According to the general manager Edmonton Oilers, Taylor Hall is coming along nicely as a hockey player. He is making good decisions and playing better than ever this season.

Maybe he’s right. But the data emphatically points to something having gone sideways in the game of the Oilers’ best offensive player.

David Staples, in his fantastic interview with Craig MacTavish last week, mentioned to the G.M. that most of the Oilers’ young stars had seen their play slide under Dallas Eakins. MacTavish disagreed, and specifically noted Hall’s play.

“I don’t think it’s accurate in what I’m seeing,” said MacTavish. “When you talk to people that know hockey very well they will tell you that they are seeing structural changes in Taylor Hall. His game is changing.”

MacTavish specifically noted Hall’s play on a goal against Tampa Bay on January 5:

MacTavish described the play as “living to fight another day” – Hall opting for the dump-in rather than trying to beat two players and turning the puck over and favourably compared that style of play to the one employed by Anaheim’s top line.

Here’s the thing: if Hall’s game has improved, over time we should see that in the results. So, for example, if he’s avoiding the bad high-risk plays he did in years past we should (all else being equal) expect to see that reflected in numbers like his shots against totals. If his two-way play overall is improved, the Oilers should be out-chancing the opposition more frequently than they have in prior years.

But that isn’t what’s happening.

The chart above looks complicated, but it’s really very simple. What we’re looking at is Oilers shot totals with Hall on and off the ice over the last two seasons.

Last year, Edmonton out-shot their opposition by six shots an hour with Hall on the ice. They were out-shot by nearly 10 shots per hour with Hall off the ice; in other words, the Oilers shot differential improved by 16 (a massive, massive number) when Hall stepped on the ice.

This year, it’s a disaster. The Oilers have actually improved significantly with Hall off the ice – instead of getting out-shot by 10 shots per hour, they’re only being out-shot by half that (and yes, that’s pathetic, but it still represents a pretty big improvement). The problem is that suddenly the Oilers are terrible with Hall on the ice. Not only are they allowing 2.5 shots more per hour than they did a year ago, but they’ve gone from generating 36 shots per hour with Hall on the ice down to 27. That’s a massive drop, a 25 percent reduction in shots for.

Well, maybe Hall’s playing tougher minutes. Except that he isn’t; we can go back and look at who Hall is playing and to quote from that linked Tyler Dellow post, “At home, where Eakins has more control over the matchups, Hall seems to be getting matchups as easy or easier than last year.”

Well, is Eakins is starting him less in the offensive zone? Nope; according to Behind the Net 55 percent of Hall’s non-neutral zone shifts were in the offensive zone last season. This year that number is 57 percent.

Hall’s still scoring, but if the shot numbers don’t improve that isn’t going to last.

The chart above shows two statistics. The first is on-ice shooting percentage, and there’s a pretty clear pattern. With Hall on the ice in three of the last four years, the Oilers have done a slightly better job than the NHL average of finishing on their shots, scoring on between 9.0 and 9.5 percent of all shots taken. This year, the total is 11.3 percent.

That means either the Oilers’ top line has evolved into the league’s best finishing line… or it’s riding a streak of goals going in that isn’t going to last.

The second number is “individual points percentage,” which is a fancy way of saying the percentage of goals a player is on the ice for that he gets points on. So a player who is on the ice for 10 goals and picks up nine points, he would have an IPP of 90.0 percent. What we’ve found looking over years of data is that some years a player picks up points on a weirdly high number of the goals scored when he’s on the ice (generally coinciding with a career year) and sometimes doesn’t get those points (generally coinciding with a terrible points year). Over time, though, those totals average out.

Sidney Crosby is the best player in the league at this (big surprise, he’s driving the offence on his line) and generally averages out in the 85 percent range. Taylor Hall topped 90 percent last year and this season is over 100 percent – somehow he has 29 five-on-five points despite only being on the ice for 27 five-on-five goals.

All of this is a long way of saying that Taylor Hall’s great point totals are not to be trusted. The Oilers aren’t generating shots when he’s on the ice. He’s getting away with it because Edmonton is scoring on a high percentage of their shots when he’s on the ice and because he’s picking up points on all of those goals and some other ones besides. History suggests, rather strongly, that these things will not continue.

So we find ourselves in a situation where the data we have directly contradicts the statements of the general manager. MacTavish sees Hall’s play as progressing nicely; the data says he’s fallen off a cliff.

Neither of these is to be dismissed lightly.

MacTavish was an excellent coach in Edmonton, something which was dramatically emphasized by what happened to the team when he was replaced by Pat Quinn. He knows much, much more about how players should execute a system than I do, and much, much more than the comments section here knows too.

On the other hand, I’ve lost track of the number of players I’ve seen riding a percentage bubble who have seen their numbers fall dramatically in the years that followed. Jordan Eberle is probably the best recent Oilers example because people will remember the debate following his breakthrough 2011-12 campaign (76 points in 78 games). In the 96 games since he has a total of 74 points, a 63-point pace over 82 games. It’s a fall that was predicted for the same reasons I’ve cited above with Hall.

I hate to disagree with MacTavish, but the data has been awfully reliable in the past and there’s really only one way to interpret it. Something has gone sideways with Hall’s game this season.