A new way to look at player value
by JustinM on Jan 30, 2011 12:05 PM EST
in Analysis
Email
PrintManaging a salary cap in today's NHL isn't simply about making sure you don't go over—apologies to Mr. Lamoriello—but about making sure you get proper value for your cap expenditures. While I am not going to discuss here what is the right and proper cap hit for someone with Deryk Engelland's skill set, we can talk about who is a relative bargain so far this year and whose salary is an albatross.
Jump with me.
I constructed an extremely rudimentary value measurement that could stand to be refined greatly, but for now we'll just call it Value. Value is defined for this purpose as GVT divided by cap hit to date expressed in millions. So, to take Sidney Crosby as an example, a GVT of 17 divided by a cap hit to date of $5.425806MM gives you about 3.13. I'm positive that someone else has already done this, and there might even be a website dedicated to it, but I haven't found such a thing.
The unquestionable winner in this contest, if you want to call it that, is Dustin Jeffrey, with a score of 26.96. However, he's only played 7 games on his nearly-minimum contract, so his cap hit to date is hilariously low. If we only include people who have been semi-regulars on the team, the scores look a little different.
Penguin Player Bargain Scores
Name GVT Cap Hit Bargain Chris Conner 2.2 245430 8.96 Alex Goligoski 8.6 1143369 7.52 Tyler Kennedy 3.1 452151 6.86 Mark Letestu 2.1 311828 6.73 Ben Lovejoy 2.2 327419 6.72 Kris Letang 14.3 2182796 6.55 Pascal Dupuis 5.1 873118 5.84 Craig Adams 1.7 343011 4.96 Matt Cooke 5.3 1122581 4.72 Maxime Talbot 2.4 654839 3.67 Sidney Crosby 17.0 5425806 3.13 Deryk Engelland 0.9 311828 2.89 Chris Kunitz 5.0 2323118 2.15 Arron Asham 0.9 436559 2.06 Brooks Orpik 4.0 2338710 1.71 Paul Martin 4.3 3118280 1.38 Evgeni Malkin 4.7 5425806 0.87 Michael Rupp 0.3 514516 0.58 Jordan Staal 0.9 2494624 0.36 Zbynek Michalek 0.6 2494624 0.24 Eric Godard 0.0 467742 0.00 Mike Comrie -0.3 311828 -0.96
Now, one must keep in mind that this does not mean that Conner is "better than" Crosby. What you can take from this table are that a) you can see who is overperforming their current salary, and b) there are definite diminishing returns when comparing cost and performance, perhaps even when talking about players of Crosby's caliber.
Small caveat as well: it's a common misconception that time on LTIR lowers a player's cap hit. It does not. The team gets a cap credit under certain circumstances when a player has been on LTIR, but it doesn't actually affect that player's salary cap number. See Cap Geek for more information, under the "How does long-term injured reserve work?" header.
62 comments
|
2 recs |
Do you like this story?
Comments
Puck (Hockey?) Prospectus frequently refers to GVS, which I think is GVT per million or money per 1 GVT, but I don’t think it’s readily listed.
I think I’ve seen someone show before that roughly, top-six forwards get twice the money per GVT than bottom-sixers (flashy offensive stars vs defensive plumbers and faceoff specialists). Obviously on the Penguins the dynamic is a bit different, with the big three centers up front and everyone else, so I suppose it’s to be expected that Sid, Geno, and Staal’s frequent linemates would outperform their cap hits. Good to see it quantified.
I’m a little surprised to see Martin and Michalek so low, but I guess when your primary duty is defense, it’s harder to look good by GVT.
My blog and Twitter, featuring coverage of the most frustrating team in the NHL
If you don't know how to use Timeonice, read this.
Behindthenet quick link to QoC/QoT/Corsi/PDO/Zonestarts
"Numbers don't lie, they just don't agree with you"--George E. Ays
I haven’t paid to subscribe to Hockey Prospectus, so I really don’t know what they’re doing over there these days. I wouldn’t be surprised if I’m reinventing the wheel, though.
Theoretically, Martin and Michalek should be on equal footing with others through GVT considering that the offensive metric and the defensive metric are calculated differently. Michalek’s and Staal’s injury problems hurt them in this area, though, and it’s why I’m considering doing a post on GVT per TOI as a way of looking at who’s doing well perhaps without us noticing.
There are always snippets at the top and every now and then an article is free.
DGVT tends to undervalue defensive contribution. Either that or the margin between players defensively is substantially thinner than the margin offensively. At least, that’s my understanding.
My blog and Twitter, featuring coverage of the most frustrating team in the NHL
If you don't know how to use Timeonice, read this.
