'05 Girls FBR Rankings - 09-26-13

Wow - so if you win, you drop in the rankings. The team you beat by 6 in head to head 2 weeks ago is one spot below you.....

As you say - in October I guess we will get a better idea where teams stack up.

Yup. Many teams outside the top 20 or so still have a small # of games (only 3 or 4) in their game history. That always throws some noise into the rankings.

Also, the games from the weekend of 5/17-5/18 were dropped from the history this week. That was a big tournament weekend, so a larger than normal # of games were removed, which jumbles the week-week movement even more.

It's too bad that Lady As were dropped but you can't control that they wanted to go play 11v11.  They seem to be doing very well in the CFBAL 04 Gold division though so good for them.  Given the fact that Sting Scarfone played in the Plano Labor Day Tournament, they should make it through the season in the FBRs by playing in one of the Thanksgiving tournaments.  Given the competition though, they should be playing with the LAs in 04 Gold.

Lots of movement overall this week in the FBRs  It still looks about right.  Thanks again BW!

Any teams decided on what tourneys they will be playing in November?

bwgophers wrote:Yup.  Many teams outside the top 20 or so still have a small # of games (only 3 or 4) in their game history.  That always throws some noise into the rankings.

Also, the games from the weekend of 5/17-5/18 were dropped from the history this week.  That was a big tournament weekend, so a larger than normal # of games were removed, which jumbles the week-week movement even more.

BW,
How come you don't keep a longer history? Seems like if you kept a 6-9 month history the rankings would be a little more accurate.

KopDad wrote:

bwgophers wrote:Yup.  Many teams outside the top 20 or so still have a small # of games (only 3 or 4) in their game history.  That always throws some noise into the rankings.

Also, the games from the weekend of 5/17-5/18 were dropped from the history this week.  That was a big tournament weekend, so a larger than normal # of games were removed, which jumbles the week-week movement even more.

BW,
How come you don't keep a longer history? Seems like if you kept a 6-9 month history the rankings would be a little more accurate.

Two reasons:

1) 6-9 months is an eternity in NTX academy
2) FBR doesn't allow me to use a sliding weight scale to give less weight to older game results

16 weeks generally will cover a full league season (fall or spring) + the usual 1 month or so of tournaments either directly preceding the season, or following the season. In other words, by the time we get to the end of fall league season, FBR will pretty much contain results from the August tournaments + the full fall league season. By the time we hit the end of post-fall tournament season around the 1st week of December, FBR will generally contain fall league season + the fall tournaments.

My experience is that results older than that typically aren't representative of a team's current level of play, and without being able to give those older results a lesser weight in the calculation, they actually cause undue influence on the calculation and make it less accurate, not more accurate.

Things will always get a bit noisy during the transition periods in winter and summer, but that's also when the majority of player movement takes place. Idle hands...