Shift Impact Percentage (part 2 of 2)
Posted 5:28 PM ET | Comments 29
In my last blog i wrote about how i feel the NHL needs newer more innovative statistics and introduced my first new stat of the year the Shift Impact Percentage or SIP%. I'm not going to recap what the statistic is composed of, for that see my previous blog. Instead i'm going to delve right into the numbers and start applying the stat so we can start to gauge what is a good SIP% and what is not. So let's get to it!
Below are several key members of the Sabres and their SIP% from last season. In addition i calculated the SIP% for six other players around the league as comparables.
Thomas Vanek .187%
Derek Roy .204%
Tyler Myers .145%
Paul Gaustad .180%
Cody McCormick .235%
Ville Leino (with PHI) .138%
Christian Ehrhoff (with VAN) .170%
Andrej Sekera .120%
Sidney Crosby .253%
Corey Perry .204%
Ryan Kessler .254%
Nick Lidstrom .165%
Shea Weber .268%
Chris Pronger .202%
Trends:
First off after running these numbers i'm even more impressed with Shea Weber. His stats were off the charts in every category. I found it interesting that Derek Roy and Corey Perry had the same SIP% and that Kessler edged out Crosby. While computing these percentages i realized both Ehrhoff and Myers led their team in total shifts, something that should provide the Sabres with a lot of stability on the back end now that they man the same blue-line.
You might be interested at the fact that Cody McCormick's shifts were far more effective than Thomas Vanek's. This points out both the good and the bad with this statistic. McCormick's numbers were great across the board, and given the ice time that he sees you can understand why Darcy and Co. were so quick to retain him this off-season. However the negative of this statistic is that it doesn't give points for goal productivity. It does account for assists because it was the only way to quantify passing on a shift, however it does not include goals because they would be redundant numbers with the shots attempted. The omission of goals however evens the playing field between a guy like McCormick and a guy like Vanek. In similar amounts of shots one will have significantly more goals than another, but according to SIP% the impact was the same. Maybe i should tweak the stat to include goals...though i kind of like how well this profiles players for their all around ability. Hence why Ryan Kessler has such high numbers.
Just like .300% is considered "very good" as a batting average in baseball, i would say .200% is that threshold with SIP. Anything above that is pretty darn good, and numbers like Webers are MVP worthy. As the season goes on i hope to come up with a few more interesting stats for you and who knows....maybe one of them will stick.
*FOLLOW ME ON TWITTER @SabreHess*
Below are several key members of the Sabres and their SIP% from last season. In addition i calculated the SIP% for six other players around the league as comparables.
Thomas Vanek .187%
Derek Roy .204%
Tyler Myers .145%
Paul Gaustad .180%
Cody McCormick .235%
Ville Leino (with PHI) .138%
Christian Ehrhoff (with VAN) .170%
Andrej Sekera .120%
Sidney Crosby .253%
Corey Perry .204%
Ryan Kessler .254%
Nick Lidstrom .165%
Shea Weber .268%
Chris Pronger .202%
Trends:
First off after running these numbers i'm even more impressed with Shea Weber. His stats were off the charts in every category. I found it interesting that Derek Roy and Corey Perry had the same SIP% and that Kessler edged out Crosby. While computing these percentages i realized both Ehrhoff and Myers led their team in total shifts, something that should provide the Sabres with a lot of stability on the back end now that they man the same blue-line.
You might be interested at the fact that Cody McCormick's shifts were far more effective than Thomas Vanek's. This points out both the good and the bad with this statistic. McCormick's numbers were great across the board, and given the ice time that he sees you can understand why Darcy and Co. were so quick to retain him this off-season. However the negative of this statistic is that it doesn't give points for goal productivity. It does account for assists because it was the only way to quantify passing on a shift, however it does not include goals because they would be redundant numbers with the shots attempted. The omission of goals however evens the playing field between a guy like McCormick and a guy like Vanek. In similar amounts of shots one will have significantly more goals than another, but according to SIP% the impact was the same. Maybe i should tweak the stat to include goals...though i kind of like how well this profiles players for their all around ability. Hence why Ryan Kessler has such high numbers.
