Why The Hamonic News Might Be A Blessing In Disguise

The biggest news for Oilers pundits yesterday, was the report that Travis Hamonic, the 25 year-old Islanders defenseman, and saviour to all of life's problems, has rescinded his trade request to Islanders management.

I may be in the minority on this, but I think the Hamonic trade proposals had been getting a little over-zealous.

Don't get me wrong, I think Hamonic is a fine player. And his $3.8 million-dollar cap hit makes him a great bargain too.

But Hamonic isn't a free-agent. The Oilers would need to ship significant assets to Brooklyn in order to get Hamonic as a return.

At this point, anyone who is familiar with what I've written on this site is probably saying: "Corey, you're about to write about how great you think Eberle is again, aren't you?".

That's right, person who's read all of my posts (AKA mom), I am.

The most common speculative trade proposal that I've seen has Eberle headed to the Big Apple, and Travis Hamonic headed to Oil Country.

I've written about Hamonic before. I still believe everything I said in that piece is true. I like the player, like the contract, but if the price tag is going to be Eberle or more, he's not worth it. Here's why:

Shot and Expected Goal Impact:

I used some Corsi and Expected Goal models to get an idea of where each player ranks, compared to other players across the league. I only compared Eberle to other forwards and I only compared Hamonic to other defencemen.

Here's a quick rundown of the statistics I used to evaluate these players.

CF%RelTM: the derivative of CF%-CF%QoT. This represents how much better a player was than the players he shared the ice with, weighted by how much ice-time he shared with each player.

xGF%RelTM: The same idea as the previous stat, but based on expected goals. This gives a higher weight to shots from high-danger locations.

Adj.CF%: CF%RelTM+CF%QoC. This stat gives a player who faces tough competition credit for who they play against. Since CF%RelTM is centered around zero and the average CF%QoC is 50%, this stat provides an estimate of what a player's CF% would be if he faced league-average competition, while also playing alongside league average talent.

Adj.xGF%: The same concept as the previous stat, but based on xG stats, rather than Corsi.

Mod.CF%: Short for Modified CF%- (((CF%RelTM+50)*0.8)+(CF%*0.2))+(CF%QoC-50). This stat is similar to Adj.CF%, but it's 80% based on outperforming teammates and 20% based on raw on-ice percentages. This might be helpful for comparing a player from a tremendous possession team to a player from an awful possession team. The reason for this stat is the notion that it's easier to outperform bad teammates than it is to outperform good ones. It's worth noting that the 80/20 split is not based on any scientific findings. It's an estimate of how a player would influence possession, without the presence of contextual factors. But it's only an estimate, please keep that in mind.

Mod.xGF%: same as Mod.CF%, but with xG instead of Corsi.

• All stats are from Corsica.hockey, and have been adjusted for zone, score and venue.
• Players needed 500 five-on-five minutes to qualify.

Hamonic

	CF%RelTM	xGF%RelTM	adj.CF%	adj.xGF%	mod.CF%	mod.xGF%
Hamonic 15-16	1.54	1.91	51.44	51.93	51.262	51.54
Rank (out of 197)	61	58	62	59	64	63

What we see with Hamonic is a player who's either a low-end top pairing or high-end second pairing defenseman. The fact that his rank doesn't change much from stat to stat means that his quality of teammates and competition is fairly standard. I'd lean towards calling him a low-end top pairing defenseman, because he managed to put up these number while playing top minutes, while a few players ahead of him have been extremely effective in their ice-time, but haven't had to carry the burden of playing 25 minutes a night.

Eberle

	CF%RelTM	xGF%RelTM	adj.CF%	adj.xGF%	mod.CF%	mod.xGF%
Eberle 15-16	3.14	4.38	53.45	54.61	53.044	54.248
Rank (out of 350)	66	49	54	42	63	47

In Eberle we see a player who is undoubtedly a first-liner quality player. Sure, he plays with good players, but he deserves to, he's a fantastic player in his own right. His relationship with McDavid last year, and Hall before that, was mutually beneficial.

But what about his defensive play? Kelly Hrudy says it's bad, is it?

The answer to this, along with many other questions in life, is kind of. His shot suppression is pretty average. According to stats.hockeyanalysis.com, his CA60RelTM was -0.80 last season. Since the start of 2012, it's -0.22, meaning that he's 0.22 shot-attempts worse than his most common linemates are defensively per 60 minutes. But while his defense is slightly worse than average, his offence still makes up for it, and more. His CF60RelTM was +7.22 last year, and it's +5.26 since the start of 2012-13.

To be fair to Eberle's critics, his defensive play starts to look worse when we factor in shot quality. Below are Eberle's relative stats, adjusted for zone, score and venue, from Corsica.hockey.

	REL. CF60	REL. CA60	REL. xGF60	REL. xGA60	REL. GF60	REL. GA60	REL. CF%	REL. xGF%	REL. GF%
Eberle 15-16	7.46	0.06	0.71	0.22	0.89	0.59	3.19	4.52	3.85
Eberle 12-16	7.13	-0.21	0.6	0.24	1.07	0.47	3.34	3.91	7.31

The main takeaway for me here, is that although Eberle's defense seems to look worse the more shot-quality is factored in, the opposite is true for his offense.

The great thing about this is that, even though on-ice goals against stats aren't terribly predictive, due to a heavy reliance on goaltending, on-ice goals for stats do a decent job at telling us about a player's offensive ability.

I've written before about how Eberle is very selective in his shot choices. He much prefers to shot from in tight than from the perimeter.

That explains why his offensive xG stats are so impressive. But you'll notice that his goal-based on-ice stats are even more gaudy. That's no fluke.

Since the start of 2012-13 (not including his 34 goal season), Eberle's individual expected goal (iXG) total at five-on-five is 52.33, but his actual five-on-five goal total is 63. If we include all situations, Eberle has 77.25 iXG and 93 actual goals in 278 games played. That's an 82-game pace of 27.4 goals.

The point of all of this is that possession stats like Corsi percentages, or even shot-location based stats, like xG% underrate a player with Eberle's elite finishing ability. Yet models based on those stats still see Eberle as a definite first-line talent.

The Oilers are likely to lose any trade in which Jordan Eberle is involved. Obviously there are still many players better than him in the league, but they are rarely made available. Travis Hamonic was not one of them, so perhaps the fact that he no longer appears destined for Edmonton is for the best.