Youth Soccer Rankings ?

[sigmabody]

If you name it by grad year, then you'll have parents who think their older kid can play in a younger division. "But he's in 8th grade...." No one working the signup table wants to deal with that ten times every hour.

My guess is we refer to them by the pair of years, same as we did 10 years ago. "14/15" means kids born the second half of 2014 or the first half of 2015.

Yep, agree with this. YY/YY is the least confusing representation, and users will know how to parse this.

I know "red shirted" kids who are a year behind in school from the start, even before rarer (but possible) grade movements. Additionally, the cutoff point doesn't 100% align with school enrollment cutoffs points anyway. School year would be far more confusing.

 

[Carlsbad7]

Yep, agree with this. YY/YY is the least confusing representation, and users will know how to parse this.

I know "red shirted" kids who are a year behind in school from the start, even before rarer (but possible) grade movements. Additionally, the cutoff point doesn't 100% align with school enrollment cutoffs points anyway. School year would be far more confusing.

Totally doable and not confusing...

 

[sigmabody]

Totally pedantic, but short category ordering from least to most confusing, in my entirely subjective opinion:
- YY/YY
- U-[Age]
- "School" year

Spelling out the date ranges explicitly is the least confusing overall, but not really "app column header" or "short team name postfix" compatible. I'd guess most people on this forum could parse/understand any of them, but it might be more challenging for average soccer parents.

 

[Code]

If you name it by grad year, then you'll have parents who think their older kid can play in a younger division. "But he's in 8th grade...." No one working the signup table wants to deal with that ten times every hour.

My guess is we refer to them by the pair of years, same as we did 10 years ago. "14/15" means kids born the second half of 2014 or the first half of 2015.

For the very rare player who was held back or graduated early I think the registration table could handle.

 

[jasportsdad]

If you name it by grad year, then you'll have parents who think their older kid can play in a younger division. "But he's in 8th grade...." No one working the signup table wants to deal with that ten times every hour.

My guess is we refer to them by the pair of years, same as we did 10 years ago. "14/15" means kids born the second half of 2014 or the first half of 2015.

Yep, agree with this. YY/YY is the least confusing representation, and users will know how to parse this.

I know "red shirted" kids who are a year behind in school from the start, even before rarer (but possible) grade movements. Additionally, the cutoff point doesn't 100% align with school enrollment cutoffs points anyway. School year would be far more confusing.

Totally pedantic, but short category ordering from least to most confusing, in my entirely subjective opinion:
- YY/YY
- U-[Age]
- "School" year

Spelling out the date ranges explicitly is the least confusing overall, but not really "app column header" or "short team name postfix" compatible. I'd guess most people on this forum could parse/understand any of them, but it might be more challenging for average soccer parents.

Agreed. Now how do we make YY/YY happen lol

 

[RandomSoccerFan]

As of 8/1, the start of the new season, the new age groups are now properly represented in SR. I imagine it will take a little while to get everything sorted out, as a number of teams seem to have some of their results tied to the older or younger team. Also, quite a number of teams that are now listed with the "12/11" type of naming scheme, aren't attached to their prior season team results. As more and more results come in over time with the new names, everything will work itself out one way or another, whether or not people are joining unlinked results.

 

[RandomSoccerFan]

Looks like the developers have now put a cap on how many detailed team lookups you can do per day. I've run into it twice now, when looking through some large clubs this new season and cleaning up some of the new unranked team entities. The counter resets every night and restores detailed access, but I wonder if any of their other unhealthily obsessed users are hitting it as well.

 

[RandomSoccerFan]

Slight correction, in case it happens to anyone else. It does not automatically reset the next day, you need to send them an email to reset it (which they seem to do promptly). Include your userID, which is shown if you go into the app settings - it lets you copy & paste it into an email.

 

[RandomSoccerFan]

I was trying to understand how much the changes in age group timing and the associated roster churn were going to affect the accuracy/predictiveness of SR this fall. I think it was generally assumed it would go down for awhile, but it was hard to estimate whether it would be by a little or by alot, and how soon it would be until it recovers as the new team setups get some games under their belt. I was pleasantly surprised.

SR was kind enough to share current predictivity #'s to figure out how big the drop in predictivity is right now due to all of thes data changes. These measurements are from games over the past 2 weeks. TL;DR; - It's both measurable and noticeable, but it's not a large amount, and is expected recover quickly.

