Most - and Least - Improved 05 Teams in 2018-19 (San Diego Group)

Kante

PREMIER
Here’s which San Diego teams have improved the most – and the least – from last season (2017-18) to this season (2018-19) so far. The table below compares offensive and defensive effectiveness of teams – i.e. how they do relative to how other teams, on average, perform.

For example, on the offensive side, in the table below, the most offensively improved team so far this year - the Strikers' - has an offensive effectiveness in 2018-19 is listed as 34%. This means that – on average – the Strikers will score 1.34x the number of goals that the team they’re playing typically allows to be scored.

So, in their most recent game on 10/20 , the Strikers played OC Surf. who on average this year have been allowing 3.67 goals to be scored. And in their match on 10/20, the Strikers scored 6 goals.

According to the offensive effectiveness algo, the Strikers were on track to score – 1.34 x 3.67 - 4.92 goals.

upload_2018-10-31_9-2-12.png

On the defensive side, Albion allows teams, on average to score 1.51x the # of goals they would typically score, the worst in the group.

In Albion's match this last weekend against Arsenal, Arsenal scored three goals. But, on average in 2018-19, Arsenal had only been scoring .83 goals per game to that point.

According to the defensive effectiveness algo, playing against Albion, Arsenal was on track going into the game to score – 1.51 x .83 – 1.25 goals.

The table ranks the teams according to the change in 2018-19, comparing how they are doing in 2018-19 for the first 5-7 games of 2018-19 vs how they did in the same number of games at the beginning of the season in 2017-18.

Both the offensive and defensive algo’s need more fine tuning (they’re updated every week), but are starting to get closer to being predictive, and the numbers can provide a pretty good idea of which teams are getting better and worse relative to last year.

By the way, the goal of providing this info is to give parents and families - who spend significant time and money on all this - more visibility/understanding into what's going on with their sons' teams. As opposed to having to read the tea leaves.
 
Here’s which San Diego teams have improved the most – and the least – from last season (2017-18) to this season (2018-19) so far. The table below compares offensive and defensive effectiveness of teams – i.e. how they do relative to how other teams, on average, perform.

For example, on the offensive side, in the table below, the most offensively improved team so far this year - the Strikers' - has an offensive effectiveness in 2018-19 is listed as 34%. This means that – on average – the Strikers will score 1.34x the number of goals that the team they’re playing typically allows to be scored.

So, in their most recent game on 10/20 , the Strikers played OC Surf. who on average this year have been allowing 3.67 goals to be scored. And in their match on 10/20, the Strikers scored 6 goals.

According to the offensive effectiveness algo, the Strikers were on track to score – 1.34 x 3.67 - 4.92 goals.

View attachment 3328

On the defensive side, Albion allows teams, on average to score 1.51x the # of goals they would typically score, the worst in the group.

In Albion's match this last weekend against Arsenal, Arsenal scored three goals. But, on average in 2018-19, Arsenal had only been scoring .83 goals per game to that point.

According to the defensive effectiveness algo, playing against Albion, Arsenal was on track going into the game to score – 1.51 x .83 – 1.25 goals.

The table ranks the teams according to the change in 2018-19, comparing how they are doing in 2018-19 for the first 5-7 games of 2018-19 vs how they did in the same number of games at the beginning of the season in 2017-18.

Both the offensive and defensive algo’s need more fine tuning (they’re updated every week), but are starting to get closer to being predictive, and the numbers can provide a pretty good idea of which teams are getting better and worse relative to last year.

By the way, the goal of providing this info is to give parents and families - who spend significant time and money on all this - more visibility/understanding into what's going on with their sons' teams. As opposed to having to read the tea leaves.

One way to provide more visibility and understanding would be to start by presenting the raw data - the WLT records before your "algos" minced them into incomprehensibility.
 
Below is win/loss for 2018-19 for reference.

The team improvement stats in the original post are to help folks understand progress made and/or lack of progress made by teams. Many times, wins and losses don't tell the whole or accurate story.

For example, if SD Surf beats the last place team in the SD group by score of 4-3, that’s a step backward for SD Surf because they should have won that game by a much larger margin. Conversely, for the last place team, despite losing, they should look at the result as big step forward.

Espola, I apologize that you found the stats incomprehensible, and will work to simplify them.


(P.S. Most folks on this forum are trying to be constructive and helpful, myself included)

upload_2018-11-1_9-55-38.png
 
Last edited:
Below is win/loss for 2018-19 for reference.

The team improvement stats in the original post are to help folks understand progress made and/or lack of progress made by teams. Many times, wins and losses don't tell the whole or accurate story about progress - or lack of progress - being made by teams.

For example, if SD Surf beats the last place team in the SD group by score of 4-3, that’s a step backward for SD Surf because they should have won that game by a much larger margin. Conversely, for the last place team, despite losing, they should look at the result as big step forward.

Espola, I apologize that you found the stats incomprehensible, and will work to simplify them.


(P.S. Most folks on this forum are trying to be constructive and helpful, myself included)

View attachment 3333

I never heard of "should have" as a statistical concept.
 
All good stuff, I like the stats...it's fun.

However while fun, I just wish the data was better or at least more complete to show an more accurate picture...perhaps by the end of the year it will look better and cover the anomalies. One such issue is that I don't think every team plays to win every game (games don't count for anything). An example would be, taking your scenario from above, that Surf beats a last place team 4-3 when they should have won 5-1 or something. Well what if the team, Surf in this case, decides to take the opportunity to play the Subs or some 06s...throws off your trends a bit (at least in the short run).

Right now there really isn't enough data to really get some good results. (OC Surf; they have played Surf twice, skewing their results a bit negative as it relates to the other teams). Let's watch as the season progresses.

Thanks again, always entertaining to see what is happening in SoCal.

Would be cool to do something like this with the Showcase data. SoCal v. NorCal etc...
Would also be cool if the DA would use data like this to build their groupings...I think we will continue to see the separation grow between the top clubs and the rest.
 
Fair points.

In San Diego, there's teams who have only played four or five games so far this year, so, yeah, small sample size. And there's anomalies, e.g., SD Surf vs LAGSD. SD surf tied LAGSD 4-4 but didn't have the three players from their starting front line, so not really fair to SD Surf to say they underperformed, and not accurate to say that LAGSD over-performed.

And agreed that things like teams not necessarily playing to win every game do occur, and might show disproportionally early on but get normalized over the course of the season.

The showcase ask for other teams is tougher since I haven't completely automated the data analysis. However, I do have some nuggets that I'll add to the showcase thread.

http://www.socalsoccer.com/threads/november-05-showcase-predictions-and-commentary.16168/
 
You think it's coocoo, yet you'd never even heard of it as of yesterday? It's just a simple algorithm that weighs teams based on how well they do. No different from keeping standings except it's the only way to compare across leagues that have limited contact. It's not perfect (and given that youth teams form and break up a lot, and some leagues like the DA are closed, there are holes in the data), but it ends up being a pretty good predictor of outcomes.
 
Back
Top