Alex Ovechkin

Photo credit: Bruce Bennett

Last season I got some flak for projecting Ovechkin would score on average 42 goals, plus or minus 8 goals over the 82-game season. The former two-time Hart winner ended up setting career lows in goals (32) and points (85), while once again suffering an early exit in the postseason.

Before anyone accuses me of being right a jinx, consider that in each of the two seasons before that he saw most of his offensive numbers decline:

Season GP G A PTS EV PP S S% TOI
2008-09 79 56 54 110 36 19 528 10.6 1817
2009-10 72 50 59 109 37 13 368 13.6 1569
2010-11 79 32 53 85 25 7 367 8.7 1688

So, despite this downturn, what can we realistically expect from one of the NHL’s best players this upcoming season?

“I fully expect Alex Ovechkin to come back into [training] camp mean as a bear,” head coach Bruce Boudreau said. “I’m sure — and it’s well documented — that individually for him, he took more criticism at the end of last year for his totals in goals and points, and not being up for the Hart, and not being up for other awards.”

Personally, I am still tinkering with my #fancystats for quantifying bear-meanness; however, my recent inklings that Ovechkin may light the lamp only 30-35 times in 2011-12 might have been a little low. Consider this a revision of my expectations.

Troy Brouwer

The Capitals added to their plethora of talented wingers by trading for net-crasher Troy Brouwer during the offseason. (Photo credit: Rich Lam)

The Capitals added depth at forward and on defense by trading for Troy Brouwer plus signing free agents Joel Ward, Roman Hamrlik and Jeff Halpern. Washington also brought in veteran goaltender Tomas Vokoun on a $1.5 million, one-year bargain-basement deal.

The defense is better than it has been in recent years. The arrival of Roman Hamrlik gives them another offensive defenseman while Tomas Vokoun’s save percentage since the lockout has been a lofty .92168, second only to current Vezina winner Tim Thomas’ .92170.

Boudreau also seems to be looking for some middle ground after his system switched from “fun-n-gun” to a more defensive style.

“I’m hoping that we can be a hybrid,” explained Boudreau. “There’s some parts we changed [last season] that I really loved. But when you’re playing like that, you have score a lot of goals [on] dump-ins and you have to score a lot of goals off the forecheck because the quick-break isn’t there. I’d like to get back to being more of a quick-break team.”

The reduction in scoring last year was also caused in part by the team seeing less “puck luck” — at both even strength and on the power play — than they had in years past. The Caps did manage to generate the same amount of shots when Ovechkin was on the ice at even strength and increase the shot volume on the power play, so if the shooting percentages progress to the mean, that alone should help give him an uptick in the boxcar stats.

Usually I run projections using similar seasons (normalized to the current season) from players at the same age and position, but Ovechkin has so few peers I decided to tackle it another way, so bear with me as we crunch some numbers.

First, at 26-years-old, Ovechkin is in his prime. Goal scorers tend to peak between ages 24 and 27 and then the goals began to fade.

Second, Ovechkin has become more of a playmaker these past few years, registering points on over 83 percent of the goals scored when he was on the ice at even strength last season, compared to less than 80 percent the three seasons prior to that. He also scored just a third of the goals himself last season, a ratio which should drift back up to his historical average of 40 percent. Add in the 5v5 shooting percentage should also rebound and the Russian Bear will score 28 even strength goals and have 34 assists.

Third, the power play should bounce back. The NHL’s 16th best unit converted on less than 10 percent of its shots on goal with Ovechkin on the ice, down from 15 percent the three years prior. That should progress to around 16 percent, and with more shots being generated on net during the man advantage Ovechkin should add another 11 goals while on the power play plus 20 assists.

That’s 39 goals total without the benefit of more starts in the offensive zone, a better off-season training regimen and plain ole good fortune.

Does this mean Ovechkin has a shot at an Ovechkin-like season? It looks like he does.

He has roughly a 46 percent chance at 40 goals or more and almost an 11 percent chance to exceed 45 goals for the year.

