G05 CAS Rankings

Due to the birth year age group change previous youth soccer rankings are no longer any good. I am attempting to build a realistic ranking model that eventually tournaments can use to help seed teams. The methodology uses the excel solver and a "best fit" algorithm that takes as input scores from all tournaments and league play to then come up with a "strength" number for each team. The difference in strength numbers tells you how many more goals better one team is than another. Due to small sample sizes early in the season there will be a few strange rankings. As more games are played things will look more normal. Teams improve their ranking by winning by more than the expected goal differential or losing by less than the expected differential (basically doing better than expected). Teams move down in the rankings by doing the opposite. I have done this for the G04 rankings and I now have a volunteer who is keeping track of all the G05 scores so I can post rankings for this age group too. We have tried to minimize mistakes (we are hand inputting scores from tournament websites) which is hard to do due to teams being listed with slightly different names in each tournament. I will only list the top 200 teams. The SRS column tells how strong your team is compared to other teams. Only games played between Cal South teams are considered. In order to include Cal North or games/teams from other regions you would need to include all teams from those regions to make rankings accurate and that is not a task I am will to take on. Game recorded begin May 7th and include the CRL Play-In games.

Here is the link:
http://coachingkidssoccer.blogspot.com/2016/08/cas-g05-youth-soccer-rankings-1.html
 
Thank you Tango,
This is great for us parents that want to keep track on where teams are without studying every tournament and in my opinion give a better ranking than got soccer which is based only on specific tournaments. A team that enters a strong tournament and gets knocked out after tying two games against top teams will get 0 points on got soccer, but these algorithms will give that same team some love for doing better than expected.

I'm not sure if you have seen the LastPlanetrankings page, but it has some nice tools you may get some tips from. I still look at this from time to time but as of now it is only for 01 through 04 girls. http://lastplanetranking.blogspot.com/2013/11/cas-g01-ratings.html

It has a match predictor so you can put any two teams through this thing and it will essentially give you a line on the game similar to gambling lines which I think is pretty neat. http://lastplanetranking.blogspot.com/2013/05/match-predictor.html

Thanks again for sharing this and the hard work that I'm sure was put into doing it!!
 
Nice work Tango, I know that takes a lot work and time. Keep up the good work it's still very early in to the soccer season, most of those team I have never heard of. Are you including CRL from last weekend?
 
Yes - includes CRL games from last weekend. If anyone sees a top team missing or anything look obviously wrong in the data send me a PM and I will take a look. You can also get a goal/point spread with these rankings by subtracting the 'Ratings' column from each other. Of course this is more accurate with higher sample sizes (more games played).
 
Love it. Like to check out the teams we are going to play before hand. FYI the Corona/Extra team at #91 is Milan Academy that recently played - and didn't score :( at the AC Brea kickoff. All future games will be under the Milan Academy name. We didn't want to pay extra money to get the name change early during the summer.

Interesting that missing our best offensive player had such a big impact on the team as I thought it is a pretty balanced roster. That and a lot of bad first touches and lack of aggressiveness. Weird you watch practice and they look great, then the game happens. I guess that happens when they are 10/11. (or even the olympics I was amazed at how many mistakes were made in the women's game).

Shout out to CDA Slammers Whittier, they kept their spacing, played support. Fun to watch. Felt good about losing to them 1-0.
 
Good stuff, Tango. Seriously.

Are you able to link to each team's results as well rather than just showing the list?
 
Love it. Like to check out the teams we are going to play before hand. FYI the Corona/Extra team at #91 is Milan Academy that recently played - and didn't score :( at the AC Brea kickoff. All future games will be under the Milan Academy name. We didn't want to pay extra money to get the name change early during the summer.

Interesting that missing our best offensive player had such a big impact on the team as I thought it is a pretty balanced roster. That and a lot of bad first touches and lack of aggressiveness. Weird you watch practice and they look great, then the game happens. I guess that happens when they are 10/11. (or even the olympics I was amazed at how many mistakes were made in the women's game).

Shout out to CDA Slammers Whittier, they kept their spacing, played support. Fun to watch. Felt good about losing to them 1-0.

Who won that AC Brea Kickoff tournament?
 
Good stuff, Tango. Seriously.

Are you able to link to each team's results as well rather than just showing the list?

If someone helps me out with JavaScript code I could. Also could attach the spreadsheet here. I assume we have that ability on this forum ??
 
Really nice analysis. And the links to drills in the web page are also great.
Thanks. Bound to be a few teams misnamed or scores appointed to wrong teams on the first iteration. So if you see anything that looks really really off on team records since early May please shoot me a PM with what you think is wrong and I will take a look.
 
Due to the birth year age group change previous youth soccer rankings are no longer any good. I am attempting to build a realistic ranking model that eventually tournaments can use to help seed teams. The methodology uses the excel solver and a "best fit" algorithm that takes as input scores from all tournaments and league play to then come up with a "strength" number for each team. The difference in strength numbers tells you how many more goals better one team is than another. Due to small sample sizes early in the season there will be a few strange rankings. As more games are played things will look more normal. Teams improve their ranking by winning by more than the expected goal differential or losing by less than the expected differential (basically doing better than expected). Teams move down in the rankings by doing the opposite. I have done this for the G04 rankings and I now have a volunteer who is keeping track of all the G05 scores so I can post rankings for this age group too. We have tried to minimize mistakes (we are hand inputting scores from tournament websites) which is hard to do due to teams being listed with slightly different names in each tournament. I will only list the top 200 teams. The SRS column tells how strong your team is compared to other teams. Only games played between Cal South teams are considered. In order to include Cal North or games/teams from other regions you would need to include all teams from those regions to make rankings accurate and that is not a task I am will to take on. Game recorded begin May 7th and include the CRL Play-In games.

