SBNation.com User Blog: six hole

Trying to quantify how unfair these "retirement" contracts are (Kovalchuk and others)

six hole — Tue, 10 Aug 2010 15:03:25 -0000

Piggybacking on posts by Crabby Appleton and cobracg, and given that the NHL may be reconsidering a few other contracts, I wanted to try to put numbers to why many of these retirement contracts "feel" wrong. I think Crabby was on to something good regarding how much cap-hit-savings is available for an early retirement. Taking another approach, I'm going to apply the GINI coefficient, a standard measure of inequality in the social sciences, to compare these contracts. At best it might suggest a bright line standard the NHL could use (e.g. "anything over .2 is an illegal contract") and at worst you might get exposed to some statistics.

First, a little math background.

I'm no stats whiz; I'll just give a few simple examples and let you figure out where to read more. The idea of Gini is to compare a set of data across time/space/people to a theoretical perfect in which everything is distributed equally. To do that you have to rank everything or everybody (by dollars, let's say) and see how different that would be from perfect equality. You wind up with a curve underneath a straight line, and the bigger the area difference between them, the more unequal:

via www.nihonkaigaku.org

For something like income inequality Brazil is at around .61, the US at .44, and Norway at .28. There's a theoretical maximum of 1 (one person owns everything, billions of others nothing) and minimum of 0 (everyone makes exactly the same amount).

Let's shift to hockey contracts. A 10-year contract where the player makes $1.45M every year has a Gini of 0. Another 10-year contract where he makes $14.5M in year one and then zero for the next nine years has a Gini of .90 (it isn't 1 because it'd have to be infinite years out of making nothing to get that). Obviously that contract doesn't fit any minimum salary rules; also you can see that Gini doesn't really approach 1 in practice. How about the most fanciful spirit-of-the-CBA-breaking contract we can imagine: $10M in year one, $500k in years 2-10. That would still be a cap hit of $1.45M and a Gini of .59. So that's a sorta-practical range from zero to .6.

Where would Kovalchuk have been? 17 years for $102M structured his way is Gini of .425. Definitely the worst in the league. His comrades:

Kovalchuk* - .425
Savard - .381
Pronger - .32
Hossa - .306
Luongo - .264
Keith - .237
Franzen - .218
Lecavalier - .214
Kiprusoff - .2
Zetterberg - .192
Richards - .154
Heatley - .115

I don't know that these are actually the 12 worst, just the ones that are commonly mentioned as fishy and a couple others that I saw and wanted to quantify. And Richards and Heatley may well not be 11 and 12 given the way the numbers drop (.192 to .154 to .115). For Pronger I think we all now know that Philly isn't getting any benefit of cap avoidance since he's an over-35. But they were clearly trying for that and screwed up. Also a few foolish people have tried to suggest Backstrom (.058) and Ovechkin (.026) belong in this fishy grouping. Uhh, they don't.

Caveats - You would expect to see some fall-off in salary as a player reaches the end of his career. Modano is about to go $4.25M, 2.25, 2.25, 1.75 in consecutive years (Gini .179). Of course that's over three contracts, to past age 40. But Kiprusoff's drop-off ($6M, 5, 1.5) seems more defensible than Lecavalier ($8.5M, 4, 1.5, 1). With a contract like Keith's it may be very defensible (from $6M in '16-17 at age 33 gradually falling to $1.5M in '22-23 at age 39). But by throwing in so many of the potential retirement years it does look very fishy, and drives up the Gini score to 6th in this list. And I guess statistically Gini is much more useful for large datasets, or in this case, longer contracts. Numbers can be thrown sharply higher by just tacking on an couple years to a short contract (ahem, Savard and Pronger). But maybe that's why Gini is useful here, in making them stand out. Kovy had to add 6 fake years to get to number 1.

So this could be the basis for some NHL-approvable contract criteria. Excluding .2 and above knocks off the 10 contracts many of us would call the most egregious. Other bright-line criteria could include 1) maximum potential savings at any point in the contract, relative to cap hit (Crabby), 2) a maximum percent salary drop from year to year in the contract, 3) a maximum percent drop from the highest to lowest years in a contract, 4) maximum contract length based on age of the player.

* The Gini calculator I used is here.

** I won't claim to have read every comment in our threads or others about this topic so sorry if others have suggested these or other criteria.

Varlamov should take every shootout

six hole — Thu, 12 Nov 2009 17:52:08 -0000

This isn't going to be analysis heavy - there simply isn't enough data to prove who's better. But I'd guess that putting Varly in for every shootout should net us 3-4 points in the standings the rest of the way. The only real downside I see is the risk of injury, since going in cold is the easiest way to pull a groin, and possibly some effect on JT60's confidence. The benefits are, well, more points in the standings, and maybe SV40 confidence. My only question is whether to put Varlamov in at the start of overtime, so as to get him warmer before the shootout, or trust that the 10 minutes before the shooting begins is enough time for him. I think there's even a case to be made that he'd be better than JT60 in overtime (less risk of screens and deflections, greater likelyhood of odd-man rushes) but that's for another time.

Fact: Caps shooters are below average in the shootout. In the last few years it's been a pretty stable group of shooters and what did we get:

'06-'07 - 12.5% vs league avg of 32.76% - (heavy on Ovechkin and Semin)
'07-'08 - 25.6% vs league avg of 32.51% - (Kozlov helped)
'08-'09 - 28.6% vs league avg of 33.71% - (Semin decent, others meh)
'09-'10 - in early returns it's 23.5% vs 31.07% - both figures are brought down by last night's game

Mostly fact: Jose Theodore isn't as good as he used to be, in the shootout or on breakaways.

