Back in February I presented an initial post on adding valued context to existing hockey possession metrics. The idea for enhancing stats of this nature was to provide additional (missing, but meaningful) context to the existing, one-dimensional advanced analytics data.
Since February I’ve been working on a series of analytics models that provide enhanced resolution and greater context for today’s standard possession metrics. The focus of the initial effort was to normalize all game metrics and make comparable throughout the season. This assisted in the identification of false peaks and valleys, helped identify quality of play, regardless of game results and assisted in identifying strengths and weaknesses in the Capitals as a team.
The purpose of this post is to provide a very brief refresher on the progress of modeling during the 2022-23 season and goals for the upcoming 2023-24 season.
INITIAL MODEL DEVELOPMENT
The first step in the model development was to factor in the strength of opposition for each game played. In short, an expected goals for percentage (xGF%) against the Anaheim Ducks is not comparable to an xGF% against the Boston Bruins. In other words, how well did the Capitals perform, considering the strength of opposition?
The first iteration of the model simply multiplied the Capitals expected goals for percentage for a game by the opposition’s pre-game winning percentage. In addition, it was determined that an average game score was an xGF% of 50.0% against a .500 team, which equates to constant value of 25. Thus, 25 is subtracted from the product of (xGF% x OppWin%) in order to derive a differential.
[xGF%(game) X OppWin%] – 25
Pretty straight forward, but now we can more accurately compare any game against any other game because we have normalized the standard game values.
The following graphic is from the initial post, which summarizes the first 53 games of the Capitals season.
You can clearly identify when and where the Capitals began to heat-up last season. By knowing this, each individual game can be reviewed to better understand why a turn for the good or bad may have occurred. More on that coming this season.
VERSION 2
Like all initial iterations of any analytics model, anomalies surface that require assessment, understanding, modification and fine tuning. Towards the end of the 2022-23 season we began to modify the initial model to provide additional resolution to the quantitative performance measures.
Enhanced Performance Index
There is no question that teams win and lose games they shouldn’t have won or lost. All of the stats, eye-tests and supporting data say the team outplayed the opposition, but because of all sorts of outside factors, including puck luck, penalties, injuries in the game, etc., the final results didn’t agree.
In the first release of the model, it was noticed that there were three or four games games where the performance score did not completely agree with the overall performance of the Capitals. As a result, it was determined that by adding the goals differential and expected goals scored differential, the final game performance score was more accurately represented when compared to the ground truth.
[[xGF%(game) x OppWin%] – 25] + (GF-GA) +(xGF -xGA)
The following screen shot from the second-generation model reflects those changes for all games since the return from the 2023- All-Star break, and provides new enhanced game scores. It also provides a color coding for each game to assist in identifying the trends of the team. [Click to enlarge].
INITIAL PLANS FOR THE 2023-24 SEASON
The overarching plan is to continue understanding the data and evolving the model throughout the 2023-24 season. Primary items of focus to start with include:
- The absence of key players from a lineup (injury or other reason) is critical context to any game data. Therefore, developing reliable multipliers for players and establishing an xGF% for a perfect (injury free) lineup and then subtracting the multipliers for each player absent from the lineup will add much-needed context to the model.
- Enhanced correlation on modeling adjustments for starting goaltenders for both the Capitals and the opposition.
- Teams get hot and cold during the season. Potential adjustments for performance over the previous 5 and 10 games will be considered and applied to opponents pre-game factors.
- Line and defensive pair performance factors – correlation of line performances and game results will provide insight as to the characteristics of what worked when the Capitals win, and what didn’t when they lose. More to come on this point.
That’s the initial plan, but there will surely be more added to the model in the coming weeks. Stay tuned.
By Jon Sorensen
Impressive your team’s statistics and Operations Analysis acumen!
A great idea to “normalize” hockey stats, with “context” to ordinary stats based on the list of contingencies like absence of key players and teammate factors, i.e. d-pairs and forward-threesomes. Good Show! Oil up your SPSS packages!
I’m still dying to see your analytical methods applied to coaching and coaching staffs. Peter Lavi was so uninspired and uninspiring in the second half of last season … that you could cut it with a skate blade. Leadership is a HUGE factor. The Terry Murray Caps were radically different from the Bryan Murray Caps. Statistically they were almost even, but the players essentially did mutiny against Terry M.
They did the same after five years of Ron Wilson — although I thought that Wilson was one of the Caps all-time best coaches, and new GM George “the Hammer” McPhee gleefully sacked Wilson in favor of the utterly dubious Butch Cassidy. Still amazes me that Cassidy eventually figured out how to be a reasonably effective coach. Took him 20 years.
Yes, leadership matters. Mikey Babcock, anyone?