Last season it was running at ~85% games called correctly (.68 predictivity), with age groups ranging from 82% to 94%. Right now it's running at 82.5% games called correctly (.65 predictivity), with age groups ranging from 70% to 92%.

 

[Karlton Athletic]

I was trying to understand how much the changes in age group timing and the associated roster churn were going to affect the accuracy/predictiveness of SR this fall. I think it was generally assumed it would go down for awhile, but it was hard to estimate whether it would be by a little or by alot, and how soon it would be until it recovers as the new team setups get some games under their belt. I was pleasantly surprised.

SR was kind enough to share current predictivity #'s to figure out how big the drop in predictivity is right now due to all of thes data changes. These measurements are from games over the past 2 weeks. TL;DR; - It's both measurable and noticeable, but it's not a large amount, and is expected recover quickly.

Last season it was running at ~85% games called correctly (.68 predictivity), with age groups ranging from 82% to 94%. Right now it's running at 82.5% games called correctly (.65 predictivity), with age groups ranging from 70% to 92%.

My eldest son’s team combined a league winning older team with a much weaker younger team for the age change, with the majority of the players coming from the good team. The Rankings app has the weaker team’s history tied to the new team, meaning we have a much weak ranking.

It will obviously correct itself in time, but I like to think other team’s are underestimating us as a result.

 

[Jamisfoes]

In tennis, they have UTR tournaments. These are events put up by the player rating company to promote players playing each other. In these events you have to play a player 1 or 2 points under your rating and a player 1 or 2 points above your rating. In tennis rating is given from 1 to 16 (16 being the top pros, 12-14 being college players)) In the end you get a wealth of data points to make your prediction model more confident.

Do you think soccer rankings will ever get big enough in the future to have something like that?

 

[RandomSoccerFan]

No, I don't. SR is already "big" enough to cover almost all competitive youth soccer teams in the country. It's already as accurate as it can be, given the constraint of only knowing final game score, and data quality issues in assigning game history to every team. It can't get much "better" at it without adding usable information in addition to what it already has. But a pretty small minority of people involved (from clubs/coaches/parents/players) have any knowledge that it exists, and even when they do, many see no reason to use ratings/rankings in the first place.

The UTR systems listed aren't much more accurate in terms of predicting a winner, but they are very much accepted by the various sports to be a part of the environment. It makes sense to set up tournaments between people by rating, and people involved completely understand.

Someone on this site, maybe even in this thread, had suggested that there be a yearly tournament where the top 20 teams nationally in SR were invited to an actual national playoff. I think that it would be a pretty fantastic (and fair) way to identify the very best teams in the country. I also think that it's never going to happen.

 

[espola]

In tennis, they have UTR tournaments. These are events put up by the player rating company to promote players playing each other. In these events you have to play a player 1 or 2 points under your rating and a player 1 or 2 points above your rating. In tennis rating is given from 1 to 16 (16 being the top pros, 12-14 being college players)) In the end you get a wealth of data points to make your prediction model more confident.

Do you think soccer rankings will ever get big enough in the future to have something like that?

Another advantage of local tennis tournaments -- players' ages are measured on the day of the tournament. No early-in-the-year birth advantages.

 

[espola]

No, I don't. SR is already "big" enough to cover almost all competitive youth soccer teams in the country. It's already as accurate as it can be, given the constraint of only knowing final game score, and data quality issues in assigning game history to every team. It can't get much "better" at it without adding usable information in addition to what it already has. But a pretty small minority of people involved (from clubs/coaches/parents/players) have any knowledge that it exists, and even when they do, many see no reason to use ratings/rankings in the first place.

The UTR systems listed aren't much more accurate in terms of predicting a winner, but they are very much accepted by the various sports to be a part of the environment. It makes sense to set up tournaments between people by rating, and people involved completely understand.

Someone on this site, maybe even in this thread, had suggested that there be a yearly tournament where the top 20 teams nationally in SR were invited to an actual national playoff. I think that it would be a pretty fantastic (and fair) way to identify the very best teams in the country. I also think that it's never going to happen.

Who would pay for all that travel?

 

[RandomSoccerFan]

Right, that's one huge issue right there. Of all the issues to deal with though, I'm not sure that's the largest. Many top ECNL teams are already hopping on a plane 3-4 times a year, top MLS N teams maybe even a bit more. Adding 1 additional tournament to that doesn't change the cost structure for them much, and this is assuming it doesn't take the place of an existing trip.

