How Valeri Nichushkin Became 2020's Biggest (and Most Controversial) Analytical Darling

The Avalanche's reclamation project perfectly combines skill and tenacity.

In February, the twin hockey statistical modellers known as “EvolvingWild” responded to speculation about Leon Draisaitl’s Hart Trophy candidacy with a list of players they considered more worthy of consideration based on the output of their Wins Above Replacement model. Most of their alternatives were uncontroversial (like Nathan MacKinnon and Artemi Panarin), but the inclusion of Colorado Avalanche bottom-sixer Valeri Nichushkin was… provocative to say the least. Suggesting that a 1 year, $850K project signing was performing better than the league’s leading scorer prompted a backlash that still lingers, and to a subset of online hockey fans Nichushkin has been a symbol of everything hubristic and incomprehensible about hockey analytics ever since. Months later, his analytical profile was responsible for his 8th-place finish in Selke voting this year, above more name-value players like Aleksander Barkov and Brayden Point, prompting yet more ferocious arguments between those who swear to god that he’s actually that good and those who can’t believe what they’re reading.

Amid this culture war, there’s one thing nobody bothered to do: actually figure out how Nichushkin managed to put up such excellent results and why he graded out as one of the best defensive players in the league this year by multiple statistical models. When most people think of Nichushkin, they think of a bust who bolted for Russia, came crawling back, blew his shot with the Stars, got bought out, and ended up on the Avalanche’s fourth line on a “show me” deal. With no gaudy point totals to signify come-back or break-out status, and the analytical models exuberant to the point of absurdity, he either flew under the radar or became a derisive meme.

Like I did with another analytical favourite, Zach Aston-Reese, I’m going to do two things in this piece: explain why Nichushkin was so celebrated by these statistical models and use the eye test to break down how he got those results. In the process I’ll not only become the first person ever to watch an Avalanche game specifically to watch Valeri Nichushkin play, but hopefully bridge the chasm a little bit and allow us to unite in celebration of the really nice season he just had.

Let’s Have It: The Analytical Case for Nichushkin

The outputs of EvolvingWild’s model, pictured above, rank Nichushkin near or at the top of the NHL defensively in each of the past two seasons, with a massive boost in offensive performance this year. This model measures offence and defence not in terms of traditional stats like point totals or hits, but in terms of how a player impacted his team’s ability to score and prevent scoring chances against when on the ice. Why are these models so fond of Nichushkin’s performance this season? It’s because his on-ice stats were magnificent. The image below shows his rank among the 329 forwards who played 500 or more 5v5 minutes in 2019-20 in each of the most important on-ice stats including goals, expected goals, and shots. I’ve included his rank in the “raw” (unadjusted) stats, the relative stats (compared to his team when he’s off the ice), and stats isolated for teammates, competition, zone starts, etc.

It’s no wonder that statistical models that use these metrics as inputs would be incredibly high on Nichushkin. When you’re on the ice for more than twice the number of goals for than goals against, that’s a pretty good sign that you’re doing something right; when it’s backed up by the underlying numbers as well it’s even better.

Two caveats have to be made with Nichushkin - the ice time and the competition. He joined a deep and talented Avalanche team as a project signing, and spent most of the year in the middle six. His low TOI is why all of his stats are either rates or per 60 minutes. Unlike his performance, his ice time is outside of his control, and he shouldn’t be penalized for it. But on the other hand, it does ring a little weird to say that a bottom six forward had an elite season or deserves a major award. The competition isn’t as much of an issue for me not only because the statistical models specifically adjust for it, but because it’s more than compensated for by his low quality of teammates; he played against top lines a little less than average, but he almost never played with top line talent.

So how did those results come to be? Microstats, tracked manually from watching the games by Corey Sznajder, tell us a big part of that story. (Percentages indicate a player’s percentile rank in that stat)

This indicates why Nichushkin was really able to take off with the Avs. In 2018-19, the Stars entered the zone with possession of the puck a below-average 58% of the time. As a result, while Nichushkin was extremely active on the breakout, he didn’t enter the offensive zone with the puck that often and frequently dumped it in. That’s not how Colorado plays the game. They led the league this year in possession entries by a large margin, carrying or passing the puck into the zone almost 75% of the time. Empowered to hold on to the puck, Nichushkin’s skills were finally fully utilized and he finished behind only Nathan MacKinnon and Nazem Kadri in terms of entries per 60 minutes on the Avs. When Nichushkin wasn’t the one to lead the entry and the puck got dumped in, he was still a huge contributor; he finished only behind Gabriel Landeskog and Matt Calvert in terms of puck recoveries off dump-ins, and only behind Matt Nieto in pressuring defencemen on the forecheck. He also ranked 59th in the league in takeaways per 60 minutes and 24th in takeaway/giveaway differential per 60. In other words, he was a puck hound all over the ice.