Behindthenet quick link to QoC/QoT/Corsi/PDO/Zonestarts
"Numbers don't lie, they just don't agree with you"--George E. Ays
by red army line on Jan 30, 2011 1:10 PM EST up reply actions
No offense
But these things really bug me. They only examine value based on one constraint, budget. In reality, there are two constraints, budget and roster space. Teams face limited roster space, having one player be able to contribute the worth of two players allows them to free up a spot for another player. This only assigns value for contributing to one constraint.
I don’t know mathematically how to express value using the second constraint, but this metric is inaccurate because it glosses over that constraint.
How do you decide when someone’s reached the “worth” of two players? Twice the replacement value? That’s putting the bar awfully low. Twice league average? That’s putting the bar awfully high.
I was using twice as an example.
When faced with limited roster space, a player who is worth x more than the average player represents value. you have x dollars and x roster spots to allocate towards maximizing your teams utility. This metric relates value only towards maximizing the utility of dollars spent. It fully ignores the second constraint.
This metric would suggest that, since both have relatively low GVT’s and high salaries.
But from the old fashioned eyeball test, I’d say that both are pulling their weight on the ice and have been worth their weight in salary cap so far.
"Game's the same. Just got more fierce."
When did I say that?
Using conventional stats, though, Michalek still has a lot to prove, injury or not, he’s got the worst +/- on the team.
It’s been good and bad with his adjustment to the Pens.
The negative is his plus/minus, a couple of defensive miscues, high number of giveaways etc.
But on the plus side: Michalek’s playing 3:47 on the top ranked PK in the league (2nd on team behind Orpik), most blocked shots on the team (82).
Hopefully as he continues to acclimate to the team and get more comfortable in PIttsburgh’s system, more postives will emerge and the negatives will fade a little more.
"Game's the same. Just got more fierce."
Yeah, what’s the rule of thumb? 50 games? I’m not panicking by any means. It may turn out to be a good thing he and Martin are under contract for so long. Let’s not forget Gonchar’s first year in Pittsburgh.
Why would we panic? Look at the team leaders in GA/G and SA/G. These guys have been HUGE for us.
Hockey Blogger at Pensburgh.com
78 defensemen have more giveaways than Michalek. It’s tough to say that he’s got a giveaway problem.
Hockey Blogger at Pensburgh.com
His on-ice shooting percentage is abysmally low (191 out of 208 defensemen). Nothing to do with Michalek, everything to do with bad luck.
Hockey Blogger at Pensburgh.com
Martin and Michalek
I agree that it seems odd that both Martin and Michalek are ranked so low. To me that just means that the rankings are too biased on offensive metrics. Decent idea though.
My thoughts on M&M: Scuderi had only 1 goal and 15 ish assists the last yr he was in a Pens uni, Michalek has similar numbers but is better on D & younger so he is worth it. Martin is playing well but I think he still is settling in and will have some big numbers in the near future – he also is really solid on D.
I think Letang, Letestu, Connor and Staal this year have been playing about 3x their salaries, respectively.
Get in the fast lane grandma, the bingo game is ready to roll!
by PensForTheWin on Jan 30, 2011 5:03 PM EST up reply actions
Nate Silver did an analysis like this for baseball players
From memory, he demonstrated that players like Arod, Pujols, etc. produce at such a rate that they can make the difference that pushes an 85 win mediocrity over the top to become a 95 win pennant winner. The reason: There are so few players like this, players that produce at such a high rate, that they are worth their very high salaries. If a team wants to contend for a championship, it needs at least one superstar.
Having Crosby on the team implies having X more wins than a replacement player? 5? 10? Considered in this way, Crosby might not have the best GVT/Cap Hit ratio, but his GVT is such that he provides wins for the team that mere mortals, whether a bargain or not, will never provide.
I’d bet the GVT/Cap Hit ratio matters most for high salary underachievers and low salary over achievers. The low salary over achievers are the kind of players all contenders need. They need them because they need the production and the need the low cap hit, which they can use to play players who are high salary over achievers and achievers. Naturally, the high salary under achievers kill teams in a salary cap league.
s.zielinski
Your last paragraph is exactly what I was trying to get at. Sid’s not really pertinent to the discussion because of course you’re going to pay a premium for skills like his.
But while in baseball they’ve determined that a win is equivalent to about 10 runs, I don’t know if they’ve come up with a similar number for hockey and goals. If I had to guess, it’d be around 6 or 7, in which case even the best players in the league are only worth about 5 wins a year. I find that hard to believe.
The number that people tend to use is 6 goals = 1 win. In reality it’s between 5 and 6, depends on the season.
My blog and Twitter, featuring coverage of the most frustrating team in the NHL
If you don't know how to use Timeonice, read this.
