Unconventional Wisdom - Part 3 Hockey Analysis Rating

We are back again for another week looking at the Pittsburgh Penguins through the eyes of some of the lesser known fancy stats. If you missed it, we previously looked at Part 1 Goals vs Threshold and Part 2 Point Shares. This week we are looking at one of the more controversial metrics, Hockey Analysis Rating created by David Johnson. Why is it controversial? I really have no idea, but its like Coke vs Pepsi with the true believers in the opposing camp vehemently opposed to it simply because it is different.

Like GvT and PS, Hockey Analysis Rating is broken down into Offense (HARO) and Defense (HARD) categories, with a third category that attempts to combine them into a single overall value (HART). It takes the concept used for Relative Corsi and expands upon it a hundredfold. Whereas the Relative stats are simply based on the difference between On-Ice and Off-Ice performance for that one specific player, Hockey Analysis Rating uses an algorithmic function that compares the player's performance both with and without every single player on his team and every single opponent he faces during the season, all in relation to TOI. It is probably the most complete metric available for trying to determine where the player stands in relation to his peers, and since it is already factored into the equation you don't need to look at it comparison to Quality of Competition and Quality of Teammate differences.

Hockey Analysis Rating was designed around a Goals based metric, but because of the popularity of possession based metrics he also included Shots, Fenwick, and Corsi based ratings. The actual calculations of the values are quite complex, comparing the player to his teammates and his opponents as mentioned above, but the final value created is based on an average player rating of 0. If you are +X then you are X% more effective than the average player. For example, a Goals HARO of +10 means there are 10% more goals scored with that player on the ice than there are without him. For Goals HARD a +10 would mean 10% fewer goals against. And of course the same applies to Shots/Fenwick/Corsi, and a negative value of HARO would mean fewer Goals/Shots/Fenwick/Corsi For while a negative value for HARD would mean more Goals/Shots/Fenwick/Corsi Against.

HARO

As we saw with both OGvT and OPS, the past 3 season of HARO show our Top 6 are right where we want them to be. Of course sample size plays a bigger role, as Bennett is higher up the list than one would have expected, but he only had 1 season worth of data whereas the others had 3. Crosby still remains in a league of his own, literally, nobody else in the NHL comes close to his offensive output. A giant leap down we come to Kunitz, and then another step down we have Dupuis, Bennett, and Malkin. Then further down but still quite high is Jokinen and Neal. Crosby had a huge peak in 2011-12, but that is likely due to sample size based on missing much of the year with concussion issues, but last year was still high enough that despite being lower than his peak could still be seen as being on the rise. Kunitz, strangely enough, had a down year in 2011-12 but had a huge bounce back last season. Dupuis was steady for two years and had a giant leap last season. Malkin was steady for two seasons but then had a really off year last season. Jokinen has been steadily declining for the past 3 seasons, and Neal had a huge leap in 2011-12 and has stayed steady if not slightly improving the past 2 seasons.

The remaining forwards all have negative HARO over the past 3 seasons. Kobasew, followed by D`Agostini and Sutter, and then a pretty huge leap down to Adams, Glass, Jeffrey, and Vitale. Kobasew and Adams have been steadily increasing for the past 3 years, and even went positive last season. D`Agostini, on the other hand, has been steadily declining. Sutter had a really bad year in 2011-12 and bounced back last season but still remained negative. Glass was on the rise for two years but had a huge drop off last season. Jeffrey had an awful 2011-12 campaign which skews his overall numbers, which have been consistent yet still negative the surrounding two seasons. Vitale was steadily awful for two seasons and then improved last year, although still quite negative.

On D we once more see Letang at the top of the list, but he is neck and neck with Bortuzzo, who likely benefited from an extremely small sample size boosting his totals. Quite a bit further down is Orpik, and a ways below him we have Martin. The remaining D are negative, with Scuderi at the head of the pack and Niskanen and Engelland a ways further down. Letang had a huge breakout year in 2011-12, and while he was a little bit of a drop off last season he was relatively consistent and still quite remarkably high. Bortuzzo, oddly enough, has remained consistently high, even seeing a slight increase, but it probably has a lot to do with his limited TOI skewing the data. Orpik was fairly consistent for two years and then had a huge leap last season. Martin had a huge gain in 2011-12 and last season wound up falling slightly, but still far above the negative value he had in 2010-11. Scuderi was consistently negative and even slightly decreasing for two seasons, but then had a huge breakout year last season to become highly positive. Niskanen had an awful 2010-11 season and improved vastly since then, steadily increasing the past 2 years. And Engelland of all people has been steadily improving to the point where he turned it around to be positive last season.

