Tags

, , , , , , , , , , , , ,

Prolific playmakers have the reputation of making their teammates better. This is a widely accepted claim, and one that I think is worth looking into. Essentially, the general idea is that certain players are able to generate high shooting percentages for their teammates, and more importantly sustain these rates. Let’s take a look at the on-ice shooting percentages of certain players when compared to the league average:

What this makes evident is that certain highly skilled players can create high on-ice shooting percentages, just as a good goal scorer can create good personal shooting percentages. This reinforces the common belief that play makers make their teammates better. In fact, play makers have a much more comprehensive overall impact. Where a good goal scorer raises his own shooting percents, a play maker raises all of his teammates; this is more valuable as this would lead to higher team-wide shooting rates.However, the large majority of players seem to have no ability to raise their on-ice shooting percentages, or at least not sustain those rates. There is also very little evidence if any at all that a player can decrease opponents shooting percentages, but that is a whole different study. At any rate, a normal players on-ice shooting percentage seems out of his control; It will fluctuate randomly, pivoting around the average, around 7.6%. For a player who has shown sustainable ability to increase their shooting percentage, it doesn’t mean that it doesn’t fluctuate randomly. It merely means that this fluctuation will now pivot around an average higher than 7.6 For example, Taylor Hall’s on-ice shooting percentage would seem to fluctuate around the percent of 9.4. The ability to raise shooting percentages of those around should be very highly valued.

Combined with possession stats such as corsi or fenwick, shooting percentages can reveal a lot about a players overall effectiveness. If we look at Taylor Hall — a player often criticised for his bad possession play, specifically so far in the 2013-2014 season — there is more to meets the eye than only his corsi or relative corsi %. Obviously the corsi percents are important, as they demonstrate ability to reduce shots against just as much as producing shots for, but from a purely offensive standpoint there is more to look at than only his differential. While Hall has been on the ice at 5on5 this season, his team has created 354 shots for, while they have given up 438 shots against — a -84 differential that is frankly abysmal. Regardless, Hall does have the advantage on his side of being one of few players that can sustain high shooting percentages. If we assume that his opponents are shooting at league average 7.6%, his opponents would have scored 33.28 goals against, which I’ll round down to 33. Meanwhile, while Hall is on the ice at 5 on 5, his team has generated 354 shots. If we assume that his teammates are scoring at his on-ice shooting percent average over the last 3 years — 9.4% — then his team will generate 33.276 goals for. This is essentially identical to his goals against if his opponents were scoring at league average shooting percentage.

I find this to be very interesting. Possession stats are so important; they are one of the most crucial stats in existence for analysing hockey. That being said, it’s also undeniable that the ability to generate and sustain high shooting percentages are incredibly important. Despite the Oilers having a team wide 43.8% fenwick percent and Hall’s personal possession numbers being awful, the on-ice shooting percentage he generates makes up for some of this lackluster play. Were he to be a good two way player and generate positive possession numbers, we would see a truly dominate force in the NHL, as we see with all-star players like Sidney Crosby.

Often, General Managers make decisions based on short-term on-ice shooting percentages. Dave Nonis, GM for the Toronto Maple Leafs, defended Leafs acquisition David Clarkson when he said “”We’re not penciling David Clarkson in for 30 goals, but anybody that has put up 30 at any time in his career has got a bit of a touch.” Here we can see some wishful thinking from Nonis. He claims, essentially, that because Clarkson had a 47 point season, it justifies expecting signing him to a 7-year 5.5 million dollar contract as he has a touch and is expected to do so again. Even if Nonis says that they don’t expect anything of Clarkson, it is obvious by his next comment that he expects Clarkson to return to the form he had in 2011-2012 and to a lesser extent 2012-2013. In Clarkson’s case, it is not his on-ice shooting percentage that has dropped, but instead his personal shooting percentage. In his only 30 goal season in 2011-2012, his shooting percentage was 13.2%. Over the previous three years, including playoffs, Clarkson’s shooting percentage is an average of 10.07%. The outlier for these seasons is his 30 goal campaign in 2011-2012, where his save percentage was an impressive 13.2%. The problem is that Leafs management seems to see this campaign as likely to be repeated, instead of the outlier in the career of an otherwise average third line goal scorer. The 5.5mil contract handed out to Clarkson is a testament of this belief.

General Managers simply must not look into outlying short term shooting or on ice-shooting %’s, as for the average player these are based more in statistical unlikelihood than they are in repeatable skill. Let’s take a look at some average players and how their shooting %’s, on ice or personal, deviate from season to season:

The league average here is hard to see, as it is surrounded by other lines, but it is essentially in the middle of the major clump. What we can tell from here is that average players, even good players, have little impact on their on-ice shooting percentages, at least not sustainably. Frazer Mclaren, for example, had an on-ice shooting percentage of 10.5% in 2012-2013, but that dipped to 0% in 2013-2014. For most players, on-ice shooting %’s have very little correlation from one year to the next. Certain players as I mentioned can sustain high shooting percentages, but to assume that an average player that has a high on-ice shooting % one year has the ability to sustain that rate is a flawed way to look at it. David Backes is having a very good year this year, with an on-ice shooting % that is hovering just below 10, but that does not mean that his GM should expect that moving forward. More likely, a GM should expect Backes to sustain closer to league average on-ice shooting %, as that is closer to his career average, and for his production moving forward to be lower than his current pace. If GMs judge players on their high water marks — which I would argue the Toronto Maple Leafs in particular do — they will find themselves overpaying players due to results in one year that they are not likely to repeat, at least not sustainably. Investing in players with good possession results is a more reliable strategy, as this is a much more consistent and sustainable ability. That being said, a mixture of players who have good career on-ice shooting percentages and good possession numbers would be a deadly combination — a combination that is rare, as players with high career on-ice shooting percentages are very scarce, especially those who raise it by significant margins.

                In conclusion, players who sustainably raise on-ice shooting percentages are incredibly valuable, and the ability to do so often inflates their possession rates more than the rate itself can express. However, an average player has very little ability to change this rate consistently — as we see with Frazer Maclaren, who has had a nearly 14% decrease in his on-ice shooting percent from last year. For this reason, although this skill is very valuable, GMs need to be careful as to which players they identify that actually have this ability. .