Be Less Wrong

 

 

Cam Charron pointed out an interesting article by Sabrematrician Phil Birnbaum yesterday entitled Eliminating Stupidity is easier than creating brilliance; a statement I think is pertinent to the state of hockey analysis in general and to the Flames moving forward in particular.

I’ve thought for awhile now that the term "advanced stats" is a bit of a misnomer given how relatively simple most corsi-based math really is: adding and subtracting, means and modes. While we can test the relevancy of these metrics with correlation coefficents and such, no one needs a graphing calculator or a stats degree to calculate or understand any of the basic, fundamental stuff.

I think there have been major strides made in terms of the math since I started writing about hockey in 2006, but I also understand there is a great deal that remains poorly understood. We’re not necessarily approaching a paradigmatic shift where a collection of stats will empower us to conculsively model each season (the future is always subject to a bit of chaos), so instead of fumbling towards perfection, the true goal of stuff like this is just to be less and less wrong over time. Or, in the spirit of Birnbaum’s article, less stupid.

To that end, hockey analysis has made a few major discoveries over the last few years that can help decision makers avoid stepping on landmines.

1.) Possession/outshooting correlates with outscoring over the long-term.

2.) Percentages tend to vary around the mean wildly, meaning variance has undue influence over save rates and shooting percentages over relatively small samples.

There are details, clauses and caveats attached to those two main issues, but overall one could conceivably avoid many of the major missteps we see NHL GM’s frequently make in their player acquisition with those principles in mind. If every NHL executive understood and respected regression to the mean, for instance, Im guessing it would save NHL owners many millions of dollars every summer.

There are a few other pretty simple principles that have arisen from modern analysis: don’t take a goalie in the first round, don’t have faith in "clutch" (or not-clutch) performances in small sample sizes, stop assuming toughness absent ability has any value on the ice, and respect the fact that performance peaks in the mid-20’s and almost always drops off precipitously after 35. Also, understand how hard goalies are to predict and how fungible puckstopping is as a commodity on the market (outside of the few established, elite options).

What rules of player acquisition do we get out of the above?

1.) Collect as many guys who can drive the puck into the offensive zone (and keep it there) as possible.

2.) Be suspicious of any output or performance that resulted from a spike in percentages.

3.) Avoid goons and ruffians who can’t play at the NHL level.

4.) Stack your roster with players in the 23-28 age range as possible. Conversely, limit risk in the 35+ age range.

5.) Concentrate on skaters (especially forwards) early in the draft.

6.) Goalies are voodoo. Limit risk by having organizational redundancy and buying cheap goaltending unless you can sign, say, Henrik Lundqvist.

That’s about it. You can color inside the lines a bit depending on circumstances and market fluctuations, but those are the broad strokes.

Which brings us to the Calgary Flames. The organization has spent a lot of time stepping on landmines (some detailed here) since about 2008 and the cumulative effect has been to blunder headlong into a rebuild, engaged in out of necessity when the iceberg proved impossible to avoid. The team may be entering the 2013 off-season with a gaggle of first round picks and scads of cap room, but they have limited opportunity to meaningfully improve the club in the short-term. Therefore, the goal for the org should simply be to keep its powder dry (no clever flood reference pun intended). 

The Flames don’t have to do anything especially clever this summer since they are a poor bet to win anything next season anyways. Instead, the primary objective should be to not be dumb – to avoid any obvious mistakes like signing Bryan Bickell for $4.5M a year or picking goalie Zach Fucale at 22. Building a great roster is about hitting homeruns here and there, but also about taking the walk once in awhile. A lot of teams with money and failing fortunes waste their time going white whale hunting in the off-season which ends up being a waste of time at best.

Consider, for example, the Oilers ill-fated quests for Danny Heatley, Marian Hossa and Tomas Vanek a few years ago as they were bottoming out. During the same time frame, they allowed Curtis Glencross and Kyle Brodziak to walk while retaining Zack Stortini. They were too centered on making a big splash than simply not being dumb.

My hope for the Flames in the immediate future is that the management group simply gets the basic stuff right. Sign Brodie and Backlund to good deals, take high-end skaters with their first round picks. Don’t chase after the flavor of the month during free agency, target players who advance play and won’t cost an arm and a leg.

This way, Calgary can improve their prospect base considerably, avoid tying themselves to any unnecessarily long or expensive new deals and position themselves opportunistically in the market with their cap space and big budget.