Behindthenet quick link to QoC/QoT/Corsi/PDO/Zonestarts
"Numbers don't lie, they just don't agree with you"--George E. Ays
by red army line on Jan 30, 2011 10:03 PM EST up reply actions
Crosby prorated
would have compiled a GVT of 34 over 82 games. That’s about 6-7 wins. And he was having a great season! To make one comparison, Barry Bonds had a WARP of 13 in the 2001 season, the year he hit 72 HRs. To make another, Gretzky’s all time best was 44 and Mario’s was 42.
s.zielinski
Interesting post, Justin — thanks.
One reason there is wide variation in value is that market for a player at any given time often determines his salary. We had to pay 8.7 million to keep Sidney Crosby, and the counterfactual (what if we had decided he was too expensive and let him leave) obviously would be much worse for our team, even if we ended up with cheaper players who gave us greater value per dollar.
This is what makes cap management so tricky — you need to maximize value when you can, but you also have to pay more for certain players because you are competing with other teams. You’re not just deciding what each player is worth and paying them accordingly — you’re deciding who you want and paying them the minimum amount necessary to keep them, which often tends to be more than their “value” according to these calculations. Yet they add more to the team than other players, and we’d be in bad shape without them. I don’t envy Ray Shero — it’s very tough to make these decisions.
I believe in Geno.
by Cari on Jan 30, 2011 9:06 PM EST reply actions 1 recs
Actually, I have the feeling that Crosby would have commanded cap max on the open market. I’m not sure what that would have been in his contract year, but it’s assuredly more than $8.7MM.
Yeah, I agree — he definitely took less to stay here, and he’s worth even more than his salary because of what he brings to a team. But it’s still the case that we had to pay him a lot of money because there was a market for him. If nobody else wanted Sidney Crosby, we could have paid him even less, and his “value” (according to this statistic) would be higher. So even though he took less to stay here, the market still factored into his salary.
I believe in Geno.
I think Sid could’ve snagged a little over $10 mil per year, so his discount was pretty solid (roughly 10%).
by mcnulty4prez on Jan 30, 2011 10:37 PM EST up reply actions
I think he could’ve easily gotten the max at the time ($12 million).
My Thoughts on the Pens & Twitter
"A player's game is a composition. Sometimes it's a line, sometimes a paragraph. Right now, Crosby's game is a novel." -Scott Burnside
by AllieLXXXVII on Jan 30, 2011 10:42 PM EST up reply actions
I was thinking they could only pay a player 20% of the cap in place at the time he signed his contract.
.
Follow me on twitter, Picasa or at Pens Through My Lens. Or don't. Whatever.
by PensAreYourDaddy on Jan 30, 2011 11:06 PM EST up reply actions
mcnulty4prez has the right #s
Penguins extended Crosby in 2007 when the cap was $50.3M/year
the max per player is 20%
so, the max Crosby could have received was $10.06M/year
everyone agrees he would have received had he asked
he gave the Penguins a 13.5% discount
he gave the Penguins a 13.5% discount
which, in practical terms, pays for the contract of Pascal Dupuis
Let’s hope we can get some more Rollback pricing when it’s time to re-sign him.
.
Follow me on twitter, Picasa or at Pens Through My Lens. Or don't. Whatever.
by PensAreYourDaddy on Jan 30, 2011 11:17 PM EST up reply actions
As would I.
My Thoughts on the Pens & Twitter
"A player's game is a composition. Sometimes it's a line, sometimes a paragraph. Right now, Crosby's game is a novel." -Scott Burnside
by AllieLXXXVII on Jan 30, 2011 11:51 PM EST up reply actions
Wow
My Thoughts on the Pens & Twitter
"A player's game is a composition. Sometimes it's a line, sometimes a paragraph. Right now, Crosby's game is a novel." -Scott Burnside
by AllieLXXXVII on Jan 30, 2011 11:17 PM EST up reply actions
Thanks for doing that math =]
My Thoughts on the Pens & Twitter
"A player's game is a composition. Sometimes it's a line, sometimes a paragraph. Right now, Crosby's game is a novel." -Scott Burnside
by AllieLXXXVII on Jan 30, 2011 11:17 PM EST up reply actions
I get that, but eyeball test for what? What eyeball test can you possibly have for seeing if a guy is a bargain or not?
by JustinM on Jan 30, 2011 10:00 PM EST up reply actions 2 recs
using the eyeball test
and then grading value on an A thru F on value – THIS year only
something like this
NAME Cap Hit Value of Contract
Crosby 8.7 A+++
Malkin 8.7 F. I know he’s injured, and I love Malkin, but he’s not producing for what he’s being paid.
Staal 4.0 Incomplete. Not enough GP.
Letang 3.5 A+++
Orpik 3.75 A
Cooke 1.8 A-
Dupuis 1.4 A
Talbot 1.05 B+
Rupp 0.85 B+
Michalek 4.0 B
Martin 5.0 B
Goligoski 1.83 A
Something along those lines…
“eyeball test” just means opinion… Malkin might not be scoring 8.7 million worth of goals, but we’re not getting nothing for his contract. F is really harsh and implies he brings no value to the team.