Just like .300% is considered "very good" as a batting average in baseball, i would say .200% is that threshold with SIP. Anything above that is pretty darn good, and numbers like Webers are MVP worthy. As the season goes on i hope to come up with a few more interesting stats for you and who knows....maybe one of them will stick.
*FOLLOW ME ON TWITTER @SabreHess*
Leave a Comment
You must be logged in to leave a comment.Blog Archive
01 Sabres add Leino, re-up McCormick....Ehrhoff Thoughts
05 Why Leino is a boom or bust move
08 FA Forwards for 3.5 to 4.5 million, what do you get?
22 Enroth and the Value of Eastern Conf. Backups
26 Northeast Division Defenseman
28 My 5 Riskiest Moves of this Off-Season...So Far
05 Why Leino is a boom or bust move
08 FA Forwards for 3.5 to 4.5 million, what do you get?
22 Enroth and the Value of Eastern Conf. Backups
26 Northeast Division Defenseman
28 My 5 Riskiest Moves of this Off-Season...So Far
You can double check if you want, but I calculated Ovechkin's at 0.392!!!!!!!! That's insane!!!!!
"it does not include goals because they would be redundant numbers with the shots attempted" That is a terrible reason for not including it... A player who takes 1000 shots and gets 2 goals is therefor more productive than someone who takes 50 shots and get 40 goals? Maybe taking shot percentage would be good to include
Of course he's just as impactful because he created 1000 rebounds for his teammates. He may only have 2 goals, but he's got atleast 100 assists with that many shots.
Oh, and mfreedman, Aulie isn't Chara 2.0, Myers is. I'm not even a Sabres fan, but they're eerily similar. Size, two-way d-men, big shot, snarly, physical...
I dislike the stat. Let's call it what it is: tracked events per shift. A player like Lidstrom suffers because he does not hit or block many shots. That's not a flaw in his game; it's indicative of the fact that his defensive style (positioning, stick lifts, breaking up passes, etc.) relies on interactions not tracked by statistics.
Furthermore, a stat should give us some insight into a player's game, or at least what typically was happening when he was on the ice. But if I said "Player X has a .233 SIP%," and you had no previous knowledge of that player's game, what could you learn from that number? Nothing. ...
Does he hit everything that moves and block a lot of shots, but doesn't provide much offense? Is he a sniper and pocket-picker who doesn't throw the body? You'd have no clue.
Other problems: It regards hits, shots, assists, giveaways and takeaways as equal "impact" events. It cannot account for plays which are not tracked. It regards a blocked shot as a "positive" (when in fact you can argue blocked shots are a negative, a result of not having possession). It does not consider the converse (how many times you GET hit, how many times you get shot ON, etc.)
I like the way you think, and I like advanced statistics. I hope you come up with some more. I just think this one is a little too "hodgepodge" to be meaningful. Rather than provide insight into a specific portion of a player's game, or realistically quantify a player's overall "value," it really just counts how much trackable "stuff" they generate.
Sorry man.Good try at making a new advanced stat, but others are wayyyyy ahead of you.
I admitted in the blog the stat has its flaws, but doesn't plus/minus? No stat individually tells the whole picture. Sven22 i think you hit the nail on the head, forget about goals, there is no way to track things like stick lifts and good positioning. So it penalizes a guy like Lidstrom.
First, SabreHess I would like to give you kudos on an obvious hard working blog and I enjoyed reading it, it is organized, has a mathematical formula for consistency and I like your articles.That said I cant get behind this particular stat. Corey Perry and Derek Roy score the exact same and well, I think that is all that needs to be said about that.That doesnt mean I dont like that you have tried, I look forward to your future entries.
@ETP95 - Firstly, every shot doesnt create a rebound... secondly, a selfish player is not necessarily a better player for his teammates, as clearly if you are taking 1000 shots and scoring twice, your shot sucks... It was an analogy to prove a point, clearly you missed it.. Thirdly, this was Aulie's first season and he played 1st line minutes... Never said Myers is bad, but don't comment on players you know nothing about :)
I live in Toronto buddy. I think I know the Leafs. I've watched them on numerous occasions, and I've watched the Sabres as well. Myers is a closer comparison to Chara than Aulie. And Aulie playing first PAIR is cause the Leafs suck!
