Over the last couple of weeks, I’ve introduced a couple of different metrics: SAF/100, SAA/100 and a ratio, shifts with at least 1+ SAF/shifts with at least 1+ SAA. I’m interested in using these metrics to take a look at some past decisions around the NHL, to see if they can shed any light on things.
I don’t know enough about Ryan Whitney to know whether he is a fitting return for Lubomir Visnovsky. I can, however, see what Tambellini’s thinking is there – if the Oilers are a ways away from being competitive, it makes sense to move wins from today to the future. The gamble is that Whitney will be a more efficient use of cap dollars between 2011-13 than will Lubo. I’m not entirely sure that this is true but there is at least a sign of thinking that makes sense.
-me, March 4, 2010
I’ve been developing some tools over the summer to help better describe what happens when a given player is on the ice. This has involved coming up with a way to look at players without zone start effects (open play Corsi%) (which has the salutary effect of also enabling us to look at specific aspects of play and how players are doing in it, which I’ll maybe get into in the future), breaking performance down into ability to generate/prevent the first shot attempt and prevent/generate subsequent shot attempts. It has, I think, provided some direction for where we should look in the future – I’m a big believer that there’s value in seeking to isolate what happens in terms of zone entries and zone exits and that we might find that that provides us with a big part of the code in generating really solid data about players, data that could be used to start talking intelligently about how some third or fourth liner will perform if he moves to another team.
I’m interested in applying these lenses to look at some past decisions and whether or not there’s anything that, if known at the time, might have been a red flag. This is a little bit of a problematic way of going about things, I acknowledge – it’s the decisions that didn’t pan out that catch the attention and you can get tunnel vision that makes you ignore the good things about the players. Even so, I think there’s some value in doing this. So let’s examine what we might have reasonably known about Lubomir Visnovsky and Ryan Whitney at 5v5 when those deals were made.
Visnovsky – 2007-08
Visnovsky had the third best open play Corsi% amongst D on a pretty bad Kings team by a pretty clear margin over number four. There are big caveats on the two guys in front of him as well – neither Dallman (34) nor Harrold (25) played very many games and they were down the totem pole in terms of 5v5 TOI, which probably suggests that they were getting softer competition. Lubo’s wasn’t super rough either – for some reason, LA decided to feed Jack Johnson to the wolves and guys who played with him got hammered when they did. That said, I think it’s fair to say that Lubo had the best open play Corsi% of anyone in LA’s top four.
Nothing too notable here – Visnovsky scores considerably better in terms of SAA/100 than the Kings other regular D but he didn’t spend a lot of time with Jack Johnson and Johnson got absolutely smoked when he was on the ice.
Again, clear separation between Lubo and the Kings regulars as far as the golden ratio goes – Jack Johnson’s number was abysmal, which probably hurt everyone who played with him.
Visnovsky – 2008-09
Lubo was traded to Edmonton in the summer of 2008 in exchange for Matt Greene and Jarret Stoll. How did he make out in that first season in Edmonton?
There were three clear tiers amongst OIler defencemen here: Lubo/Grebeshkov, Souray/Gilbert and then the rest.
The SAA/100 numbers are pretty flat – not a lot of difference amongst the various defencemen.
We see here what we saw with the open play Corsi% numbers – three tiers: Lubo/Grebeshkov, Souray/Gilbert and then the rest. It’s a funny thing, the way that life unfolds – if Steve Tambellini had never happened and Kevin Lowe and MacT had kept running/coaching the Oilers, it wouldn’t have surprised me if they’d found a way to create a playoff team by 2009-10. The 2008-09 was just brutal in terms of depth and that’s something that Lowe and MacTavish had a pretty good track record in terms of solving.
Instead, Tambo did happen, MacT was fired/resigned, the world spun off its orbit and four years later, the Oilers have Taylor Hall, RNH, Nail Yakupov and Darnell Nurse. Lots of top end talent but the same fundamental problem that the team faced at the end of the 2008-09 season: finding depth.
Visnovsky – 2009-10
This was Lubo’s last season with the Oilers, as he was moved to Anaheim at the trade deadline for Whitney. Open play Corsi%:
Again, there are three tiers of defencemen that you can see: Lubo/Souray, Gilbert/Grebeshkov/Smid/Staios and then Strudwick et al. I’m probably being generous to Souray in slotting him as part of a tier with Lubo – Lubo pretty clearly scores the best amongst the Oilers D by this metric.
This is kind of interesting to me, this graph, because there are a lot of horrible numbers on those 2009-10 Oilers – nobody’s under 150 in terms of SAA/100, which is abysmal. Lubo and Souray are the best of the defencemen, clearly ahead of everyone else except Steve Staios which is funny because Pat Quinn didn’t seem to like Lubo.
I’ll throw in an extra graph here, comparing SAA/100 for the Oilers D who were on the team in 2008-09 and 2009-10, because I think it’s awfully interesting.
If you want to talk about what changed from MacT to Pat Quinn, one of the things that changed was that once the opposition got that first shot, they became significantly more likely to get another. This was true of all the defencemen. If this data had been available in real time, it would have been interesting to get the coaching staff’s thoughts on why that might have happened, what changed from the preceding year, etc.
The golden ratio chart:
Hmm. Lubo’s a pretty clear winner. There’s no zone start effect in this data so this is pretty clean stuff and he’s pretty clearly at the head of the class. Let’s contrast MacT’s team with Pat Quinn’s team again:
This is awfully interesting – on balance, I’d say that the 2009-10 Oilers D performed better at this than did the 2008-09 Oilers. Lubo, Gilbert and Souray were all in the same ballpark, Staios and Smid improved considerably and Peckham, Strudwick and (somewhat worryingly) Grebeshkov took big steps backwards.
