This is the second part of a two-part series investigating whether faceoffs help you win in the NLL and if so, how much. In part I, we discovered that faceoffs do matter to some extent, in that teams that win more than half the faceoffs in a game tend to win that game a little more than half the time. Now we’re onto the “how much” question, and here’s where the math gets a little heavier.
To help us with this question I have called on Cooper Perkins, the Seals play-by-play announcer, stats geek, and the creator of LaxMetrics.com. Cooper is great at breaking down stats in ways I wouldn’t have thought of so I was hoping he could add some interesting insight, and he didn’t disappoint. The rest of this article and the data and graphs were all provided by Cooper. Thanks to him for joining me on this faceoff adventure.
This isn’t the first occasion in which Graeme has written about the relationship between winning faceoffs and winning games in the NLL. He’s done a tremendous job, and I’m excited to offer some further context to his findings on behalf of the LaxMetrics.com blog. Consider this a lacrosse media crossover!
As Graeme’s Part I of this exploration indicated, from 2006 to this point in 2022, teams that won more draws than they lost enjoy a winning percentage in those games over .500. As the percentage of faceoff wins climbs, so too does each team’s winning percentage, at least to a point. There seems to be some degree of diminishing marginal returns for faceoff success.
But when we consider league-wide data spread across more than 15 years, we lose a little bit of the nuance of each individual club in the process. Personnel matters in more places than just at the faceoff dot. In the same way that some teams are built from the backend forward and others are built from the front end back, there are a handful of organizations that have prioritized faceoff success in their roster construction.
In the spirit of LaxMetrics.com statistical analysis, we’ve come up with a means of calculating the impact of faceoffs in terms of “wins”. The methodology is fairly simple: goal differential is the key to predicting a team’s record, and the more possessions a team has, the better their chances of scoring more goals. Faceoffs impact the number of possessions a team has, thus playing a role in how many goals it scores.
Our first step involves figuring out how many possessions each team averages per game. We are defining “Rough Possessions” as the sum of a team’s total shots (including shots that miss the goal) and its turnovers. Obviously, this is an imperfect representation of the total number of possessions in a game. There are other outcomes to a possession than just a shot or a turnover, but the vast majority will end in one or the other. Perhaps if the NLL adds play-by-play tracking to its stat collection in the future, we could have a more precise number of possessions per game at our disposal.
Secondly, we need to calculate each team’s “Scoring Rate”, which is simply its goals scored per game average divided by the “Rough Possessions” number described above. This percentage gives us a solid idea of how often a team converts its possession into goals, which is crucial in establishing the relationship between faceoff wins and goal differential.
The next step is to find the average number of faceoffs each team wins per game. If we then compare each team’s faceoff performance to the league average, we can get an idea of how much better or worse at the dot a team is than its peers. By subtracting the league average faceoff wins per game from a team’s average, we can see how many possessions the team adds to its offense via faceoffs.
When we take that number of faceoff wins per game above the league average and multiply it by the team’s “Scoring Rate”, we can see how many goals a team adds to its scoring average by winning faceoffs. We can then add these values to each team’s goal total and plug it into our Pythagorean Win-Loss formula [Goals Scored^5.65/(Goals Scored^5.65 + Goals Allowed^5.65)] . This tells us how many “wins” a team can add or subtract from its 18-game record projection as a result of their faceoff success.
The results shouldn’t be surprising. What we find is that faceoffs matter more to some teams than to others. Specifically, faceoffs matter the most to teams that are either really good or really bad at them. They matter far less to teams that are roughly average at the dot.
Predictably, Albany and Halifax this year have the highest “Wins Added” numbers, which reflects the tremendous work of Joe Nardella and Jake Withers respectively. Additionally, Colorado, Georgia, Saskatchewan, and Rochester are all particularly bad at winning faceoffs, which is reflected in their negative “Wins Added” values. Functionally, these teams are losing nearly a full win from their 18-game record projection. Among the league’s worst faceoff teams, the only outlier is Buffalo, who has the league’s worst faceoff percentage, but ranks 10th in “Wins Added” (-0.466). We can attribute this de-coupling to Buffalo’s league-best scoring rate (15%), as the Bandits convert possessions into goals at a rate that minimizes the impact of lost possessions on faceoffs. The opposite is true of Rochester. The Knighthawks struggle at converting possessions into goals, meaning that the possessions lost in faceoffs are a significant problem for them.
As we can see, translating faceoffs into “Wins Added” puts into perspective how valuable the possessions gained by winning faceoffs can be. Aggregated over a full season, a team like Albany, who is in the playoff hunt, can expect enough added possessions to add a full win to their record. In their case, the gravity of an extra faceoff-fueled victory could be the difference between the Firewolves making or missing the playoffs. The inverse is true for Georgia. Could their season-long struggles at the dot cost them enough possessions to lose a full win off their projections? Both outcomes are possible.
In the end, the LaxMetrics conclusion to the question of “do faceoffs matter?” is simple: faceoffs matter a lot if you’re either really good or really bad at them. We saw the evidence of this phenomenon at work earlier this season when Albany upset San Diego 13-12, winning 25-of-30 faceoffs in the process. Mathematically, the extra possessions accrued via faceoff almost surely lifted Albany to a win.
But matchups of the best and the worst are rare. Most faceoff battles are between fairly evenly matched, mediocre opponents. In those cases where faceoffs wins aren’t wildly skewed in one direction, their importance is limited to situational value throughout the game. Faceoffs might matter a lot to Albany and Colorado, but they don’t matter much in the big picture to at least six of the other 12 teams in the NLL.