Tag Archives: Scoring Chances

Edmonton Oilers, Fancy Stats, NHL, Walter Foddis

Are the Oilers measurably better?

Recently, Jonathan Willis of the Edmonton Journal’s Cult of Hockey, argued that there is no evidence that the Oilers have improved from previous seasons. Based on the shot metric comparisons he used, that was a logical conclusion. However, there are other ways to assess progress, which I indicated in an earlier post, especially early in the season. In particular, a Progress Index can be derived by comparing the shot metrics of a game, or series of games, to the previous season’s series against a specific team. For metrics, I compare Weighted Shots (WghtSh%; 1 point for goals, 0.2 points for shot attempts); Shot Attempts (SAT%; blocked, missed, and shots on goal), Scoring Chances (SC%; defined by war-on-ice), and High-Danger Scoring Chances (HSC%; i.e., shots from the slot area). (All data is collected from war-on-ice.)

Direct comparison to the previous season series accounts for quality of competition. For instance, against an elite possession team like Los Angeles, you would not expect the Oilers to improve from a dismal 45% shot attempt differential to a respectable 50% SAT (i.e., break-even). Rather, you would expect something more incremental, such as improving to a 47% or 48% SAT. In this prior post, I show how progress indices are computed for the Oilers first two games. I have done this analysis for every game to answer the question, “Are the Oilers measurably better?” The table below shows with coloured bars whether a shot metric improved (blue), worsened (red), or did not change significantly (no colour).

Reading the bottom two rows, we see that the average Progress Indices turn out positive! Although the Weighted Shot metric improved, I find it difficult to describe in straightforward language, but I can describe the other metrics. On average, the Oilers have increased their shot attempt differential (SAT%) by almost 6 per 60 minutes (+3%) compared to the 2014/15 season series against these teams. With regard to higher quality shots, the Oilers increased their Scoring Chance differential by just over 6 per hour (+6%). Finally, there is a slight improvement in the highest quality shot, High-Danger Scoring Chances, of 0.5 per hour.

Although we see improvement in overall shot metrics, what we don’t know is if the improvement is that more offense is being generated, or better defense is involved, or both. To tease apart offense and defense, we look at shot metrics for and against, respectively. An increase in shot metrics “for” means the Oilers are finding ways to generate more shots, especially quality shots, which will translate into more goals. A decrease in shot metrics “against” means the Oilers, as a team, are doing a better job in suppressing the team’s offense. So which is more responsible for the improvement: Offense or defense? My intuition was offense, but I was wrong!

Although the average progress indices for offense has improved a little (+1.1 SAT/60; +1.55 SC/60), most of the improvement in the differentials is coming from defense! In particular, compared to last season against these same teams, the Oilers have allowed close to 5 (4.78) fewer shot attempts, 4.62 fewer scoring chances, as well as 1.29 fewer high-danger scoring chances per hour. Are you surprised? I was. So it seems that the combination of new personnel and McLellan’s systems have made more of a difference defensively than offensively, although both have improved. This is something any Oilers fan wants to see. We all know that the Oilers are not a playoff team, and that are greatest weakness is our defensive corps, but given that our team defense has improved, that’s good news!

Special Teams

The above analysis is equal-strength (5v5) data, which is about 80% of the game. What about the other 20%; special teams? Early in the season, special teams are best measured using Shot Attempts For in assessing the power-play and Shot Attempts Against to measure the penalty kill. From 2012 to 2015, the Oilers’ power-play has ranked 27th as measured by goal differential and 24th as measured by Shot Attempts For (SAT_F = 89.4 per hour). Notably, though, under coach Todd Nelson for the latter part of the 2014/15 season, their PP goal differential was in the top 10.  This young season, the Oilers’s PP units are generating shot attempts at rate of 93.5 per hour, which ranks ranks 19th. In terms of quality scoring chances (high-danger zone), the power-play ranks 11th with 20.6 high-danger scoring chances per hour. Thus, compared to previous 3 seasons combined, this season’s PP looks to be generating more offense.

Curious to see whether the 1st unit (Nugent-Hopkins/Hall) or 2nd unit (Mcdavid/Yakupov) is performing better, I looked at the their respective shot atttempt generation per hour. The 1st unit is generating more offense, with the Oilers pumping out shot attempts at rate of just over 106 per hour with Hall & Nugent-Hopkins on the ice. With McDavid and Yakupov on the ice, the Oilers are generating about 90 shot attempts per hour. When comes to high quality shots, Hall, Nugent-Hopkins, and McDavid have similar metrics with with a high-danger scoring chance rate of about 25 per hour. With Yakupov on the ice, this metric drops substantially to 13 per hour.