HARD

Defensively, our best forwards over the past 3 seasons were primarily our Bottom 6, which is to be expected because that is where you want your defense focused. Like HARO, Bennett seems to be taking advantage of small sample size to climb unnaturally high up the chart. He is quite handily at the top, all by himself, with a fairly steep drop to Vitale and Adams, and then Jeffrey, Glass, and D`Agostini. Sutter falls in even further behind, with Dupuis coming in at the end as the last forward with a positive HARD. Vitale had a huge season defensively in 2010-11 but has since fallen and has declined the past 2 years. Adams, on the other hand, was declining for two seasons but had a huge breakout last year. Jeffrey had a pretty awful 2011-12 campaign which skewed down the data of his decently high and consistent surrounding seasons. Glass and Dupuis apparently had a big drop off in 2011-12 but a huge bounce back season last year. D`Agostini had a big gain in 2011-12 but was on the decline last season.

The remaining forwards all have negative HARD. Jokinen leads the way with Neal not far behind, a bit further down is Kobasew, with Kunitz, Crosby, and Malkin bringing up the rear. Jokinen went from an abysmally negative 2010-11 season to a huge leap in 2011-12 but fell again last season, but still remained positive. Neal had a poor showing in 2011-12 but with consistently good years around it. Kobasew has been on the rise the past 3 seasons and as of last year was comfortably positive. Kunitz had a really poor 2011-12, but his surrounding seasons were mildly positive. Crosby and Malkin were on a decline into the massive negative for two years but had a huge bounce back last season to be massively positive.

On D the top HARD is Niskanen, by quite a wide margin oddly enough. Then we have Martin and Scuderi as the only other positive D. Orpik, Engelland, and Letang are lumped together slightly negative, while Bortuzzo and his small sample size brings up the extreme rear. Niskanen had a pretty poor 2011-12 season but was consistent and quite highly positive in the surrounding years. Martin, Orpik, and Letang also had a poor 2011-12, but a huge bounce back to be massive positive last season. Scuderi had a huge peak in 2011-12 and fell into the far negatives last season. Engelland has a great 2010-11 followed by a dismal 2011-12, but bounced back this past year, although he still remains negative. And Bortuzzo's numbers are skewed by limited sample size and especially an incredibly low 2011-12 season, but last year he had bounced back into the positive range.

HART

Looking at the combined data, we see the same trends we saw with HARO, our Top 6 at the top of the list and the rest trailing behind in some manner. Crosby leads the way as the most dynamic two-way player on the team, he is followed by Bennett and his sample size inflated numbers, then Kunitz, Dupuis, Jokinen, Neal, and Malkin. Crosby has been steadily improving for the past 3 seasons, while Kunitz and Dupuis bounced back from a poor 2011-12 season to have a breakout last year. Jokinen had an excellent 2011-12 season sandwiched between two consistently mediocre seasons, while Neal and Malkin followed the same progression as Kunitz and Dupuis but with much smaller numbers.

The Bottom 6 includes 3 with positive numbers, D`Agostini, Adams, and Sutter, with Kobasew, Glass, Vitale, and Jeffrey rounding out the bottom. D`Agostini had an excellent 2011-12 season but has since declined, but still remained positive. Adams and Kobasew have been steadily improving the past 3 seasons, climbing out of the negative hole last year. Sutter had two steady seasons and saw his number jump last year. Glass had a poor 2011-12 season but actually climbed up into the positive last year. Vitale also had a poor 2011-12 season and also improved, but remained negative. And Jeffrey had two decently positive years sandwiched around one extremely poor 2011-12 season.

On D the best two-way D is Letang, with a big leap down to Orpik and Bortuzzo, followed by Martin and Niskanen, and then barely breaking even with Scuderi. Engelland is the lone negative and actually the lowest on the entire team. Letang had a little bit of a decline for 2 seasons but had a huge leap last year. Orpik had a poor 2011-12 but a huge bounce back last season. Bortuzzo's data is skewed by sample size, but he does show signs of improving. Martin and Niskanen had two consistently poor season that were slightly on the decline, but a huge bounce back season last year. Scuderi had an decent 2011-12, climbing out of the hole the previous season, but then fell again last year, although he remained positive. And Engelland had a huge drop off in 2011-12 and then bounced back to just barely break into the positives last year.

Goaltenders do not get a HARO, because despite the fact that there are rare instances of goalie goals and some plays where a goalie walks away with an assist, by and large they have no impact on the offensive output of their team. They do, however, get a HARD score based on their contributions to preventing goals. Fleury leads the way with a slight positive result, while Vokoun is slightly in the negatives. Fleury had a poor 2011-12 season, but he bounced back with quite decent numbers last year. Vokoun, on the other hand, has been steadily improving the past 3 seasons, to the point where last year he actually crawled out of the hole.