Lacrosse Analytics

D1 Mens Lacrosse
Post Reply
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Lacrosse Analytics

Post by laxreference »

Wanted to set ourselves up over at this new forum (RIP LaxPower) so that we have somewhere to discuss the quant angles to the 2019 season.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

First thing we wanted to do this year was to get the strength of schedule calculations done for each D1 team. Same as before, we used our Lax-ELO ratings to come up with an average S.O.S. for each team. The top 10 turned out to be a fairly predictable lot:
SOS_visualization(20190107).png
SOS_visualization(20190107).png (26.67 KiB) Viewed 6578 times
Of course, calculated this way, SOS is equal parts a measure of how you've chosen to schedule and the games you are forced to play because of conference ties. To get a clearer picture of the teams that went out of their way to schedule a difficult slate, we can recalculate the rankings without any conference games included:
SOS_NonConf_visualization(20190108).png
SOS_NonConf_visualization(20190108).png (25.15 KiB) Viewed 6578 times
The full write-up of this analysis can be found here.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
Drcthru
Posts: 555
Joined: Mon Jul 30, 2018 5:57 pm
Location: East bank of the lower Willamette

Re: Lacrosse Analytics

Post by Drcthru »

Interesting that neither Yale or UVA is on either list!
Everyone wants to change the world but, no one wants to do the dishes.
bearlaxfan
Posts: 893
Joined: Wed Aug 29, 2018 2:38 pm

Re: Lacrosse Analytics

Post by bearlaxfan »

Some of the non-conference weakness is due to regionality. The immediate OOC opponents for, as an example, Yale, will be other Connecticut schools that would not usually be in the top 15 in the polls, while UMd has Towson and Loyola as consistently polling programs, with UMBC the outlier. How many athletic depts will commit the $ to travel so its lax program can get to the opponents needed to build up SOS, if travel is required? Some, but not every.
Typical Lax Dad
Posts: 32144
Joined: Mon Jul 30, 2018 12:10 pm

Re: Lacrosse Analytics

Post by Typical Lax Dad »

Mid week games have a regional impact. Stiffer competition in mid Atlantic vs Northeast.
“You lucky I ain’t read wretched yet!”
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

bearlaxfan wrote:Some of the non-conference weakness is due to regionality. The immediate OOC opponents for, as an example, Yale, will be other Connecticut schools that would not usually be in the top 15 in the polls, while UMd has Towson and Loyola as consistently polling programs, with UMBC the outlier. How many athletic depts will commit the $ to travel so its lax program can get to the opponents needed to build up SOS, if travel is required? Some, but not every.
Great point. I guess you really need to account for 2 factors to really isolate the program's choices: conference strength and proximity of talented opponents. Fair point.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

Since there isn't any data with which to analyze the new shot clock (until Friday...), I wanted to lay out a framework for thinking about how it plays out. The central question of this season is whether the shot clock ultimately ends up benefiting offenses or defenses. And of course, this is down to many different factors. It will also play out with different teams in difference ways.

That said, there are a few things that we do know. First of all, the number of shots per game has been declining over the past 6 years, albeit by only a few shots. To me, this indicates a general increase in conservatism on the offensive end, probably owing to the rise in uber face-off guys.

Another helpful trend is that the number of goals scored per game has been extremely consistent.
GoalspGame.png
GoalspGame.png (4.36 KiB) Viewed 6166 times
That is helpful because it makes it more likely that an increase or decrease in goals per game can be attributed to the new rules.

I also think that this is a good example of the limits of analytics and data analysis in general. Everyone would love to know the optimal way to attack the new rules. But with no data to work off of, advanced stuff has no clue. Thankfully, there are lots of experienced coaches to lean on...

The full post is here.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

I did a poll to see what people thought of the new shot clock rules. Specifically, the question was around whether we expect to see more or less goals, on average, because of the new rules. Personally, my gut is more goals. 75% of the voters agreed. Frankly, no one knows though until the games get underway...TOMORROW!!!!!

Still, there are some empirical ways that we can think about the question of the shot clock. For example, of our top 20 offenses, I ranked them based on the percentage of their goals that came after 80 seconds of possession. Of course, personnel changes from year to year, but to the extent that these numbers are based on program tendencies, you might expect the teams that relied on longer possessions to have to adapt more than those that didn't.