Last season, the Oilers’ penalty kill–as measured by Shot Attempts Against–ranked 12th (SAT Against/60 = 95). This season, the Oilers allowed shot attempt rate is worse at 99, which ranks 21st. Our top penalty killers (by ice-time) last season were Gordon, Hendricks, Fayne, & Ference with a SAT Against of about 99 per hour. This season, the top 4 are Letestu, Lander, Klefbom, & Sekera with a combined SAT Against of 107 per hour.

Still too early to evaluate the goal-tenders because of too small a sample size, I’ll say a few tentative words. Unfortunately, to this point, the Oilers’ goal-tending tandem has not performed well, despite their strong starts. Talbot’s adjusted save% is ranked 24th (among goalies with a minimum of 7 games played) and Nilsson’s adjusted save% is ranked 42nd out of all 60 goalies. This means that despite the team’s improved defensive, the below-average goal-tending hasn’t allowed the Oilers to capitalize with fewer goals against.

Final Notes

I was surprised by the improved team defense and the poor goal tending performances. With Talbot, I was very hopeful that the Oilers’ goalie woes were behind them, but it seems this is still a question. Time will tell. Let’s hope the improved defense, which I attribute mostly to coaching, will continue as players internalize further the systems they’ve been taught.

Hope you found this informative. Please leave any comments or questions below. Thanks for reading.

Edmonton Oilers, Fancy Stats, NHL, Walter Foddis

Brief Analysis of Justin Schultz’s 2014/15 Season

6 Comments

When I analyzed the seasons of Mark Fayne & Andrej Sekera, my analyses were dense. I was still learning about the new data–data that went beyond the “first wave” of analytics, like shot-attempt differential (a.k.a. Corsi). I was still deciphering what “right” analytical pieces were needed for a comprehensive, yet readable player profile. Recently, an Oilers’ fan asked me if I could give an analytical overview of Schultz’s season. The overview had to be brief because I was posting to a Facebook thread. Much to my surprise, I think I managed to pull it off. I include this analysis below. I am curious to hear from readers if they find the analysis, as brief as it is, understandable and informative.

Brief Analysis of Justin Schultz’s 2014/15 Season

Take-Home Point: The Oilers were only slightly worse defensively with Schultz on the ice, but he appears to contribute enough on offense to compensate.

I begin with the broadest measure, shot attempts (SAT), and then refine the analysis to Scoring Chances (SC) and High-Danger (i.e., the slot area) Scoring Chances (as defined by war-on-ice). Unless otherwise stated, all player data originates from war-on-ice.

Shot Attempt Metrics

I computed the Oilers’ score-adjusted shot attempts generated (SAG) and suppressed (SAS) per 60 minutes, relative to the team’s average (RelTM). Combining SAG and SAS, I also computed a shot-attempt differential per 60 minutes (SAT60), Schultz lead the Oilers’ defense with a SAT60 RelTM of +7.2. Next was Oscar Klefbom with +4.5. To clarify, +7.2 means that with Schultz on the ice, the Oilers generated just over 7 more shot attempts per 60 minutes compared to the team’s average. Defensively, Schultz’s SAS60 RelTM was -2.2, which was third lowest (lower is better) behind Klefbom (-3.7) and Jeff Petry (-2.2).

Scoring Chances & Net Goals

In terms of scoring chances, defensively Schultz was only slightly better than the team’s average with the Schultz-iced team allowing 0.14 fewer scoring chances (SC) per 60 minutes. Offensively, though, Schultz again lead Oilers’ defenders with the team generating 3.5 more scoring chances per 60 minutes. Marincin was second with 3.25 and Klefbom was third with 2.8.

In terms of the High-Danger Zone shots (the slot), the Oilers allowed 14% more shots than league average. With Schultz on the ice, the team allowed 2% more shots from the slot. So when it comes to allowing high quality shots, Schultz is a little worse than the team. Again, though, Schultz tends to compensate with more offense. Indeed, the team generated 16% more shots from the slot with Schultz on the ice. Because of the additional offense with Schultz on the ice, the team is actually expected to out-score the opposition. Including all scoring zones, when Schultz was playing, the team’s expected Net Goals was +0.32 per 60 minutes, which translates into +8 goals.

Passing Metrics

A common pattern in Schultz’s shot metrics is his apparent ability to generate more offense. Because of Ryan Stimson’s Passing Project, we have direct measures of a player’s passing and shot metrics. Here’s the glossary for the graph below.

