The Next Great Lacrosse Statistic
Or: Why I don’t actually hate stat nerds
There are many battles being fought in the lacrosse world right now. This is far from the most important of them and I hate it in the same way that one hates a sore calf that never heals right plagues you for the rest of your life. But, like serial killers and true crime podcasts, things that we hate can also interest us. To me, the next frontier in lacrosse (besides an ever-increasing and intensive level of recruiting coverage) is the application of advanced statistics to the sport.
Now, that is not to say that such things don’t already exist. There are plenty of advanced team and individual stats that are being applied right now on hundreds of coaches’ computers. Things like offensive and defensive efficiency numbers have been around for a while now, but I’ve always been less interested in those broader team-based statistics than I am in the individual ones. We all know the basics, right? Goals, assists, groundballs, face-off percentage, save percentage, shot percentage, caused turnovers - these are the numbers that everyone rattles off. They’re black and white, staring you in the face as valuations of a player’s direct contribution to the success of their team. Except for turnovers - turnovers don’t count as long as the player has some sick highlights! Oh, oh, or a huge social media following! Right, guys? Right? Guys?
The Premier Lacrosse League’s stat usage has been spearheaded by their own internal team and utilized most heavily by Joe Keegan in his articles and newsletters. He has been very gracious in sharing any and all specific advanced data that I have asked for and if you’re subscribed to this newsletter you should subscribe to his as well and then listen to the LacroCity podcast episode that is in the making. The shot charts for players are useful and interesting if you’re into tracking player positioning and shot selection. Conversely, traditional NCAA team stat usage is not great. It is beyond basic and offers zero additional insight. This is just another example of how pro lacrosse is superior.
My favorite stat to try and chart/create for lacrosse players is, depending on the context, inherently negative in nature. I am weirdly obsessed with the replacement player metric for professional lacrosse players. I have tried to break it down in numerous ways with the aid, and scorn, of Inside Lacrosse’s Patrick McEwen since the summer of 2020, and the nuclear winter of our lives, began. We have come to agree that the replacement level attackman (and offensive midfielder) is the 21st best-rated player in the league. Now, that’s WILDLY simplified (mostly for my tiny dude-bro brain to relate to Patrick’s gargantuan science brain) but if you think about it - it’s pretty simple to formulate. Seven teams, three [starting] attack/o-mid spots per team, and the worst-rated starter is replacement level.
So, how do we solve for player ranking? Well, we (and by we I mean Patrick because I just said “yeah, that sounds good” to everything) take every player and weigh their points per game, goals per game, assists per game, turnovers per game, and their true shot percentage against each other. Here is the result of that:
And here is the one for [offensive] midfielders because why not?
Does anything...jump out at you, dear reader?
One of the things I despise about covering pro lacrosse is the blind fandom that exists for players. Perhaps despise is too strong...let’s say blind loyalty for players at the next level just because they played at someone’s favorite college or are from someone’s hometown. It is nigh-impossible to write anything critical of any player without someone that knows them roaring down the lane of Twitter or Instagram to dunk on you with the ferocity of an anthropomorphized cartoon tiger. That’s partially a byproduct of lacrosse being so small and thus - sorry for this but not really - small-minded when it comes to player evaluation. Conversely, there is zero joy in writing that someone has had a bad game (other than the euphoria that comes from conjuring the perfect metaphor/simile/image connected to it). It’s not always necessary to point out when a player is horrible because then lacrosse turns into soccer *ominous foreshadowing* but it is a necessary part of the discourse. Still, the truth is relative when it comes to stats. I’ll give you an example.
Player A led his PLL team in points, was third in the league in assists, and a top ten leading scorer.
Player B shot 18%, was third in the league in turnovers, and had six goals in five games.
(It’s the same player.)
Now, if I told you that player was Rob Pannell (because it is) how would you propose that I begin to break down his season? Do I start with the good parts and gloss over the bad bits? Do I just write about the good things? Do I make jokes about Rob never wearing a shirt like I usually do? Ideally, all of that sounds very fun. For me.
Obviously, Rob Pannell is a phenomenal lacrosse player. However, his basic stats can be manipulated to fit different narratives. It’s much harder to swap narratives with advanced stats - that’s a big part of what makes them advanced. The context is already included. This is why I want to share with you one of my favorite potential crossover stat applications: expected assists.
What’s an expected assist? The cut and paste definition is that an expected assist from soccer/futbol (or “xA”) is:
Obviously, lacrosse doesn’t have the raw data - positionally or tracking wise - to measure the latter piece of that explanation.
Yet.
However, different types of advancement have occurred in other sports that give me hope for [speedy] progress. Because I guarantee you that Rob Pannell would have led the PLL in xA in 2020. Take for example fellow SubStacker Ryan O’Hanlon’s article on how two of the least likely of bedfellows made sweet statistical love and birthed a new model of player evaluation:
https://nograssintheclouds.substack.com/p/how-a-former-minor-league-baseball
First: Please follow Ryan’s SubStack, his stuff is awesome even if you don’t like soccer.
Second: The idea of using classifications of playstyle for pro lacrosse players is, I must admit, intoxicating. For better or worse I have always wanted to break down the ideal complement of a six-on-six unit and how certain players work well in certain formations and how hybridization still exists in the modern era of specialization and role adaptation, but...it’s still far enough away that I can scream and wave at it frantically whilst it’s back remains turned.
But then...then I read that piece on DAVIES and I think of the future of measurable metrics and it seems so tantalizingly close. The issue isn’t an application of the numbers; it’s a lack of raw data to extrapolate. Lacrosse, for all of its monies and sundries, suffers from a lack of investment in areas such as individual player performance. While the business of evaluations is picking up - see also the new Inside Lacrosse Evaluation program which I have consulted on and contribute to - the money is largely confined to the margins of the recruiting market right now. But imagine a day where advanced player evaluation is applied to the highest levels of the game past college and into the pros.
We are years away from graphics like this to explain and/or show a lacrosse player’s strengths and weaknesses like this:
BUT - we are not far from following the trail blazed by pioneers in search of a better way to talk, describe, and evaluate players of all sports. You can hate the slow progression of stats and love the evaluation of the player. I’ve never been against statistics; I’ve just always been looking for the perfect one. Even if - no, especially if - it proves my whole approach to be wrong.