analytics.soccer: ECNL - SW - 05/04 - weekly game predicts

Kante

PREMIER
we're back... whew... it's been a bit.

the Kante algorithm is probably a little creaky, a little stiff etc. etc. but once it stretched out, it seems not to be worse for wear.

for those who aren't familiar, and asking "wtf is this?!?!?", we used to post game predicts every week for several age groups, mostly on the boy’s side. the goal was to apply some objective criteria to help families measure progress not by wins and losses but by how did their kids do relative to historical expectations.

if team A is playing team B, and team A is good and team B is not, winning is kind of meaningless. team A is expected to win.

instead, the algorithm predicts, based on historical data, that team A will score for example, four goals and that team B will be held scoreless.

if team B then either scores a goal or holds team A to three goals, that then is progress for team B relative to how team B has done in the past, and progress is good.
________

some caveats. the algorithm usually takes about ten games before it’s as accurate as it will get. So far, we have about six games for most ECNL 05/04 teams. You can do the math.

typically, after ten games, the algorithm will be accurate for goals scored by the home team about 75% of the time, and accurate for goals scored by the away team about 65% of the time.

so, the algo is not perfect (alas, since EPL allows betting…) and can be a little temperamental, but it’s often close enough to make things interesting.
________

speaking of interesting. Two items.

hearing back from parents in the comments section about how the games went after the fact also makes it interesting. So, the more feedback and commentary the better.

also, and maybe this will be interesting and maybe not. we – there's three of us- are working on a website/service that will provide some of the value that youthsoccerankings.us used to, guidance for college recruiting and some other things.

please go check us out at www.analytics.soccer. We’d appreciate your comments and feedback. It’s a work in progress, so a little messy, but we have some big ambitions (as you’ll see) and the more feedback we get, the better.
________

without further ado/verbosity, here’s predicts. First team listed is the home team.

Our Bring the Popcorn Match
(i.e., so competitive that you should bring the popcorn)

Sunday, 10/30: SDSC Surf vs LAFC SoCal => Predict: 1-2 LAFC SoCal winner


Our Play the Kids! Match
(i.e., match is not expected to competitive, thus expected winner should play its younger players up, no disrespect to other team. just is what it is…)

Sunday, 10/30: West Coast FC vs San Diego Surf => Predict: 0-4 SD Surf winner
________

Saturday, 10/29

Legends vs SDSC Surf => 1-2 SDSC Surf winner

Sporting CA vs. Slammers => 3-2 Sporting CA winner

Pats vs LA Breakers =>1-0 Pats winner

SD Surf vs Heat FC => 3-2 SD Surf winner

Del Mar vs LAFC SoCal => 1-2 LAFC SoCal winner


Sunday, 10/30

LA Breakers vs Slammers => 1-1 Tie

Beach FC vs Pats => 1-3 Pats winner

Del Mar vs Heat => 2-2 Tie
________

Good luck this weekend! (it’s nice to be back…)
 

datamec

BRONZE
Definitely excited to seeing you doing this again. I remember reading your posts for the 05 age group back in the day when the DA was in full swing. Was always impressed with how accurate the predictions were even when I thought they would be way off.

It will be interesting how the algorithm works for this age group as from my experience it is a crap shoot who will show up on any given weekend for each team. With school events, college visits, injuries, and normal life I know the team my son plays for we never know who will or won’t be available for any given game.
 

Kante

PREMIER
here's predicts vs actuals:

The Algo had a rough weekend, probably shaking off some rust from being out for so long. Like MJ when he went go play baseball for a bit...

Would love folks' feedback and commentary on the matches, and why the results may have differed from what was predicted. FYI, here's the cirteria we use for evaluating success:

1) Did the Algo predict the right result?
2) Did the Algo get the right margin of victory?
3) Did the Algo get the right home team score?
4) Did the Algo get the right away team score? (this, btw, is the hardest piece to predict

here's predict vs actuals. predicts are in black, actuals are in blue. all observations, unless otherwise stated, are from looking at the scores (i.e. we weren't there). Predicts for this weekend will be out tomorrow. Meanwhile, while you're waiting, go check out www.analytics.soccer

It's in process but we'd love you to poke around and tell us what you'd like to see...

____________

Our Bring the Popcorn Match
(i.e., so competitive that you should bring the popcorn)

Sunday, 10/30: SDSC Surf vs LAFC SoCal => Predict: 1-2 LAFC SoCal winner
Hope the popcorn got brought, this one looks like it was tight. LAFC SoCal won 1-0.

So, not so bad for the Algo to start things. Anyone who was there have any thoughts/comments?