Here is the link:
http://coachingkidssoccer.blogspot.com/2016/08/cas-g05-youth-soccer-rankings-1.html
Impressive work!
 
Solver was not calculating ratings for 1/3 of teams for some reason. So I fixed the problem and refreshed the rankings. Should be much better now, but of course there are some teams with smaller sample sizes (few games played) that will get better slotted later down the road.

Top 10 (now looks like)
1. LAGSB Elite
2. SD Surf EGSL Academy
3. SoCal Blues Draluck
4. Strikers FC - OC EGSL
5. Eagles
6. Beach FC - LB White Perez
7. LA Galaxy San Diego Elite Gurley
8. SoCal Blues Kale
9. IE Surf Premier RB
10. TFE S.E. LA

Same link as above to see full list.
http://coachingkidssoccer.blogspot.com/2016/08/cas-g05-youth-soccer-rankings-1.html
 
Are EGSL spring league games included? How about games against 2004 teams?
Judging by the number of games for our team it doesn't look like the EGSL spring league games or games against 2004 are included. I'm sure that would move things around a little bit and I'm sure would at least improve our teams SOS.

Might be about time for a page just with past and future tournament links.
 
Currently, not included... (* because it would quadruple the complexity)
- EGSL Spring League games
- Friendlies
* Games against teams from other age groups (ie 04, 06)
* Games against other regions (ie - Cal North, Hawaii, Washington, Nevada, Arizona teams)

Included...
CRL Play-In games
CRL League games
All tournament games
Fall season games (when they start)
 
Due to the birth year age group change previous youth soccer rankings are no longer any good. I am attempting to build a realistic ranking model that eventually tournaments can use to help seed teams. The methodology uses the excel solver and a "best fit" algorithm that takes as input scores from all tournaments and league play to then come up with a "strength" number for each team. The difference in strength numbers tells you how many more goals better one team is than another. Due to small sample sizes early in the season there will be a few strange rankings. As more games are played things will look more normal. Teams improve their ranking by winning by more than the expected goal differential or losing by less than the expected differential (basically doing better than expected). Teams move down in the rankings by doing the opposite. I have done this for the G04 rankings and I now have a volunteer who is keeping track of all the G05 scores so I can post rankings for this age group too. We have tried to minimize mistakes (we are hand inputting scores from tournament websites) which is hard to do due to teams being listed with slightly different names in each tournament. I will only list the top 200 teams. The SRS column tells how strong your team is compared to other teams. Only games played between Cal South teams are considered. In order to include Cal North or games/teams from other regions you would need to include all teams from those regions to make rankings accurate and that is not a task I am will to take on. Game recorded begin May 7th and include the CRL Play-In games.

Here is the link:
http://coachingkidssoccer.blogspot.com/2016/08/cas-g05-youth-soccer-rankings-1.html

Tango, I noticed that you stated the rankings were based on CalSouth vs CalSouth games so I am confused as to why the EGSL Spring games were not included in the mix if all of those games were CalSouth teams? I figure the more data points you get into the algorithms the better. Also in the rankings can you include any movement from week to week in a separate column. Example from the most current ranking would be like Strikers OC EGSL slipped from 4th to 5th so that would mean a -1 in the movement column. Something similar to what they do on the NFL Power Rankings on ESPN. Thanks for doing this.
 
Tango, I noticed that you stated the rankings were based on CalSouth vs CalSouth games so I am confused as to why the EGSL Spring games were not included in the mix if all of those games were CalSouth teams? I figure the more data points you get into the algorithms the better. Also in the rankings can you include any movement from week to week in a separate column. Example from the most current ranking would be like Strikers OC EGSL slipped from 4th to 5th so that would mean a -1 in the movement column. Something similar to what they do on the NFL Power Rankings on ESPN. Thanks for doing this.

Thanks for the input. I was advised by some of the subject matter experts on these boards in private conversations not to go back and include EGSL Spring games as they were not considered serious games. I was told just going back to the CRL Play-In games was the appropriate thing to do. If you feel otherwise then make an argument for it and maybe others can chime in as well. It will take a "very" good argument to change.

There already is a column that shows the previous weeks rankings. If you prefer another column showing +3 instead of the 17 and 14 then I can do that. Just let me know what format you want. I can't add too many columns otherwise the formatting gets messed up.
 
I can't speak for any other teams and obviously some experts on the subject must know more than me, but I know our team took the games fairly seriously. Obviously it being spring league and most of these teams first real games with the new age group there was a lot of player movement and experimenting going on, but I don't know that should be considered as non serious games. Scdsl is based on developing and not supposed to have emphasis on winning, so would that also be considered non serious? I'm not one to say what you should and shouldn't track. All that you are doing is great and much appreciated, I'm just throwing in my 2 cents on the spring league from my perspective.
 
Back
Top