We’ve calculated a lot of things around here. We have stats for the distribution of effort with a team. We split possessions up by how they start. We can tell you which games had the best (read: lowest) game tension scores. I could go on; we have a lot of metrics.
One thing we haven’t calculated, until now, is shooting percentage. I think it’s because shooting percentage is such a basic stat. You can find it, for each team, on the NCAA stats website. And because it didn’t seem novel, we didn’t see much reason to incorporate it into our stats stable.
And in one sense, that is right. Why should we produce a page on our site that shows you exactly the same information you can get elsewhere. The problem is that shooting percentage affects outcomes so much more than something like play shares.
So by not feeling the need to have an explicit page dedicated to shooting percentage, we accidentally overlooked a key component in understanding why wins and losses happen.
The great eraser
Some of you are probably sitting there thinking: “wow, this guys not too bright.” And that is a fair sentiment. But at least we have seen the light. And it came in the form of a Dragon and Panther.
See, in my never-ending quest to provide useful quantitative information to lacrosse fans, I started doing game recaps with a very stats-heavy undercurrent. The first two that I did were for High Point’s win over Virginia and Drexel’s win over Michigan.
And in both recaps, I noticed that the loser won all sorts of statistical categories. UVA took more shots than High Point (47 to 40). Michigan won more faceoffs than Drexel (22 to 17) . UVA committed fewer turnovers (17 vs 25). The Hoos had more TOP (54% to 46%). Michigan picked up more groundballs than Drexel (47 to 35).
In short, so many of the things that we think of as being crucial to winning were in favor of Michigan and Virginia, which runs counter to the fact that both Drexel and High Point actually won the games.
And when we drilled into that, it was clear that the reason was shooting percentage. High Point shot better than Virginia (35% to 28%). Drexel shot better than Michigan (38% to 32%).
With the help of some posters over at FanLax, it became clear to me that I needed to incorporate shooting percentage into LacrosseReference in a more formal way.
Adjust it to trust it
And that was my morning. It is easy enough to calculate shooting percentage for each team, but a) that exists elsewhere and b) it doesn’t really tell the story well enough.
For starters, shooting percentage is an offensive metric that is influenced by the opponent’s defense. Conversely, the shooting percentage that a team allows is as much a factor of their defense as the opponents offense.
The advantage from shooting percentage is really a collective advantage. How well do you do in putting shots past their goalie and how well does your goalie do in preventing their shots from going in?
So it only makes sense to create an adjustment to shooting percentage to reflect that. The simple way to do it is to determine how much better you are than the average. Creating a metric that combines defense and offense is a little tougher.
What we settled on is the following.
- Step 1 – Calculate a team’s offensive shooting percentage
- Step 2 – Determine the gap between a team’s defensive shooting percentage allowed and the average team’s offensive shooting percentage
- Step 3 – Add this gap to the team’s offensive shooting percentage to get our adjusted shooting percentage
In practice, it works like this. Let’s say that the average team’s shooting percentage is 28% (we are not weighting by shots taken). And then let’s say that Team A’s shooting percentage is 24%. Not great you might say.
But when we look at Team A’s defensive stats, we note that they have collectively held opponents to 15% shooting. Whoa, that’s pretty good. It’s 13 percentage points better than the average defense (remember, the average offense is at 28%).
When we add that gap (13 percentage points) to Team A’s offensive shooting percentage, we get 37%. Their offensive is a bit below average, but their defense is way above average. Collectively, they are better than the average team at making sure the ball goes (or doesn’t go) where they want it to go.
Let’s approach this from another direction. Let’s say that there are two teams: Team A has an ok offense, but an amazing goalie where Team B has a few sharpshooters on offense and an ok goalie. By adjusting offensive shooting percentage to account for goalie play, we can compare the two teams.
Or if you want to think about this relative to other statistical categories, it helps answer the question of: how on earth does Team B beat Team A when Team A wins more groundballs, dominates time of possession and turns the ball over less often?
Adjusted shooting percentage is sort of like special teams in football. It’s not a great analogy because obviously shooting and goalie play are much more important in lacrosse than kicking is in football. But if you think of everything that happens between 2 shots being taken in a lacrosse game as “normal play”, then it makes a bit more sense.
A football team may dominate a game, but be felled by poor special teams. In a lacrosse game, a team may dominate most of the important categories, but end up losing because the other goalie stood on his head.
What’s the plan
For now, we are going to keep track of our adjusted shooting percentages. It will be interesting to see how they correlate with overall wins as well as something like the adjusted efficiency numbers that analyticslacrosse.com puts out.
Longer term, this might even be a cool way to augment our live win probability page content. We shall see…