Our Play the Kids! Match

(i.e., match is not expected to competitive, thus expected winner should play its younger players up, no disrespect to other team. just is what it is…)

Sunday, 10/30: West Coast FC vs San Diego Surf => Predict: 0-4 SD Surf winner

Yup, SD Surf is like the honey badger, they just don't care.... ;):eek::rolleyes::);)

SD Surf beat the spread, winning 5-0. Right about now, the Algo's feeling a little cocky... just saying
________

Saturday, 10/29

Legends vs SDSC Surf => 1-2 SDSC Surf winner

And then there's this result.. SDCS Surf making the algo look bad... yeah, they won, and yeah the Algo said they'd win. But, guys winning 6-1 when the Algo called a 2-1 win? That's just rude

(but tbh, never have liked the 05 Legends since, back in the day, they had a bunch of guys who were clearly like, three years older than their stated age, but that's a whole other thing from years back with no place in this family-oriented forum)

Sporting CA vs. Slammers => 3-2 Sporting CA winner
Geez, now the Algo is feeling little ridiculous... Sporting CA win 6-0

Pats vs LA Breakers =>1-0 Pats winner
Ok, this is better, Pats win 3-2. Algo got the result and the margin...

SD Surf vs Heat FC => 3-2 SD Surf winner
Genuine question to the peanut gallery. Anybody catch this game? Were the Heat missing any offensive players or mids?

Algo got result and home team score, SD Surf won 3-0


Del Mar vs LAFC SoCal => 1-2 LAFC SoCal winner
Am ambivalent here. Algo says he nailed it, but did he? LAFC Socal win 4-2. Maybe, kinda...

_____

So Saturday outcomes for the Algo?


results: Algo went 5-0 for results on Saturday
margin by predicted winning team: Algo went 1-4
home team score: Algo went 1-4
away team score: Algo went 1-4

Not incredibly awful, but not our standard. The Algo feels good about the results. but really, it feels more like the first time we asked the Algo for help betting on EPL games...

Sunday, 10/30


LA Breakers vs Slammers => 1-1 Tie
LA Breakers win 4-0.

&*^&*%&*%(^%

We mean, congrats! LA Breakers for a nice, unexpected win, you guys should feel really good about your progress

(8*(&*&(^(*&^*&^)


Beach FC vs Pats => 1-3 Pats winner
wtf! Beach wins 4-2.

(8*(&*&(^(*&^*&^)

seriously? we mean, congrats! LA Breakers Beach FC for a nice, unexpected win, you guys should feel really good about your progress


Del Mar vs Heat => 2-2 Tie

Evidently, the Heat got stuck in the same traffic that got the got, scrolls up... the Heat goal scorers on Saturday, so yeah guess a couple of players were missing.. Del Mar wins 2-0
______


So Sunday outcomes for the Algo?

results: Algo went 2-3 for results on Sunday
margin by predicted winning team: Algo went 1-4
home team score: Algo went 2-3
away team score: Algo went 1-4

Again, not incredibly awful, but not our standard. The Algo is also a little more humble...
______
So net net, work to do. More games will help the accuracy though. Predicts will be out tomorrow.

In the meantime, go check out
www.analytics.soccer go kick the tires a little, it's what we would call pre-alpha, but should give a flavor of where we'd like to go. let us know what you think and what you want to see... Thx!
 

Kante

PREMIER
here's predict for this weekend. hopefully, the algo does a bit better than last week.

lots of fairly tight games so no popcorn or play the kids games. good luck this weekend! (and yes, good luck even to the Legends!;))

Saturday
West Coast FC vs Del Mar: tie 1-1
SDSC Surf vs LA Breakers: tie 1-1
Legends vs Pats: Pats win 3-1
LAFC SoCal vs SD Surf: SD Surf win 2-0 (one of us might go to this game)
Beach FC vs Slammers: Beach win 4-1

Sunday
LA Breakers vs West Coast FC: tie 1-1
Slammers vs Del Mar: Del Mar wins 2-1
 

justneededaname

SILVER ELITE
Something that I think is hard for the algo to predict in this age group is the roster for any given week. Not sure if other teams have the same issue, but the PSAT, SAT, homecoming, and college visits have had a significant impact on the week-to-week results for our team.
 

justneededaname

SILVER ELITE
Science and the World Cup: how big data is transforming football

One clear lesson that has emerged from data analysis is that players shouldn’t shoot when they’re far from the goal. “If you look at any league in the world, the distance from where players have taken shots was much higher ten years ago,” Ramineni says. “That all happened because data-analytics people have started saying: ‘Why are you shooting from there? It’s only a 2% chance!’ ”

 
Top