My conservative projection for Ovechkin in 2011-12:

Season GP G A PTS EV PP S S% TOI
2011-12 78 39 54 93 28 11 328 11.9 1603
  • SeanL

    Neil, just admit that you didn’t set up your monte carlo machine properly for last seasons’ projection. Given the data present at the time, you should not have predicted a goal total lower than the lowest season total of Ovechkin’s career.

  • Peter Hassett

    What?

  • Chrisrdiachok

    BOOM ROASTED. Seriously though, shut up.

  • SeanL

    Are you seriously insinuating that your prediction was correct? Mr. stat geek, you should know how your “monte carlo machine” works by now. I cut you slack because it seemed like it was one of your first attempts last season. But by now you should have realized that such a prediction should not yield an output less than the lowest output of the input data.

  • http://twitter.com/ngreenberg Neil, RMNB

    He scored 32 goals, so of course it was accurate. A Monte Carlo Sim works by putting inputs IN. Obv, I knew the RIGHT inputs. By now you should realize that when it comes to projections, I am pretty darn good.

  • http://twitter.com/ianoland Ian Oland

    FYI: I’m the one responsible for linking the word “Russian bear” to this video. http://www.youtube.com/watch?v=A7zU2YDNHeo&feature=youtu.be I couldn’t help myself.

  • Livia

    These numbers are very encouraging. Thank goodness for regression to (Ovi’s) mean, plus solid new players, and a balance between offense and defense. Can’t wait for the season to start!

  • SeanL

    A flawed projection can still end up being “accurate” by that measure. You took Ovechkin’s healthy, “fun and gun” numbers and magically came out with his trap game, fat/hurt/lazy/whatever Backstrom numbers. You did nothing to factor in regression to the mean, or his decline in shooting %, or the team’s decline in PP%, or anything else. Yet you still had decreased stats.

    I realize you’re proud of the work you do here and that you put tons of effort in to RMNB, ESPN, and WashPo. But that doesn’t mean that every single prediction you made was done correctly. Your conclusion did not fit the data. Real world factors that you didn’t account for helped skew the real world data to match your projection.

    You’re a stat guy. You know that variance exists. You know that, while it’s unlikely, a result can fall multiple standard deviations away from the mean. That’s exactly what Ovechkin’s season did, and it happened to line up with your conclusion of a down season. You got lucky.

    What you fail to mention is that Ovechkin’s season fell outside of your standard deviation. Even you have to admit Ovechkin’s season was abnormal. I’m not sure why you’re trying to use an abnormal, and potentially (hopefully) and outlier, season as a justification for your previous analysis.

  • http://twitter.com/ngreenberg Neil, RMNB

    Haters gonna hate.

  • Peter Hassett

    Oh you’re so proud

  • HMA

    Neil,

    Nice work as always.

    What would be interesting is if you ran your simulator on previous years and see where they compared to the actual results.  (2nd season based on rookie season, 3rd season based on first two…etc.)

    What were the odds that Ovi would score 65 goals in ’08?

  • JCinNL

    I think you’re both right to an extent.  I give you credit for predicting a notably sub-par season when few people were.  But a 17-goal range for his goal production was given and Ovechkin’s output still fell outside that margin.  And an ‘over’ on an o/u of 37 was recommended – also a miss.  So regardless of how the statistical analysis was done, it didn’t end up being ‘right’ in terms of forecasting his eventual production, even within a fair-sized range.  And one should always acknowledge that correct answers can result from flawed analyses, regardless of whether that actually happened in this case (it might be helpful to make clear what other inputs went into the sim, if there were any beyond SOG and S%).
    My statistic-free, ‘gut’ prediction for the 2011-12 season: 46 goals, +/- 4.  I might also try my hand at a proper sim before the summer is out, as you once suggested others should do.

    Thanks for the post though – I am enjoying the site.

  • Taylor

    Interesting…. But a 0.8% chance he breaks 50 goals? Your predictor doesn’t pass the logic test.

    All that it would take would be a 360 SOG season with a 14% shooting percentage. Which is well within the realm of “somewhat likely.”