By the Numbers: Jacob Trouba is developing into an elite defender

This last summer, Jacob Trouba supposedly requested a trade from the Winnipeg Jets due to feeling that given the chance he could play as a bonafide “number one” defender in the NHL.

Since his return, Trouba dominated the NHL and it shows in the numbers.

Judging where a player performs well statistically is not that hard. One can find which areas a player does well and where they do not fairly easily. Do they score well? Does the team out shoot well with them on the ice? Etcetera.

Advertisement - Continue Reading Below

Appropriately combining how all these factors interact can be a lot more difficult. This is what DTMAH’s WAR model attempts to do. It combines the many statistics a player has and estimates their impact on the game.

Trouba has trended well in WAR since his entry into the league. Per minute of play, he placed 75th in the league for defenders in his rookie season, then was 69th and 37th the following years. He had grown from an elite second pairing defender to an average top-pairing defender right before our eyes.

What about this season?

Unfortunately we do not have access to this year’s WAR. We can look at each component, however, and make some estimates from that information.

First we can look at good ol’ fashion shot metrics for 2015-16 to present:

Jets Defenders

Advertisement - Continue Reading Below

This simply looks at how much better a team does controlling events with the player on the ice versus on the bench. For example, if a team controls 52 per cent of all shots with a player on the ice and 50 per cent with him on the bench, then the player is scored as a +2. 

What these numbers suggest is that the Jets do better in all five types of shot metrics with Trouba on the ice versus when he’s on the bench. Now, we do have to account for the fact that 75 per cent of Trouba’s games had Dustin Byfuglien, a bonafide top-pairing defender, as one of the defenders on the ice while Trouba was on the bench. This makes it harder for Trouba to be a plus-player in relative metrics.

Trouba’s “all shots” number suggests his impact on shot quantity is positive but smaller than perhaps his impact on shot quality that we see with the “expected goals” number. Expected goals are simply unblocked shots with shot location and a few other adjustments added.

Now this does not account for usage, which we can use DTMAH’s XPM (Expected Plus-Minus) for:

Advertisement - Continue Reading Below


Numbers are from DTMAH’s database. Note for graphic: Zero is not league average but replacement level.

Advertisement - Continue Reading Below

XPM is a more advanced expected goal model than the Corsica version above. XPM’s expected goals adjust for more quality factors than Corsica’s version, like regressed shooter history (Stamkos historically finishes more shots than anyone else given all other factors) and strength state (5v5 versus 5v4). XPM then accounts for a player’s usage (linemates, linematching, coaches on both teams, schedule, zone starts, etc.) at the play-by-play level.

Ultimately this makes XPM very powerful, like a super Corsi. It does not account for all of a player’s impact in the game but a major portion. For example, XPM accounts for 54 per cent of a defender’s offensive impact on DTMAH’s WAR and 100 per cent of their defensive impact.

Not only has Trouba been the Jets’ best defender in XPM, but per minute he ranks as the 24th best defender in the entire NHL for this season in impacting shot quantity and shot quality differentials.

As we noted, XPM does not account for all of WAR. The other component is BPM (Boxscore Plus-Minus), which combines many of the individual factors a player inputs into the game:


Points/60 is point production per sixty minutes. TOI.QoT is the average ice time of the player’s linemates. Giveaways and takeaways /60 are just those two statistics relative to ice time. iFenwick is simply a player’s individual non-blocked shots, with these two components being their expected shooting percentage given shot quality factors and their shot production volume per sixty minutes.

I have listed the above stats in the order of their relative importance in impacting BPM. For example, 5v5 points per sixty minutes is actually the combination of goals, primary assists, and secondary assists per sixty minutes. Combined these components account for over 60 per cent of BPM. I aggregated them together since for defenders, unlike forwards, the relative importance of all three of goals, primary assists, and secondary assists are fairly similar.

The bulk of BPM’s value comes from adjusting a player’s point production relative to their quality of teammates. We see here that Trouba has been slightly above average in playing with the Jets best players, but his scoring has been relatively way better, with a top-10 5v5 point production rate.

While Trouba struggles in giveaway production, this is not unusual for players that carry the puck often. These players are typically the best on the team and so the impact of giveaways on BPM is minimal, like all the other remaining factors. Trouba has been a bonafide top-pairing defender in takeaways and shot volume generation, while being slightly above average for his shot quality generation.


We know that Trouba places 23rd overall in XPM, which is 54 per cent of WAR’s offense and 100 per cent of WAR’s defense, thus far this season. He has impacted the Jets very positively in shot volume and quality when accounting for his usage.

However, we can only estimate Trouba’s BPM impact. BPM is predominately driven by point production, which Trouba has been a top 10 performer while only receiving slightly advantageous minutes in performance. Aside from giveaways, the remaining factors range from above average to top-30. Overall we expect Trouba’s BPM to be slightly better than his already very impressive XPM.

This suggests that we could be looking at a top-20 defender year for Jacob Trouba. Looks like he was right when he said he was ready to be a number one defender.

All numbers courtesy of Corsica unless otherwise stated.


  • ps241

    Impressive article Garret. Seems like my eye test is matching the numbers. It will be interesting to see what the Jets do when Myers returns because we have spent the entire season without either Trouba or Myers in the line up.

  • MBfromMB

    Hopefully he decides to stay long term now that hes playing top min on the Right side.

    Garrett can you please look at the jets roster and do some sort of defensive zone coverage comparison? Im not good with analytics. TSN’s article about grading the Jets players got me thinking of what is the Jets real problem. I find the Jets have the offense nailed down (8th in the NHL for GF) but I want to know if the Goaltending is the biggest problem that Chevy needs to fix (I think its part but not all) I think its players not being defensively responsible but I cant tell if its a whole team structure thing or if there are certain players that just suck at playing a defensively responsible game. Thanks