their higher ranked team(s) are off that weekend.

and....

- if you are playing a particular "shirt" in a LEAGUE or next year's QT, don't assume that if they are significantly higher ranked than your team, that its an enormous, gigantic upset if you beat them.

True?

There will probably be many people out there who disagree with me on this, but my answer to your question is a big, fat, FALSE.

IMO, guest players have less of an impact than most people want to believe. Can they help? Sure. How often are they complete game changers that drastically alter the final outcome of a game? Again, IMO, not very often. I think they are a convenient excuse used by people to justify why their DD's team didn't get the outcome they were expecting/hoping for in a game.

bwgophers wrote:

KnKsDad wrote:bwgophers wrote:
canaryman wrote:
BW, where does Tier 2 start and stop?



"The ranking tiers do not line up exactly with the rankings. However, the general breakdown of the tiers for the '04's is currently:

Tier 1: RASE and DT South (only unbeaten, un-tied teams)
Tier 2: Remainder of top 10, except TSNR
Tier 3: TSNR + majority of teams between 11-30
Tier 4: Majority of teams between 31-50
Tier 5: Majority of teams between 51-85
Tier 6: Majority of teams between 86-95


Out of curiosity, please explain your specific logic on TSNR?"

The above was from last weeks thread. I'm still curious.

Prior to 4/23, TSNR's record against teams in Tier 2 was 0-2 (0%). They did not belong in Tier 2 based on that fact.

After 4/23, TSNR's record against teams in Tier 2 was 1-2 (33%). At this point, I look at their record against Tier 3 teams, which was 4-1 (80%). Becomes a bit of a judgement call at this point. In these cases, I look at what assigning the team to each Tier does to the overall record of 1 Tier vs. another. In this case, I decided that the overall better case scenario was for TSNR to be in Tier 3.

As of 4/30, Piper has been relegated to Tier 3, based on the loss to TSNR on 4/20 and their loss this past weekend to SRFC. With that change, TSNR's record against Tier 2 is now 0-1 (0%) and against Tier 3 is 5-2 (71%). Turns out my decision to keep TSNR in Tier 3 after 4/23 was the correct one in this case.

Thanks for the explanation. This may be a dumb question, but are there presently any teams that are ranked 11-30 that are considered Tier 2?