  • CC% and CC/60 = Corsi Contribution (or Shot Attempt Contribution), which are individual shot attempts, primary passes leading to shot attempts, and secondary passes leading to shot attempts. These are given as a percentage (i.e., proportion of shot attempts a player is involved in when on the ice) and per sixty minutes. These metrics tell you how much offense goes through that player while on the ice and also how often they contribute.
  • Composite SAG and SG represent the total number of shot attempts and shots a player generated from both primary and secondary passes per sixty minutes. SAG/60 is solely for the player’s primary passing contributions.
  • Entry Assists represent the number of controlled entries a player assisted on. This is determined by the number of passes in transition (prior to entering the offensive zone) that was recorded for each player.
  • SC Contribution% and SCC/60 are identical CC% and CC/60, but represent only the scoring chances a player was involved in. Passing data for scoring chances was combined with War-on-Ice’s scoring chance (link to definition) data to arrive at a player’s total number of scoring chance contributions. SC SAG/60 represents the number of scoring chances set up from a player’s primary passes.

When it came to generating shot attempts and shots, Schultz was pretty close to the average defender (e.g., Corsi Contributions/60, Composite Shots-on-Goal/60). However, in terms of directly contributing to Scoring Chances through primary passes (SC SAG/60), Schultz reached an “elite” level. He ranked 8th overall among NHL defenders. Moreover, looking at scoring chance contributions overall (i.e., the last 3 metrics), Schultz ranked among the best in the league (over 90th percentile). Thus, there is not only more offense with Schultz, but higher quality offense.

Summary

Relative to the team, Schultz’s defense metrics suggest that he isn’t a terrible liability, and in fact, not that bad at all. This contrasts with his defensive gaffs in highlight reels. I also vaguely remember Schultz’s blunders, but I also know that human memory is unreliable and biased. (That’s why I rely on analytics to rescue my fallible memory with objective data.) Although Schultz is not that bad defensively, relative to the team, we all know the team’s defense was bad overall (e.g., 26th in score-adjusted Scoring Chances Against). In short, he’s average as a defender on a defensively weak team. Still, Schultz’s offensive ability compensated for his middling defense, which we clearly saw in the team’s expected Net Goals (+8) and increased Scoring Chances with Schultz on the ice.

Recommendations for Schultz to Succeed

What does Schultz need to succeed? First, he needs to be more consistent in executing his defensive responsibilities. At times, it seemed like he fell back into old habits. Easier said than done when under pressure, but I think he needs to work on the mental discipline to execute what he intuits to be the “right” play, which includes the discipline to implement what McLellan and his coaching staff will have taught him. This is more of a psychological battle than mistakenly being labelled as “lazy.” When a player is overwhelmed, they tend to feel their options are limited. But if they have enough practiced learning in different situations, this would give them confidence when a similar game-situation occurs.

Second, I would suggest avoid matching him up against tougher competition. On the road, though, this may be next to impossible.

Third is what I consider the must-fix-above-all-else element: Schultz needs to improve his ability to read plays. This is possibly linked to my first point about mental discipline. I recall seeing him look lost at times and wondered how the coaching staff was going to help him correct it. Then I would see this “lost in the wilderness” situation repeatedly, which lead me to wonder about Schultz’s “Hockey IQ.” Specifically, I questioned whether he was cognitively capable enough to read developing plays, decide where he should go, and what he should to to be most effective. Can McLellan and crew help him break through this apparent psychological barrier? (Disclaimer: I don’t know what media pundits and hockey people mean by “Hockey IQ.” If someone can define it for me in a formal way; that is, a way that it can be measured reliably, I would be a happy analytics camper.)

Fourth, Schultz cannot be hesitant to use his body. (His physical hesitation, accompanied by a compensatory stick-reach while bent over, has lead to the derogatory term, Jultzing.) I think being more physical is what Schultz was alluding to when he spoke of “playing with an edge.” By physicality, I don’t expect him to separate opponents from the puck with big hits, but I do expect and want him to make it more difficult for opponents to execute their shots and passes. Duncan Keith of Chicago had 16 hits last season. You’d think he get more hit by random chance alone. Yet, he is one of the best defensemen in the league. Victor Hedman of Tampa Bay also hits very little; less than 1 hit per game. Thus, effective defense is possible without splattering players along the boards.

Finally, being partnered with an effective defensive defender would be ideal. Klefbom is on his way to being a strong defender (as well as an offensive contributor), but we also have to remember he only has 60 NHL games under his belt. Despite being in the league for over 200 games, Schultz still needs a mentor! Pairing him with a veteran like Fayne in certain situations (e.g., softer competition), could help Schultz progress.

Thanks for reading and please let me know what you thought of this brief analysis. (At least, briefer than what I normally post.) Also, what do you believe needs to happen for Schultz to succeed?