So keep an eye on these teams as we kick off the season tomorrow.
2018HighestPctOfGoalsafter80secondsTop20OffEff (1).png
2018HighestPctOfGoalsafter80secondsTop20OffEff (1).png (28.09 KiB) Viewed 6090 times

The full post is here.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
Henpecked
Posts: 1128
Joined: Wed Aug 29, 2018 9:02 am

Re: Lacrosse Analytics

Post by Henpecked »

Laxreference,

This statistic is fascinating to me. Having watched every game Delaware played last year, I noticed that there was absolutely no urgency to get the offense started quickly. DeLuca essentially ran one midfield line out there the whole season and they would just casually circle the ball two or three times around the horn till the refs invariably put on a shot clock.

I saw the same thing with Bucknell who ran primarily three middies all season. Will Sands would get two or three touches behind the net before launching into the offense. In both cases the teams seemed to be milking the clock and getting middies much-needed rest for the first minute of every possession. And in both cases, these long possessions often ended in goals.

Having watched some Delaware games in the fall and this spring, I have noticed a lot quicker pace and a not too discernible drop off in offensive efficiency. I have also seen many more combinations at midfield which is probably necessary to play at this pace.

Interesting to see these stats.
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

Henpecked wrote: Thu Jan 31, 2019 4:36 pm This statistic is fascinating to me. Having watched every game Delaware played last year, I noticed that there was absolutely no urgency to get the offense started quickly. DeLuca essentially ran one midfield line out there the whole season and they would just casually circle the ball two or three times around the horn till the refs invariably put on a shot clock.
Always good when the numbers bear themselves out in real life. Your point is exactly why I tend to think that the new clock will increase overall goals. My working hypothesis, and this has been borne out by some of the other work I've done, is that coaches are more likely to artificially slow their offenses than encourage them to go fast. And I suspect it's because coaches are a conservative species by default, so the chance to rest your tired legs while not necessarily hampering your offense was too good to pass up.

Not having that luxury means offenses will tend to play at their more "natural" pace this year, which shouldn't cause a decrease in efficiency. That's the theory anyway.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

I've got the live win probability engine up and running for the Utah/Vermont game if anyone is interested. It takes the game stats as they come in, calculates the win probability of both teams and presents it, along with a running tally of the players with the highest game scores. Assuming everything goes to plan that is, haha...
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

Here are all the live win probability links from today's games. We were able to get it working for all but two games.

Updated list of Live Win Probability Links

- Furman vs Duke
- Mercer vs North Carolina
- Maryland vs Bucknell
- Rutgers vs Lafayette
- Penn State vs Villanova
- Ohio State vs Cleveland State
- Detroit vs Jacksonville
- High Point vs St. Bonaventure
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

I went through, as I do every week, and calculated the top statistical performances of opening weekend. The idea here is to distill the box score stats down into something that can be compared via a single number. I use the expected goals added (EGA) concept that we came up with a few years back.

Full post is here, but a few highlights were:
  • Josh Stout becomes the first Utah lacrosse player to ever grace our list of top individual performers. Pressure's on Bonnies...
  • Alex Jarzembowski from UD Mercy put in a strong performance at X (15 FO wins) to help the Titans overcome Jacksonville in one of the better games of the weekend
  • Rutgers and Duke are well represented, which you might expect in games where their offenses did pretty much whatever they wanted.
And the nifty infographic...
PlayerGame_visualization(20190206).png
PlayerGame_visualization(20190206).png (58.27 KiB) Viewed 5802 times
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

Win probability engine is up and running for Duke vs High Point.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

I saw some discussion over on the Cornell boards about my Wisdom of the Crowd post. They had asked to see the breakdown of "% of goals scored after 80s" for all the NCAA tournament teams from last year.

So I expanded the post a bit to incorporate a few more cuts of the data. You can find it here. And here is that specific visualization...

2018HighestPctOfGoalsafter80secondsNCAATeams.png
2018HighestPctOfGoalsafter80secondsNCAATeams.png (40.24 KiB) Viewed 5690 times
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

In case anyone is watching and wants to do some 2nd screening, I've got the win probability page up and running for Syracuse vs Colgate. Nail biter for all you Orange fans.
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
laxreference
Posts: 1099
Joined: Wed Aug 29, 2018 3:54 pm
Contact:

Re: Lacrosse Analytics

Post by laxreference »

Win probabilities are live for Bucknell vs Bryant
Data Engineer/Lacrosse Fan --- Twitter: @laxreference --- Informed fans get Expected Goals, the new daily newsletter from LacrosseReference
Post Reply

Return to “D1 MENS LACROSSE”