'05-'06 - .833 save% for MTL (5/6) in shootouts
'06-'07 - 1.00 save% for COL (6/6)
'07-'08 - .786 save% for COL (22/28), 6-1 record
'08-'09 - .727 save% for WSH (16/22), 4-4 record thanks to below average shooters on team

As of last year he was still above the league average. I think most of us would feel that by the end of last year, in the playoffs, and even into preseason that JT60 was no longer above average on breakaways. On numerous occasions we needed a big stop on a breakaway and didn't get it. Most of us have also been pleasantly surprised by his play this year, and somewhat disappointed by SV40's. Of course by now their GAA and SV% are very similar. I'm not going to read too much into his 0/2 in shootouts this year, I'll just state that he's at league average for a team that needs an above average goalie.

Conjecture: Semyon Varlamov is one of the league's best on breakaways.

Obviously he's described as "very athletic and explosive" in every scouting report, and our own eyes can tell us his movement is incredible and reflexes quite good. He does make mental mistakes and he does give up goals in bunches sometimes. But my eyes also say he's well above average at making the big save on the breakaway, with plenty in last year's playoffs. I only remember him conceding one this year - 10/29/09 Todd White ATL - but I could be off and the definition may be subjective. It's too early to use his 13/14 performance in this year's shootout as proof, but it seems to fit the narrative. As a young athletic goalie if you only give him this one thing to concentrate on, with no other distractions, he's one of the best. The risk of goals in bunches or deteriorating confidence means nothing here because if he's given up two we've probably already lost the shootout. The risk of continued failure is irrelevant.

Fact: The Capitals need to pick up points in shootouts. The Caps are right with the best in the league and getting Game 7s at home is important.

But we've consistently left points on the table. Last year it was with a 4-5 record in shootouts, '07-'08 was 4-6. We're 2-1 so far, we haven't lost by more than a goal all year, and it looks like we're a good bet to keep going into overtime at a good rate. Assume we're in nine more shootouts this year. With JT60 I think we're 3-6, with SV40 I think we're 7-2. That's four massive points in the standings - could be less, could be more.

Let's look at our competition. MAF and Johnny have staked PIT to a 4-0 record behind 13/14 saves. Their shooting is a nice 6 for 13 too. NJD and MTL are 3-0 thanks to 8/9 and 7/7 respectively, and great shooting. These teams are getting it done. PIT and NJD are our main competition at the top of the east right now, unsurprisingly.

Advice: Get Varlamov into every shootout. Somehow. Just do it.

Yes, the Caps were particularly bad at secondary scoring last year

six hole — Tue, 27 Oct 2009 22:06:43 -0000

I was going to put this in CapsChick's earlier thread but it was just too long and different. I fiddled around with some numbers from '08-'09 just to see if there would be something notable in the numbers for the top tier teams. Basically similar to GINI coefficient (measure of inequality) to see how balanced the scoring is for the top nine teams, with 99 pts or more (calling that the Caps peer group). This is a bit different than secondary scoring (and my headline is misleading) but may point us in the right direction.

If someone can recommend a good way to post an excel spreadsheet anonymously I'm happy to share, not that there's anything too special in the numbers.

First I took all scorers with more than 10 pts for each team, ranked them by points scored, then normalized. So perfection would be everyone on the team scoring exactly the same (something above 10) no matter how many total points the players had. The Caps look by far the most unbalanced compared to the others. The unit numbers don't mean much but it went WAS (.215), PIT (.197), PHI (.189), NJD (.186), BOS (.178), DET (.165), VAN (.155), SAN (.153), and CHI (.147).

10 pts is probably low but it gets you towards a full measure of all "scorers" on the team while getting rid of most everyone who only played a few games.. I considered doing another cutoff at 20 or 30 pts but instead decided upon top 10 scorers for each team. By this approach the Pens are even slightly more unbalanced than the Caps (.150 vs .146), the Hawks, Sharks, Wings, and Bruins look very balanced (.067 to .079), and the Canucks, Devils, and Flyers are in the middle (.120 to .127).

This is partly just putting numbers onto an obvious result you can "see" by looking at the scoring lists.
And one number doesn't tell you everything - the Pens skew at the top thanks to Malkin and Crosby, while the Bruins had a ton of people getting 10+ pts. Problems include not accounting for games played, trades during the season, not accounting for lines/position in each game, getting an exact definition of secondary scoring, and the normalizing which gives the Wings 774 pts the same weight as the Devils 625. This is all psuedo-scientific at this point.

But while I do think there is a "there" worth looking at, I also think looking at the '09-'10 season so far (9-12 games) might be premature. Unless you're willing to rigorously define secondary scoring for each team, break apart line combinations, and do hard work for each game. In which case if that's what you're after do it now, instead of for the whole season.

Marcus Johansson, down on the gård

six hole — Tue, 27 Oct 2009 01:12:07 -0000

Marcus Johansson, down on the gård

After 12GP he's 6-4-10, high scorer for FÄRJESTAD BK, and good for 17th in the league so far. And he's known as "Mackan."

Mackan goal rescued a point

One of the youngest on the team is also one of the best. Marcus Johansson's acknowledgment to 1-1 against Lulea rescued a point and now he also leads the team's internal points.

http://translate.google.com/translate_t#
http://stats.swehockey.se/