thanks gabe. the charts really drive home the point.

He made, what, $27MM in Chicago over four years? And then $3.75MM this year with Edmonton? Not that it matters a lot but you seem to have him making too much money.

What’s a win worth in terms of dollars in the NHL?

It seems to me the correlation is ruined by a handful of teams screwing up with the six contracts in the bottom right. Other than that there seems to be a good trend of paying for performance. You could chalk up that up to a half dozen ‘dumb’ management groups instead of a general inability to value goaltenders.

Instead of just a graph, is there a table with all the data points somewhere?

I did not need to see this chart to conclude that some GMs are stupid and over pay players based on one year of outstanding play (happens all the time and isn’t limited to goalies) but I am not sure it is to the extent you show here. There are flaws in your analysis with the most significant being that you don’t account anywhere for who that goalie is playing for. The same goalie would have a lower save percentage on a weak defensive team than a great offensive team and unless you factor that sort of thing into your chart it isn’t telling a complete story.

Elsewhere on this blog you have a post about how coaching style influences shot difficulty and how Jacques Lemaire coaching style greatly limits shot difficulty against. Interestingly, Niklas Backstrom’s save percentages in his 4 year career are .929, .920, .923, and .903. The first three of those were under Lemaire, the fourth was for Todd Richards. Coincidence or no? Did Lemaire’s system make Backstrom look better than he actually was or was this year was just a bad year for Backstrom? How many other goalies are good goalies playing behind bad defensive teams/systems? How many other goalies are mediocre goalies playing under great defensive systems? All these sorts of things need to be considered.

Do many GMs not fully take this kind of thought process into account when evaluating goalies (and players) and thus make mistakes? Sure, happens all the time. Did Minnesota made a mistake signing Backstrom to a big contract? Maybe, but maybe not because he only had one good season, but because they didn’t take into account the Lemaire effect.

After taking a second look at the salary vs save% chart I think your conclusions aren’t quite so obvious.

-There are 7 goalies with a save percentage under .910, only one of them made more than $10M.

-Of the bottom 14 goalies in save percentage, 12 of them made less than $10M.

-Of the 17 goalies who made more than $10M, 11 had a wins above replacement greater than 5.

-Of the 26 goalies who made under $10M, 17 had a wins above replacement under 5.
That actually tells me that there aren’t a lot of huge mistakes being made by GMs.

To me there appear to be 6 goalie contract mistakes – those goalies earning >20M and have a wins above replacement <~5). Take out those 6 contracts and it appears you have a fairly decent relationship between salary and wins above replacement.

I too would be interested in seeing a table/chart with goalie names.

Elsewhere you (and others) have noted that it’s best to take 4 years worth of EV data to get a good sense of a goalie’s talent. And this is assuming that they’ve played a starting role somewhere. Here it seems to suggest that you’re best bet to not overpay are good, young goalies, at least able to get near .915 at EV. But for young goalies, you might not have that much data, or at least not have it all at NHL level. (I assume that there are inefficiencies in projecting from junior, AHL, or other leagues, but correct me if I’m wrong.)

Thus, it seems to me that the inefficiency is partially due to the fact that, as a GM, you might have to come to up with a contract decision based on less than optimal data. No?

For example, if you were the MTL GM last summer, and you thought to yourself you’d like to move one of your goalies to save money, and maybe pick up a large winger from someone. Esp. since you’re not sure you’ll want to sign both at the end of this season (both RFAs), esp. b/c you’re worried the price tag for both could go up. You’ve got Price with about 2795 NHL shots under his belt (.912, not sure EV vs SH splits), 310 AHL shots (.906) and 5349 WHL shots (.915), and an ELC of $2.2 million. Added to that, you know that Price has all kinds of good image pedigree (top 5 pick, WJC, made a good first impression). Halak, still on a two-way for 775,000, has 1697 NHL shots (.914 overall), 2009 AHL shots (.930), 516 ECHL shots (.932), 1442 QJMHL shots (.913). He’s also older by two years, and a 9th rounder, although also a WJC goalie.

Let’s say a player you want is available for one or the other: do you think you have enough info to make a call on this? If you moved Price, and picked up a Raycroft for your back up, you could spend the million+ on another position.

I’m not saying that MTL should have thought in these terms, it’s just an example of what kind of decision might be available at the time, and with what kinds of information.

I think that other inefficiencies come up because you need to manage image (i.e., this is a good team to play for, we care about our players and develop them, etc.), and sending away a top five pick when he’s only 21-22 might not be the best thing for your image, esp when the other guy isn’t totally proven in fan/media eyes.