 

[Jamisfoes]

No, I don't. SR is already "big" enough to cover almost all competitive youth soccer teams in the country. It's already as accurate as it can be, given the constraint of only knowing final game score, and data quality issues in assigning game history to every team. It can't get much "better" at it without adding usable information in addition to what it already has. But a pretty small minority of people involved (from clubs/coaches/parents/players) have any knowledge that it exists, and even when they do, many see no reason to use ratings/rankings in the first place.

The UTR systems listed aren't much more accurate in terms of predicting a winner, but they are very much accepted by the various sports to be a part of the environment. It makes sense to set up tournaments between people by rating, and people involved completely understand.

Someone on this site, maybe even in this thread, had suggested that there be a yearly tournament where the top 20 teams nationally in SR were invited to an actual national playoff. I think that it would be a pretty fantastic (and fair) way to identify the very best teams in the country. I also think that it's never going to happen.

I think when the top MLS next team and the top ECNL team 5 miles from each other haven’t played each other in 3 years, it’s a missing data point that could be improved by having them play each other.

It doesn’t have to be a tournament that requires traveling. It could be a soccer rankings sponsored Local scrimmages.

 

[RandomSoccerFan]

Competing top leagues have zero incentive to play each other at the moment, and quite a bit of incentive not to. Soccer Rankings isn't going to be the issue that turns the tide.

A soccer team plays 15-30 different opponents for the entire year, and has "missing data" when compared to all other xxx,xxx teams. That doesn't stop them from being ranked/rated within the entire pool. There will always be the argument that "Well, they didn't play directly so we can't be sure". Which will never be resolved to anyone's satisfaction, regardless of any math/statistics/probability that one chooses to fully grasp or instead to ignore.

 

[Jamisfoes]

Competing top leagues have zero incentive to play each other at the moment, and quite a bit of incentive not to. Soccer Rankings isn't going to be the issue that turns the tide.

A soccer team plays 15-30 different opponents for the entire year, and has "missing data" when compared to all other xxx,xxx teams. That doesn't stop them from being ranked/rated within the entire pool. There will always be the argument that "Well, they didn't play directly so we can't be sure". Which will never be resolved to anyone's satisfaction, regardless of any math/statistics/probability that one chooses to fully grasp or instead to ignore.

How did we get here where top teams are afraid/ need incentives to play each other? This is the why other countries laugh at us.

 

[Larzby]

How did we get here where top teams are afraid/ need incentives to play each other? This is the why other countries laugh at us.

I think the top teams from the lower-ranked league would happily play the teams from the higher-ranked league. The reverse might also be true, as well, but the leagues discourage it by having intra-league events (champions leagues, showcases, etc) that take away many of the free weekends that might otherwise be used for inter-league games, and also drain the pocketbooks of the families. It's honestly brilliant. Also infuriating.

 

[Larzby]

In tennis, they have UTR tournaments. These are events put up by the player rating company to promote players playing each other. In these events you have to play a player 1 or 2 points under your rating and a player 1 or 2 points above your rating. In tennis rating is given from 1 to 16 (16 being the top pros, 12-14 being college players)) In the end you get a wealth of data points to make your prediction model more confident.

Do you think soccer rankings will ever get big enough in the future to have something like that?

Matches between tennis players of different UTRs happen all the time, not just a result of tourneys especially set up for that to happen. If you sign up for a tourney you're not ready for, you lose 6-0, 6-0, in the first round and then go back and try to find events closer to your skill level. Youth tennis has some problems (mostly with access), but at least there's no way to hide from the other top players (other than by calling balls out, but that eventually catches up to people)

 

[Jamisfoes]

I think the top teams from the lower-ranked league would happily play the teams from the higher-ranked league. The reverse might also be true, as well, but the leagues discourage it by having intra-league events (champions leagues, showcases, etc) that take away many of the free weekends that might otherwise be used for inter-league games, and also drain the pocketbooks of the families. It's honestly brilliant. Also infuriating.

Even outside of league season, what I see this summer is increasingly mls next and ECNL teams are hiding in predominantly mls next or ecnl tournaments. It’s like the two leagues are afraid of playing each other. I am tired of playing the same teams that we already face during the season.

Soccer ranking can help promote inter league competition by encouraging teams in close proximity to play each other once a year. They can even change the algorithm to place more importance on results between opposing leagues.