The first of Tomas Plekanec’s 15 seasons in the NHL came in 2005-06, which coincided with the league’s introduction of a shootout. The shootout was not kind to Plekanec. Among all NHL players with at least 20 attempts, Plekanec’s career shootout percentage of 13% is the lowest league history. A deeper look into Plekanec’s shootout usage reveals an interesting pattern: Seasons # Years Plekanec Shootout Attempts 2005-06 to 2014-15 10 23 2015-16 to 2018-19 5 0 Each of Plekanec’s shootout attempts occured in his first 10 seasons, and his success rate was so bad in that timeframe that h ..read more

Most often, running backs in the National Football League are compared by taking summary statistics from each play. Traditional metrics include yards-per-carry, expected points added per play (EPA), and win probability added per play (WPA). While comparing players based on these their averages is the straightforward, means can be sensitive to outlying observations and distributions with heavy tails. Unfortunately, yards, EPA, and WPA are all strongly skewed, which means that taking averages may not be an optimal approach. As part of the NFL’s Big Data Bowl, participants dove deeper into ball c ..read more

Air yards plots are popular and informative ways for assessing the accuracy of quarterbacks at different depths of the field. As examples, Josh and Ben have both done excellent work in highlighting how using where a quarterback was throwing to tells us more of a story than simply looking at completion percentage alone. Here’s one example of an air yards plot, using a pair of players from the 2019 season. The blue player tends to be slightly worse than the red player at short throws, better than than the red player at distances around 20 yards, and similar for longer distances. Both players ar ..read more

As part of the NFL’s technology conference held last week, I presented to club staffers various approaches on how R could be used to explore NFL data. Here’s a slightly abridged version. This specific audience featured both novice and expert R users. Given that a semester’s long course would be needed to truly cover the breadth of how R could help explore football, note that this particular document is more an overview than anything else. Finally, most of the code is not meant to be run from an outsiders’ perspective, and instead uses a few internal databases. I’m sharing publicly in case it s ..read more

How do you estimate outcomes under a set of rules that don’t (yet) exist? In this post, I’ll walk through how the NFL league office used resampling to estimate metrics related to an offseason proposal related to overtime. Background Back in March, the NFL’s competition committee and owners debated the merits of a rule change, proposed by Kansas City, that would allow both teams the opportunity to possess the ball at least one time in overtime. This would create a scenario unlike any other in football. If the game starts with an opening drive touchdown, the receiving team would: Need a touchdo ..read more

It’s NHL playoff season, which means that, given the relative league parity, every call and no-call comes under a microscope. In the third period of close games, for example, a penalty call alone can decrease a team’s chance of winning by 10 percentage points or more. One characteristic of NHL games that a few prominent members of the media have picked up on is that penalty calls tend to “even-up” within a game – that is, it is unlikely that either team ends the game with a substantially larger number of penalties than its opponent. Kevin and I, as well as a few others (ex 1, ex 2), have cover ..read more

Brian, Alex, and a host of others have done nice work recently when it comes to creating better scatter plots using images in ggplot in R. In this post, I’ll show how easy it is to use the ggimage package and do the same. What data to use? Ron, Sam, Max, and the folks at nflscrapR have built a pretty amazing tool to analyze football play-by-play data. They’ve nicely stored csv’s that summarize several play-level and team-level characteristics relating to each play. In my example, I’ll use this data to contrast team run and pass performance, as measured through expected points added. For more o ..read more

Taking a step back Most of sports analytics involve decisions made on the margins. We debate fourth down decisions that, across a season, add up to roughly a a third of an expected win. We compare the merit of Cy Young candidates, when the relative difference of each candidate’s WAR equates to nothing more than rounding errors. We tell NHL coaches to pull their goalie a few minutes earlier, knowing that, in more seasons than not, not an extra point will be gained by going so. Thus, perhaps it’s worth taking a step back to analyze what can be a substantially more dominant but often unquantified ..read more

One way in which professional sports are relatively fair is that, in each season, teams are almost always given an identical number of home games. This seems like an obvious way to run a sports organization, until you remember that postseason berths in NCAA hoops and football often hinge on incredibly unbalanced schedules. Playing at home is a benefit, and while the reasons for the overall advantage are somewhat up for debate, there are obvious and unique advantages that make playing at home different each sport. In our cross-sport paper, Ben designed a pretty amazing chart that highlights, on ..read more

In the 2016-17 season, the Washington Capitals dominated the NHL like few teams in recent history, winning the Presidents’ Trophy with 118 points. The Caps entered a 2nd-round playoff series with Pittsburgh as decent-sized favorites (58 percent), and sure enough, Washington outplayed its rivals. In each of seven consecutive games, the Capitals outshot the Penguins, finishing with 70 total more shots on goal. Unfortunately for Washington, not enough shots turned into goals, and because hockey games are decided by goals, it was the Penguins that moved onto the Eastern Conference Finals. The Caps ..read more