With the boring theory part of SAA/100 out of the way (except for one point, which I’ll deal with below), it’s time for the fun stuff: results. After I posted the theory post, I got a bunch of requests on Twitter for data relating to specific players. It turns out that it’s kind of fun to look guys up, so I figured I’d do a post with some numbers for specific players that touches on some things.
Vegetables first though. In a comment to the previous post, which was picked up by Tom Tango, a concern was expressed about the first shot effect. Tom expresses it rather succinctly, so I’ll quote him:
if you have the NY Eve Russians v Canadiens on the ice, they wouldn’t need that second shot, because the first shot would have produced a goal. So, they are “bad” at creating multiple shots, only because they made sure that they got a good first shot off.
Now, keep in mind that this is a defensive stat. So if there is a NY Eve Russians team out there, a team’s only playing a small part of its schedule against that team. Also keep in mind that it’s questionable in today’s NHL that teams exist that pass the puck into the net, as the old Russians used to do. It’s also questionable just how much the Russians actually did that – my buddy Pat McLean did a breakdown of the 1972 Summit Series, where the Russians were supposedly much the same way, and counted scoring chances. This was his conclusion:
This was Canada’s series from start to finish. Only in game five, which Sinden would later call Canada’s best game (?!), were the Canadians outchanced (and badly for that matter) at ES. Throw that game out and the ES numbers are even better looking for Canada.
Bad goaltending, poor discipline and some bad luck made this a much closer series but even with these factors Canada was the dominant team. Put together the roster that swept the last three games and play them in an eight game series against the Soviets one hundred times and Canada wins the series between seventy five to eighty five times. Seriously. I don’t doubt that for a minute. And I never thought that would be the case.
Not at all the narrative that has been hammered into a generation of Canadians: they were more skilled but our guys had heart and they were godless Commies who didn’t have heart so that’s why we won. We outchanced them like crazy according to Pat but we had lousy goaltending and a taste for dumb penalties. I digress.
Since 2007-08, Tampa Bay has the worst rate of allowing goals on open play shot attempts – 5.0% of shot attempts against the Lightning turn into goals. The best rate has been that posted by the New York Rangers – 4.0% of shot attempts against them turn into goals. You’ll notice that this is awfully low.
My thinking is that this is unlikely to affect things too much. The problem with talking about records when a team scores first is that right away you’re talking about their record in games in which they scored at least one goal. There aren’t a lot of goals in a hockey game. There are a lot of shot attempts.
The other thing is that some of those goals are going to come following a sequence of shot attempts. That’s an interesting question and one that I may delve into in the future – do a disproportionate number of goals come on shifts with multiple shot attempts?.
What this objection is really saying is “What if the numbers are skewed by the fact of some shot attempts turning into goals?” It’s probably worth doing a quick check on at some point but given that a) the vast majority of shot attempts don’t turn into goals and b) some of those goals will come following multiple shot attempts, I have serious doubts that it makes any significant difference.
* * *
So – onto some requests and guys who I looked up last night. If none of this makes sense to you, go and read the post leading up to this one. I’m going to generally present two graphs for each player: his raw SAA/100 and then where it ranks on a percentile basis. As I noted in the previous post, SAA/100 has been drifting upwards and it’s happened enough that what might have been poor in 2007-08 is good now. For that reason, as well as to make a sense of scale available to people, I’m including both.
It’s not a mistake that there’s no bar in the S.J. 2013 column for Murray’s percentile – he was the worst in the NHL, so there’s no percentile ranking showing. So, if you’re a San Jose fan who thought that Murray fell off the face of the planet, here’s a pretty strong data point in support of that. You sort of wonder why Pittsburgh traded for him and, after they decided to let him go despite spending two second round picks to get him, why Montreal decided to sink resources into him. I’d be shocked if he’s anything more than a third pairing guy in Montreal by the end of the year.