Based on the macro-level results and microstats, I think it’s likely that Dallas was not the right place for him. His forechecking and - for lack of a better word - time-wasting ability led to very nice defensive results, but in playing a low-event dump-and-chase style with stalwarts like Radek Faksa his offensive skill was not being fully taken advantage of. It didn’t help that he had horrible puck luck, scoring zero goals on a still pretty meager 5.5 expected goals. Colorado was the perfect spot for him, and the results reflect the benefit of a change in style. But let’s back that up with the good old eye test.

The Eye Test

When a no-name player has excellent statistical results, there’s often an assumption that his excellence “in the spreadsheets” is unreflected by how he looks on the ice. That couldn’t be further from the truth with Nichushkin. Unlike most bottom-six players with excellent defensive results (your Aston-Reeses, Folignos, and Riley Nashes), he’s a speedy, skilled player who is a lot of fun to watch. He’s tenacious on the forecheck, comfortable carrying the puck, and when he scores it’s often a highlight-reel effort. He honestly looks like the player described by this 2013 pre-draft scouting report:

Nichushkin is a tremendously gifted, big forward who has the skills and finesse of a Russian skater, but the grit, forecheck, and physicality of his North American counterparts.

Watch this goal against the Blackhawks (and enjoy the awkward call from the Chicago commentator):

The Hawks win the draw but Nichushkin immediately bolts to the puck, blows past both defencemen, and finishes on the backhand. In two games that I watched of his, he repeated that “if you lose the draw, book it to the point at full speed” play many times, which often caught defenders off balance and led to a hasty and ill-advised plays on their part.

I compiled a set of snapshots of Nichushkin’s performance in a game against the Toronto Maple Leafs from December (powered by InStat Hockey):

As you can tell from these highlights, the core of Nichushkin’s game at both ends of the ice is pressure. When he’s out there, no inch of ice is guaranteed - he’s going to use his speed and reach to terrorize opposing defenders and press the attack. He’s not an overly ambitious passer, but he has the hockey IQ to make the high percentage pass to keep the cycle alive and maintain possession. Overall, he plays like a high-end talent determined to make a grinding third line better rather than playing like he’s on a top line; in the entire game I never saw him make a stupid play or try to do it all himself. His line spent the majority of the game cycling the puck in the Toronto end, although he did get some chances to impress defensively with a well-timed stick check deep in the slot and some nice hands to deal with an attempted breakout pass in his skates.

Conclusion

One of the fundamental divides between most hockey fans and the analytics community is how they conceptualize defence. For the former, defence is something that takes place in your end of the ice and consists of a set of active plays that you do (steals/hits/blocks/etc.), while for the latter, defence is anything that prevents your goalie from having to face quality scoring chances. A forecheck in the offensive zone, a smart stick to keep the play alive, the right amount of pressure on a defender trying to break out the puck - that’s all part of the overarching goal of defence. So when a model declares Nichushkin the most impactful defensive forward in the league that doesn’t mean he’s lifting sticks like Mark Stone or making perfect pokechecks like Patrice Bergeron. It means that - regardless of how he does it - he’s keeping the puck far away from his net.

To my eyes, Nichushkin is a player with a lot of skill who’s smart enough to know exactly how to be extremely effective on a bottom six line. He uses his puck skills to enter the zone with possession, his speed to forecheck, his size to protect the puck in the zone, and his hockey IQ to get to the right spots and keep the cycle alive. Like some other forwards with strong defensive metrics, he plays a fairly low-risk style, but the difference is that he has the talent to generate offence despite that. The question is whether this would translate in a top line role. Nichushkin was so good in his role this season that his WAR totals ranked among stars who played much more than he did, but that doesn’t guarantee that those results would scale up or even remain steady if given 20 minutes a night. As curious as I would be to see him get a real shot, Colorado’s middle six seems like a match made in heaven for him. Whether or not he’s on your imaginary MVP and Selke ballot, he deserves a lot of credit for the season he just had.