I believe in Geno.
didn’t mean that 71 has brought nothing to the team this year, just that his CONTRACT vis-a-vis his output is poor.
That’s really what the metric above is at least attempting to get at — how good are the contracts?
They do this by dividing GVT by the Cap Hit in $.
But it doesn’t work.
It suggest that, for example, Connor is 3x as valuable as Crosby.
It just means we get more per dollar from Conner, which makes sense because Conner isn’t paid very much. And we didn’t have to compete with other teams to sign him, so his price wasn’t increased by competing offers. It doesn’t mean Conner is more valuable than Crosby. Crosby is obviously much, much more valuable. It just suggests that Crosby isn’t 16 times more valuable than Conner (which is what you might think looking only at their salaries).
I believe in Geno.
It's hard to put any player on team as a number of times more valuable than Conner
but for just throwing shit out there… I would say we are better off with 1 Crosby than 16 Connors so :)
I'll take Sather's eyeballs from the 80s
They don’t work so good these days
I'm expecting 100% improvement from the Lions this season.
I like the info JustinM but I think it’s incomplete. To see if Sid or Geno are earning the $8.7 you have to compare them to other players in the $8.7 ballpark around the NHL not lower paid team mates. The same with Conner compare him to other $500k NHL players.
by threeheadedmonster on Jan 31, 2011 9:04 AM EST via mobile reply actions
Michalek & Martin
Interesting post, Justin. Thanks!
Re. Michalek & Martin (this is sort of a general response after reading the comments above …):
I think the first thing to remember is the time it takes for defensemen to adjust to a new system — which is typically longer than it takes for forwards to adjust. This was definitely the case for M&M, although it’s easy to forget amid the rough start for the Pens, and the subsequent monster streak the team went on. It might be interesting to compare their first/second half numbers, or this year versus next year (let me know when your time machine is complete). Also, while I think there would be too many caveats to include with this data, it would also be interesting to consider what their numbers were last season with their respective clubs.
Second, as Cari mentioned, there’s a dimension that isn’t quite captured here by the “signed as a free agent” vs. “brought up the ranks”. The interesting comparison here seems to be Orpik vs. Martin. Although, again, I think the comparison will be more interesting after a full season with the club.
I'm looking forward to reading Crosby's novel.
Just for fun, I took a look at how these numbers work out for a few of our friends around the league.
Player, Bargain (GVT, CapHit):
Kovalchuk, 0.477 (2.0, 4.193548)
Boogard, -0.0978 (-0.1, 1.022177)
Kessel, 1.325 (4.5, 3.396774) [Le Sigh]
Seguin, 4.063 (2.3, .566129) [Le Double Sigh]
Hall, T, 8.302 (4.7, .566129)
Stamkos, 31.43 (17.3, .550403)
St. Louis, 3.906 (12.9, 3.302419)
Gonchar, 1.156 (4.0, 3.459677)
Ovechkin, 1.95 (11.7, 6.0)
Backstrom, 2.135 (9.0, 4.214516)
Semin, 2.16 (8.1, 3.744194)
Lots could be said about these numbers, both in defense of players, or in mockery, but I’ll just point out one thing: Stamkos is still on his entry-level contract. Which he is obviously out-playing — even if you add in the bonuses.
I'm looking forward to reading Crosby's novel.
Yeah, I think entry-level contracts shouldn’t be compared directly with standard contracts. There’s no way the comparison will be apples to apples.
No, I think that illustrates quite effectively how powerful ELC contributors really are.
"Essentially, all models are wrong, but some are useful" George E.P. Box
by Knee high to a duck on Feb 1, 2011 6:20 PM EST up reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on PensBurgh
FanPosts
Community blog posts and discussion.
Recent FanPosts
In Association With
Latest Puck Daddy - NHL from Yahoo! Sports
- NHL Playoff Beard Watch: Blood, Sweat and Beards of unsung unshaven heroes
- Devils guarantee they’ll downplay 1994 conference finals for Game 6 vs. Rangers
- Dodger Stadium Winter Classic; 10 worst NHL fans; ‘While The Men Watch’ debate (Puck Headlines)
- Lightning GM Yzerman suggests he won’t be trading for Luongo, strengthens position for Luongo trade
- Marek Vs. Wyshynski Radio: Ron Hextall, Luke Richardson and Rangers/Devils
Latest stories and headlines from NHL.com
- Despite loss, Phoenix Coyotes build solid foundation in the desert with run to West finals
- Worth every penny: LA Kings D Drew Doughty comes through with stellar playoff performances
- Coyotes finish off best season ever
- Doughty enjoying breakthrough playoff for LA Kings
- Maple Leafs sign defensive prospect Petter Granberg to entry level deal