Jeff Petry/Ladislav Smid
There’s a lot of interesting stuff in there, I think. For all the talk about guys like Ladislav Smid maybe taking a step back in 2013, his SAA/100 was almost identical, year-over-year. I’ve talked a lot this summer about how the Oilers performance changed last year and Smid took the same hit that a lot of guys took – his Corsi% fell from 47.6% to 44.1%. It’s intriguing to me that this doesn’t really show up in his SAA/100 – he was basically the same as he was last year, in terms of how many shot attempts the opposition got after the first one. Logically, this suggests that teams were getting that first shot a lot more. This points, again, towards a change in what the Oilers did in 2013 that made it easier for teams to get that first shot.
You’ll note that Petry’s number kind of bumped upwards this year. I’d guess that this happened when he was separated from Ladislav Smid. I included Justin Schultz because I think it’s instructive. There’s lots of muttering about what his second contract should look like now that he and Nail Yakupov are next in line. I’m less sold on Schultz than I was on RNH and Hall because I don’t think that the underlying numbers are really there for Schultz yet. Is a number one defenceman (even in the Mike Green mold) or is he more of an offensive whiz with defensive deficiencies that won’t resolve? I’m not critical of where he’s at; I just think people should be realistic about where he’s at. I’m looking forward to see how he does in his own end this year.
Not sure I have too much interesting to say here. Pretty consistent up until last year, his first in Ottawa, when he played with Eric Gryba and Erik Karlsson. Maybe there’s a positive correlation between “partners named Eric/k” and SAA/100.
That 2007-08 Thrashers outfit was one of the great tire fires in NHL history. It’s interesting how Hossa’s actually put up decent numbers as far as SAA/100 since getting out of Atlanta. Obviously, numbers are going to be team dependent to an extent and it may be that wingers have less of an effect on this than most. If you’ve got bad centres and a defence corps that can’t get hold of the puck and move it, you’re probably not going to score well here.
The New York Rangers: 2008-09 – 2013
This…honestly, this needs an entire post and maybe more. It will get one.
Breaking: the numbers say Alex Pietrangelo’s pretty awesome at this aspect of playing defence. The challenge in using numbers a lot of the time is looking at the number, understanding what it says about performance and then kind of challenging yourself to come up with ways in which the number might lie to you. It’s hard to think of a way in which the numbers could be lying about Pietrangelo here – he plays against the best and there’s no awesome defenceman hanging around who might be carrying him. It’s hard to find anything to dislike about the guy.
Age comes for us all. I’ve alluded to ratio of shifts with a SAF/SAA – Lidstrom aged a bit better there. He was still a good defenceman when he left the NHL but he wasn’t what he was. No shame in that but that 2010-11 Norris Trophy might be best understood as a parting gift.
Sam Gagner/Shawn Horcoff
I thought that this was a fitting pair on which to end things because it lets us start getting into thinking about the differences about players. Horcoff’s had a reputations as reasonably good player in his own end; Gagner has had a reputation as a guy who’s a bit small and loses position on guys from time to time. Sure enough, a difference shows up in the data. Keep in mind, Horcoff probably played against a higher class of player too, particularly in the early years, so the difference between the two of them is probably larger than it seems.
Also of note: look how both guys posted awful numbers during the Quinn year and then reverted back to much better numbers the following year. Ultimately, as I’ve said time and time again, what I’m interested in is finding way to isolate differences between players, to shine a light on questions like “How much does the difference between Sam Gagner and Shawn Horcoff defensively cost you?” I think that this stuff is a step in the right direction.
I’ll end this here for now – I might throw up another one of these over the weekend. Requests outstanding include Keith Ballard pre/post Vancouver, Jordan Staal, John Erskine, Joe Thornton, Patrice Bergeron, Erik Cole, Ales Hemsky, John Carlson, Marc Andre Bergeron, Andy Greene and a comparison of Lubo and Ryan Whitney. If you’ve got one, feel free to drop it in the comments and I’ll see what I can do.Email Tyler Dellow at firstname.lastname@example.org