G08 State Cup President's Bracket

Supermodel56 said:
So any predictions on who moves onto Quarters/Semis/Finals? ;)
Click to expand...
Updated rankings for the 4 groups. They combined A and B but still say 1st place team plus three wildcards, so is that 2 out of A/B with the other 3 1st place teams and 3 wild cards? or 4 1st place teams and 4 wildcards?
G08_StateCup.png
 
Carlsbad10 said:
Updated rankings for the 4 groups. They combined A and B but still say 1st place team plus three wildcards, so is that 2 out of A/B with the other 3 1st place teams and 3 wild cards? or 4 1st place teams and 4 wildcards?
View attachment 586
Click to expand...
Three wildcards. First place of first half of A/B and first place of second half of A/B go through. The A/B group has been rearranged compared to what you have above.
 
rainbow_unicorn said:
Three wildcards. First place of first half of A/B and first place of second half of A/B go through. The A/B group has been rearranged compared to what you have above.
Click to expand...
I just ranked the teams 1-6 for A/B now that they combined them, are you saying that the way they had it originally with a Group A made up of Blues, Albion, and LA Galaxy OC (Irvine Slammers) and a Group B of Legends, LAGSB, and Southwestern each get one spot. I wouldn't think so the way the bracket is combined on calsouth.
 
Supermodel56 said:
Nice! The overall ranking does not include strength of schedule, does it?
Click to expand...
Not necessarily, but neither did the BCS which was a similar type algorithm. The methodology uses a "best fit" algorithm that takes as input scores from all tournaments and league play (~1400 games) to then come up with a "strength" number (SRS) for each team. The difference in SRS tells you how many more goals better one team is than another. And you move up and down the rankings (your SRS increases or decreases) based on how you do vs. what the algorithm predicts for that game. So you play a team with a higher SRS and beat them and your SRS increases. More recent games are weighted more than older games and SOS is calculated by adding the SRS of all your opponents and dividing by the number of games and ranking each team. For example the Blues played the toughest schedule by this calculation and you can see that with 23 games against teams ranked in the Top 10.
 
Carlsbad10 said:
I just ranked the teams 1-6 for A/B now that they combined them, are you saying that the way they had it originally with a Group A made up of Blues, Albion, and LA Galaxy OC (Irvine Slammers) and a Group B of Legends, LAGSB, and Southwestern each get one spot. I wouldn't think so the way the bracket is combined on calsouth.
Click to expand...
Yes. The group A teams will only play against group B teams and vice versa. That way you get an equal opponents field from where you can determine top group A and top group B teams. They probably combined into an A/B group so that parents in other groups wouldn't complain that teams in group A and B have a better chance of making it to knockout rounds.
 
rainbow_unicorn said:
Three wildcards. First place of first half of A/B and first place of second half of A/B go through. The A/B group has been rearranged compared to what you have above.
Click to expand...
Ahh, the ladder would have helped us, 2 teams from A/B. Not sure how that works to have the 2nd team from AB play the 2nd wildcard rather than winner of C and winner of D getting a WC or second place team. WC #2 is the spot to play for. ladder.png
 
Carlsbad10 said:
Not necessarily, but neither did the BCS which was a similar type algorithm. The methodology uses a "best fit" algorithm that takes as input scores from all tournaments and league play (~1400 games) to then come up with a "strength" number (SRS) for each team. The difference in SRS tells you how many more goals better one team is than another. And you move up and down the rankings (your SRS increases or decreases) based on how you do vs. what the algorithm predicts for that game. So you play a team with a higher SRS and beat them and your SRS increases. More recent games are weighted more than older games and SOS is calculated by adding the SRS of all your opponents and dividing by the number of games and ranking each team. For example the Blues played the toughest schedule by this calculation and you can see that with 23 games against teams ranked in the Top 10.
Click to expand...


Interesting! Where is the data being pulled from? Wondering what it would look like if strength of schedule was based on wins (say weight by differential up to 3 goals) since a lot of teams don't go for the blowout wins... that also tends to skew the data, if you're playing weaker teams during the season or weaker tournaments, you're going to have a much higher SRS.
 
Supermodel56 said:
Interesting! Where is the data being pulled from? Wondering what it would look like if strength of schedule was based on wins (say weight by differential up to 3 goals) since a lot of teams don't go for the blowout wins... that also tends to skew the data, if you're playing weaker teams during the season or weaker tournaments, you're going to have a much higher SRS.
Click to expand...
The data is pulled from each league and tournament website. I think a max goal differential of 3 would be way too low as the rankings takes into account 159 G2008 teams in Calsouth and with no max goal differential, the max "spread" is Strikers with a 22 goal spread over Team #159. Capping the limit on goal differential will punish teams that play in weaker leagues, SDDA, and Coast for example and favor those teams in SCDSL that play teams at their level every week. For fun, I ran the algorithm at a max of 6 goal differential and compared it to no max differential for the T30 below:

max6.png
 
Carlsbad10 said:
The data is pulled from each league and tournament website. I think a max goal differential of 3 would be way too low as the rankings takes into account 159 G2008 teams in Calsouth and with no max goal differential, the max "spread" is Strikers with a 22 goal spread over Team #159. Capping the limit on goal differential will punish teams that play in weaker leagues, SDDA, and Coast for example and favor those teams in SCDSL that play teams at their level every week. For fun, I ran the algorithm at a max of 6 goal differential and compared it to no max differential for the T30 below:

View attachment 590
Click to expand...
Really well done, and I agree completely that the CSL teams playing in unpartitioned brackets (all teams are placed in bronze at this age group) do better in the "no-cap" table because they are able to run up scores against some very weak teams in their bronze brackets.

Any chance of reposting the entire table for all teams? Capped or uncapped? It will be very interesting to compare them pre and post State Cup.
 
Daniel Miller said:
Really well done, and I agree completely that the CSL teams playing in unpartitioned brackets (all teams are placed in bronze at this age group) do better in the "no-cap" table because they are able to run up scores against some very weak teams in their bronze brackets.

Any chance of reposting the entire table for all teams? Capped or uncapped? It will be very interesting to compare them pre and post State Cup.
Click to expand...
G08full.png
 
You must log in or register to reply here.
Back
Top