Blind Stats Test: Matvei Michkov vs. Owen Tippett

The world’s information ecosystem is increasingly becoming siloed. Every news entity and social media platform is designed to maximize user retention and keep you in one place. Echo chambers form, at least in part, to confirm what users are looking for which fulfills the goal of keeping user’s attention.

This creates an air of distrust and typically forces retreat further from the middle. In the context of sports, it typically means that ideology drives takeaways. With information tailored for each individual user, it can be a challenge to break outside of what algorithms assume users would like to read or watch.

In a sports context, this leads to aggressive factioning. Within large fanbases, common ideologies or opinions form and cordon segments off from each other. These splits stem from priors and overarching ideologies. In the case of the Philadelphia Flyers’ ongoing Rick Tocchet conversation, the handling of Matvei Michkov is proving a durable topic.

With the Flyers still a bit of a longshot for the postseason, Michkov’s development is the third rail of discourse. The forward is playing 1:55 fewer per game on average in his sophomore effort and spending time on the left, a position he’s less familiar with.

Litigating how Tocchet uses Michkov falls on deaf ears. The Flyers head coach has repeatedly deflected when pressed on Michkov’s role and it’s led to social media unrest. Now, social media isn’t the most apt bellwether of public opinion. The 90-9-1 rule surmises that 90 percent of social media users lurk, 9 percent post occasionally and 1 percent post the majority of content.

With that said, there are certain ways to break through that noise. Blind resumes are a strong way to remove your preconceived biases. Just the raw numbers and what players have done on the ice.

Below, you’ll see a comparison of Matvei Michkov and Owen Tippett. Their numbers will be blindly attributed until the end of the article where the full statistical profile for each will be revealed. All numbers are two year averages.

Under the Hood

It would give away the store to the neurotic to lead off with raw goals and assist totals, so we’re gonna start out with underlying numbers. Diving into production is a hairy business and requires a nuanced view. The underlying metrics to follow are a means of describing specific game actions.

Individual Rates

The key stats to understanding contemporary hockey are oriented around creating scoring chances. A scoring chance is any attempt at the net including shots on goal, blocked shots, missed shots, posts and saves. While not all scoring chances are shots on goal, all shots on goal are scoring chances. Teams and players that have a majority of scoring chances in a game are generally assumed to have had control because they were able to have the puck more than their opponent.

Going a level deeper, Natural Stat Trick logs individual data of game situations which brings us to the subsets to describe our respective players.

ICF: Individual chances for, all scoring attempts by a player per 60 minutes Shots: Shots on goal per 60 minutes ixG: The expected goals value of a player’s scoring chances per 60. Based on historic shot tracking data of how likely a scoring chance is to result in a goal iHDCF: Scoring chances classified as high danger based on where they come from on the ice, these are scoring chances most likely to result in goals

Player A is a volume approach player, he averages a higher rate of individual scoring chances and shots on goal than Player B. When Player A is on the ice, he is inclined to let the puck fly and get it going on net. In comparison, Player B is a bit more selective in when he chooses to take scoring chances, but he is often in slightly better position on the ice to take them.

There are plenty of players across the league with above average shots that don’t need the highest possible quality of scoring chance to score, but the easier the attempt, the more likely a goal is to follow. In theory, a plus-shooter getting the puck to more dangerous areas should result in goals more often. In addition, those higher quality shooters often have the capacity to score on lower quality chances more often than their less skilled contemporaries.

Rate of offense

Building outward, beyond individual play, we can look at how the Flyers at large as a team play with Player A and B on the ice.

We’re expanding from individual chances outward to the entire combination of teammates now. The rates and percentages above are tied to each player and the other four players on the ice at the same time as them. As we broaden the statistical pool, the numbers are a bit noisier, the context of a line’s responsibilities and the quality of competition come into play even more than on an individual level.

In this case, Player B’s group on ice is producing at a higher rate of offense entirely across the board while getting slightly worse goaltending, which is a good trade off for the Flyers. Philadelphia sits in 23rd of goals scored in all situations and 20th of 5-on-5 goals, meaning offense is at a premium.

The Share of Offense

While creating offense at a high rate is beneficial to eventually scoring goals, there is a greater context to evaluate. It’s great to create offense frequently, but if that trade off comes at the expense of the other team having more offensive opportunities, there needs to be a recalibration.

When it comes to the share of Flyer offense when our players are on the ice, the picture is different. Here, Player A and his linemates have a higher share of the underlying offense. Yet, when it comes to actually winning the minutes, Player B’s group does a better job. This is with the duo having pretty even shooting and save percentage luck.

This is likely a product of player B’s higher rate of dangerous offense on both an individual and group basis. Since Player B’s group creates high danger scoring chances more often, they have a better chance of converting them even though Player A has a better share of them.

It’s a bit confusing, but it’s tied to rates. That’s why it’s important to scale out minutes over per 60 minute basis to account for differences in workload. Now, this can make numbers wonky for players who’ve played significantly fewer minutes.

I’ve intentionally avoided disclosing the minutes workload to this point because that would give away which player is which.

The Reveal

Player A has played 2313 minutes over the last two seasons, player B has played 2280 minutes. That 33 minute difference is about two games worth of work respectively for each player, but it does give away the results.

Player A is Owen Tippett, Player B is Matvei Michkov.

With relatively similar minute allocations, Michkov has an edge in the rate of offense he and his teammates generate. The gap between the two isn’t huge and some of it can be attributed to the per-game minutes difference between the two wingers, but it is real.

Throw in the fact that Michkov has a 13 point advantage while creating more dangerous offense and it’s not unrealistic to say the sophomore forward is a better player at the moment. Points are not the be all end all in an evaluation, but that in conjunction with the edge in rate of offense leads one to believe Michkov has more room to squeeze.

This isn\'t meant to be a dig at Tippett either. Both players are amongst the best forwards on the Flyers. If anything, Philadelphia\'s goal needs to be acquiring more quality running mates to bring the best out of both Michkov and Tippett as soon as possible.

Talking Points