I promised people some graphs on Sharp and what "sheltered" means from an advanced stats perspective. After the jump, I will present them.
"Zone starts" measure the ratio of how often a player starts a shift in the offensive zone compared to the defensive one. If a coach was rolling lines, all the players on the team would eventually have very similar zone start ratios. And that number would reflect the overall quality of the team. Good ones having generally more offensive zone starts compared to weaker ones.
What we know is that for zone starts, individual players do not "regress to the team mean". Instead coaches give some players more than the average offensive zone starts and others more defensive ones. There are generally two reasons for this. First, as often as is possible, you want to put your best offensive players in the offensive zone and best defensive players in the defensive zone. The other reason is that a coach wants to protect a defensively weak player. So coaches will put those players on the ice as far away from their own goal as possible.
Now some people new to this concept think players with the most offensive zone starts are the weakest defenders. Basically, if they were more of a two way player, their zone ratio would be more equal. This is usually NOT the case. In fact most offensive zone starts generally go to the best offensive players. You are trying to score, too. So there are only so many opportunities to "shelter" players.
So how do you determine which players are in offensive situations because they are good offensively and which players are more protected because they have less of a two way game? It is about who you play against. Coaches use the offensive zone starts against the other team’s least offensive players to "hide" their weaker defensive players. They simply don’t throw the defensively weak players out there against the other team’s top lines. Even in offensive draw situations. So you are looking at the ratio of "offensive zone starts" to the "level of competition played" to determine who are the sheltered players.
note: I think you have to click on these images in another window to see them a little better.
So what I did was graph "zone starts" to "Corsi Relative Quality of Competition." The first of these graphs is for the Chicago Blackhawks 2011/12 season. I’m just looking at the top 13 or so Hawks forwards by games played.
The best defensive forwards tend to have more defensive zone starts and play against the other team’s better lines. So as the player moves up this graph they play against tougher opponents. As players move to the left they get more defensive zone starts. This means players in the upper left hand corner are your "defensive stoppers." And as you can see in the graph, the Hawks players in the left corner are exactly who you would think they would be. Dave Bolland is there and the rotation of three wingers that spent the most time with him; Bickell, Shaw and Frolik.
Players at the Bottom of the graph are pretty much your 4th line guys. I’m going to ignore Sharp for a second, so your bottom line players are Mayers, Hayes and Brunette. Brunette played on the top lines more often than the other two and that is why he is above them in the graph.
So now we are going to look at the relationship between the other two centers, well three actually, since Kane and Kruger split time at 2c this year. If we were to compare Kruger and Toews by only looking at zone starts, we would think that Kruger is the better defender. What we know is that it’s really about the ratio of zone starts and who you play against. Toews plays against the other team’s top talent more often than Kruger or Kane. So looking at this graph you would conclude that Toews gets more offensive zone starts for his offense rather than his lack of defense.
Comparing Kane to Kruger, though, shows the difference in those two players. Kane got significantly more offensive zone starts than Kruger. Kruger played more on the bottom lines so his QoC numbers are slightly lower but he got way more defensive zone starts. So Q pretty much used the second line differently when Kruger centered it than when Kane did.
Now let’s look at the wingers Stalberg, Hossa and Sharp. Stalberg tended to play against the higher level of competition, Hossa tended to get less offensive chances and play a slightly more two way game. And that leaves Sharp.
Patrick Sharp is on the graph in what are the cushiest shifts for a top 9 forward on the team. He is getting the most offensive zone starts against the easiest competition available. This is either your best offensive player; or one of your weaker defensive ones. And if he was the best offensive guy he would tend to be more to the right than some of the other wingers.
Patrick Sharp played 43% of his time with Kane, 32% of his time with Kruger and 27% of his time with Toews. Kane had a slight advantage of playing 32% of his time with Toews. But that really isn’t enough to offset this observation. How is Sharp’s QoC numbers so far below all three centers? He didn’t play on the 4th line. The only way, is to conclude that each center tended to get easier shifts when playing with Sharp and tougher competition when playing without Sharp.
And just an observation on Hossa, his 2011/12 season saw a drop in his QoC compared to previous seasons. You will see that in next graph. So Hossa isn’t playing as tough of opponents this season as he has in the past. That is something to watch for as Hossa gets older.
Anyway, looking at the graph, Hayes is an example of a sheltered player playing on the 4th line. Where Sharp is on the graph is the position where top 9 players are thought to be "sheltered." And in this case, Sharp is so sheltered that he even ended up below a frequent 4th liner in QoC.