Again – had Quinn coached the 2010-11 Oilers, this might have been fuel for a pretty interesting discussion. What, if anything, was done differently from 2008-09 to 2009-10 that let some players perform better? Was there a way to retain that benefit while simultaneously cutting down on the gong show in the Oilers’ end once a shot attempt had occurred? Who knows.
Whitney – 2007-08
Hal Gill and Mark Eaton didn’t play a ton with the 2007-08 Penguins, so the real regulars are the guys from Gonchar to Scuderi, inclusive. Whitney’s Corsi% was nothing to write home about, both as a raw number (44.1%) and within the context of his team.
Nothing too notable here, although it’s probably worth mentioning that a 149.5 SAA/100 in 2007-08 is pretty awful. It does appear to have been a teamwide issue though.
The ratios are kind of interesting. Whitney was always sold as a puckmoving defenceman, a guy like a Gonchar or a Kris Letang (who was a rookie in 2007-08). His numbers here were nowhere near what happened when Gonchar was on the ice and somewhat worse than Letang as well. Whitney’s kind of in a group of guys who aren’t particularly reknowned for their ability to make plays with the puck.
Whitney – 2008-09 (Penguins)
Yikes. That’s horrific.
Those are pretty brutal SAF/100 and SAA/100 numbers too. Anything over 150 SAA/100 is really bad – Whitney was at 160. Teams were just unloading on the Penguins when he was on the ice.
And, to complete the set, Whitney’s shift ratio is significantly worse than everyone else on the Penguins for that season.
Whitney was coming off of surgery that cost him 31 games at the start of the 2008-09 season but I think it’s pretty reasonable to say that there wasn’t much when he came back that said that he was a very good defenceman. He sticks out as being markedly worse than the rest of the Penguins D. When Anaheim acquired him, Bob Murray said this:
He’s a big kid who can skate and move the puck. The game is evolving, and you better get the puck out of your own zone. We have good forwards and you have to get the puck to the forwards. Whitney can do that. He also is very good on the power play. He can shoot the puck and sees the ice very well. I don’t know if you guys can understand how tough it is to find a puck-moving defenseman. I guarantee you it’s harder to get a top-four defenseman than a top-six forward.
Awfully hard to see these skills having any positive impact on Whitney’s numbers – it just doesn’t seem to have been happening.
Whitney – 2008-09 (Ducks)
As a trade deadline acquisition, Whitney didn’t play a lot of games with the 2008-09 Ducks but, as with the Penguins, his numbers were pretty bad relative to the rest of the team’s defencemen:
Whitney – 2009-10 (Ducks)
New season, same problems. Again Whitney’s got an open play Corsi% that’s a) bad and b) not particularly special relative to his teammates. Aaron Ward came in after Whitney left and was just months from retirement at this point; his open play Corsi% was better than Whitney’s.
The SAA/100 number isn’t terrible but the SAF/100 number is nothing special. There’s something odd about this – it’s a trend with Whitney. For an offensive defenceman, his skills didn’t seem to translate into generating multiple shot attempts in a shift.
And there’s the ratio problem again: he’s ever-so-slightly worse than Aaron Ward, who’s comparable in the sense of being a good talker but they aren’t really supposed to be comparable players.
What do I take from all of this? Well, if someone asked me to summarize the data on Lubo, I’d say that he’d pretty clearly been the best puck possession D on three bad hockey teams at the time that the trade was made and, despite the Oilers being poor, he’d posted numbers north of 50% in terms of Corsi% that seemed to be tied to an ability to generate more shifts on which his teams took shots than shifts on which they allowed shots.
Whitney though…man. There’s just nothing about him that stands out in a good way. He’s played on a good team in Pittsburgh and posted unimpressive numbers, both in the context of the team and in a raw sense. If you look at his numbers, you conclude that his team’s spending more time in its own end than the other team’s end over and over and over.
It’s obvious that the Oilers scouts liked Ryan Whitney; it’s equally obvious that Anaheim’s scouts liked him before they gave up Chris Kunitz, a player so good that he might knock Taylor Hall off the Olympic team. Pittsburgh must have liked what they saw as well when they signed him to a big contract. I’m not sure how one would go about reconciling his possession numbers with so many team’s pro scouts watching him.
That’s the question that the Oilers should have been asking though: why does a player who the scouts love so much not have more of a positive effect on the results when he’s on the ice? I’ve got a couple of theories as to why this might be – hockey teams seem to love a player with some size who can make a pass (I’m thinking of Jack Johnson); they seem to sort of lose the ability to accurately weight the various components of a player’s game. It’s a theory, nothing more, but there do seem to be a certain class of players with whom there’s a total disconnect between the results and what teams think of the player.
I’m hesitant to condemn anyone with this sort of analysis because I’m making it knowing who things turned out: Lubo’s received Norris votes since the trade and Whitney, despite being the younger player, barely found a contract this year. That said, facts are facts and, had the Oilers had this information in 2010, it might have injected a discordant note into their internal discussions that caused some efforts to reconcile this data with what the scouts think. Maybe they would have ultimately decided that the scouts were right, maybe they wouldn’t have but a team that’s got more relevant data for its analysis is going to make better decisions over time than a team with less.Email Tyler Dellow at firstname.lastname@example.org