Look it up, the only thing Aulie is closer to Chara than Myers (on a per game basis) in PIM/game, and he's only 18 seconds closer at that. Myers is closer on pts/game, hits/game, blocked shots/game, TOI/game, shots/game, takeaways/game, giveaways/game, and size-wise as well. Myers is so much like Chara it's not even funny.
Holy shit are you still on this?? Its a fucking picture, get over it... I never said Myers wasn't more like Chara, in fact I never said anything about Myers Period! Fucking stupid people drive me nuts.. Leafs defense was terrible? Aulie played his fiorst year on the 1st LINE! He made minimal mistakes all season and was a huge asset. He beat out Schenn, Kaberle, Beauchemin, and Gunnarson for that top pairing. That is an accomplishment
1st, he only played 40 games - not even half a season. 2nd, he only played 19 minutes a game - 3rd on the team, so basically he was 2nd pair.
Yep, you're right, that must mean he is an awful player.. 3rd most in your first year player must mean you are destined to become a career AHLer, what was I thinking
Here are the Top 10 NHL Players when rated using Shift Impact Percentage:
Matt Martin - NYI - .43056
Paul Bissonnette - PHX - .39612
Alex Ovechkin - WSH - .39219
David Van Der Gulik - COL - .38028
Ben Holmstrom - PHI - .375
Dale Weise - NYR - .37362
Ryan Callahan - NYR - .37230
Ryan Reaves - STL - .37209
Nate Guenin - CBJ - .36842
Tim Kennedy - FLA - .35365
Interestingly enough, theres only one skilled guy on the list (Ovechkin) and even though hes super talented, he still throws the body around alot. Which is why i think this stat is favored towards player that dont get alot of ice time but go out and hit everything that moves. SabresHess, you need to come up with a way of valuing each stat this incorperates to make some more valuable than others so you dont get results like Matt Martin is the most impactful player in the league.
Maybe something like 2 points for a goal, 2 for a primary assist, 1 for secondary assist, 0.5 for a hit, 1 for shot, 1.5*takeaway-giveaway, and 0.5 for blocked shots. TubbyTreat, who are the leaders with these numbers?
You guys who are bashing this stat are missing the point... its not a "tell all" stat. Its to be taken in context with all the other stats out there like goals, TOI, ect ect ect. But good job Hess ... you put a lot of time into this. Enjoyed reading it.
Thanks for the comments. Tubbytreat i appreciate you doing the math, but realistically more than half of those players would not have qualified due to lack of games at the NHL level last year. I thought about weighing certain categories more than others, but then the weight would be arbitrarily assigned by myself and who am I to determine what is more important? I'll kick the tires on tweaking the stat and come back with some new ones by the start of the season.
ETP95 - There aren't stats collected for primary assist vs. secondary assist on NHL.com
SabreHess - I wrote a little app to scrape the NHL stats and then plugged your formula in to process the results. I can do this for all 891 players who played a single NHL game last year. I think to resolve some of the lesser players making the Top 10, we have to give weighting to the number of games played per number of possible games in a full season.
@Tubby- I have some interesting weighting ideas that should refine the stat, i'll PM them to you and if you can throw them in your app we can sift through the results.
I tried the following formula:((Hits Blocked Shots Assists Total Shots Takeaways-Giveaways) / Total Shifts) * Games Played / Total Possible Games = Shift Impact PercentageAlex Ovechkin - WSH - .37784
Matt Martin - NYI - .35705
Brian Boyle - NYR - .33707
Cal Clutterbuck - MIN - .32343
Shawn Thornton - BOS - .31987
Chris Neil - OTT - .30647
Dustin Byfuglien - ATL - .29677
Tuomo Ruutu - CAR - .29376
Dustin Brown - LAK - .27439
Ryan Callahan - NYR - .27241
Don't you guys know about Corsi, or QualComp, zonestart%?Do some research.