Now looking at two seasons ago we see very similar things. Bolland and his wingers are in the left corner. Dowell as the 4th liner is at the bottom. Kopecky, Stalberg and Brouwer who spent time jumping between the top six and bottom six are in the middle; Kane in the more offensive situations, and Hossa is still for this year, taking on a more defensive role against a higher level of competition. And that brings us back to Sharp.
Sharp is again in the most one dimensional offensive only position on the graph. He got by far the easiest top 9 shifts for forwards. And forwards like Stalberg and Brouwer ended up with higher QoC numbers even though they spent a bunch of time on the 4th line. And the QoC difference between Sharp and Toews at center is substantial.
If Sharp was played like a two way player, his QoC number would be much closer to the Toews range. A coach has to work at getting that kind of separation. And he has to be doing it on purpose. A two way center who played more often with Hossa than Kane should simply have a higher QoC; at least closer to Toews. And, in fact, if Sharp was primarily playing against the other team’s second line and Toews was playing against the checking line, then Sharp’s position would be above Toews on the graph.
So looking at the Cup year is interesting. Q has yet to fall in love with Bolland as a top checking line player. Q still has him playing against the other team’s top lines but much more in a two way role on the second line. Madden spent more of this season as the checking line center. The roles really weren’t decided until the end of the year and in the playoffs. However, Versteeg, Ladd and Brouwer are in more defensive roles. Hossa, Sharp and Kane are in the more offensive ones.
And by the way, this is one monster team. Even your defensive players are getting 55% offensive zone starts; which of course, is simply dominant.
And one other comment, zone starts for centers are sometimes misleading. Looking at Madden, he has many more defensive zone starts than his wingers. The reason for that is Q was using Madden as a second center in defensive zone draws. So it can be difficult comparing centers from different teams. Some centers will appear more offensive and play higher levels of competition when in fact they just aren’t being used as much in that second center role. So you have to be careful with that.
Anyway, now back to Sharp. Outside of your 4th line guys, Sharp has gotten the easiest combination of zone starts and quality of competition compared to any of the top 9 guys. And this is part of a four year trend. A top two way forward doesn’t end up here four years in a row.
Now let’s skip a year and look at the season Patrick Sharp was voted 4th overall for the Selke trophy. Notice something different? Patrick Sharp is playing with DAVE BOLLAND. He is over there on the left in a checking role on the Bolland line. The two players most "protected" this year are Martin Havlat and Patrick Kane. Havlat is SO protected he is below the 4th liners in QoC. And Toews is in the primarily top offensive role.
So what happened?
How did Sharp go from a perceived two way player in 2007/08 capable of being a Selke nominee, to spending the next four years playing as a primarily one dimensional offensive player? Well there a number of possibilities:
1. Q is a complete moron, playing a premier two way player in a position normally reserved for a one dimensional offensive player. And doing it 4 years in a row.
2. Sharp was paired with Kane and Toews for the 2008/09 season. And he was not capable of repeating his sterling performance of the previous year. So Sharp modified his two way game to get more scoring chances and become a more one dimensional offensive player. I mean scoring goals gets you all-star games and big contracts, right?
3. Or maybe, he never really was all that much of a two way player and Dave Bolland carried and covered for him? And that let Sharp give an offensive punch to the Hawks checking line. Basically, was Sharp "Havlat" before Havlat was moved onto Bolland’s line the following year? And did Q get the idea of moving Havlat to Bolland’s line from watching tapes of Sharp playing there the previous year?
The Last four years
In 2008/09, The Hawks moved Sharp from the checking line and he played him with Toews and Kane. And they played primarily in as cushy an offensive role as possible. Since that time Kane and Toews have moved up to higher levels of competition; Toews much higher. Sharp has for the most part stayed at the same level. And obviously this shouldn’t be the case if Sharp was such an exceptional two way player.
Compare Kesler to Sedin on the Van graph. They are much closer on the QoC axis and in fact Kesler on the second line is above Sedin on the first. That is where a two way second line center typically appears on the graph. If you want to understand why Q doesn’t play Sharp at center, just compare the VAN graph of Kesler and the CHI 10/11 graph of Sharp.
So for as long as Q has been the coach of the Chicago Blackhawks, Q has played Sharp as primarily an offensive player. And Q has NOT given Sharp much of a two way role. And that has been true for four straight years now, which is kind of important. It is not coincidence when it is four years running. Q plays Sharp in that one dimensional offensive role on purpose. It is NOT an accident. So either Q is a complete moron or Sharp isn’t (at least anymore) the two-way player some people think he is.