A Hot Shooter Keeps About a Third of It

On Tuesday, the last day an NHL team could hand out an eight-year extension, Ottawa gave Drake Batherson eight years at $10.75 million a season, the largest contract in the franchise's history. He is coming off a career year: 33 goals and 71 points in 79 games.

He scored those 33 goals on 167 shots. That is 19.8%. In the 908 NHL shots he had taken before last season, he scored on 12.8% of them.

Whether Ottawa overpaid is a separate question. A contract buys eight years of a player, and Batherson's assists, his minutes and his role do not depend on his shooting luck. This piece is about what happens to that seven-point gap next season, because Batherson is not unusual. Thirty of the skaters being drafted this fall are carrying one just like it, and six of them go in the first twenty-five picks.

What a hot year turns into

The question is simple. When a player's shooting percentage jumps well above what he has done for his career, how much of it survives into the next season?

Read the top row. Forty-three players shot at least seven points above their career rate. The next season, forty of them scored at a lower rate, and the group lost more than a quarter of its goals per game.

Read the middle row too, because it is the control. Players who shot close to their career number did it again (9.9% then 10.1%) and were a coin flip to score more or fewer. Shooting percentage is noise around a number the player already owns, and the further a season wanders from that number, the harder it snaps back.

It works in both directions. The bottom row shot 7.1% and came back at 10.9%, and more than four in five of them scored more goals the following year.

Nothing else moves

I did not expect this part to be so clean. If a hot shooting season were a player genuinely getting better, you would expect everything around it to rise with it: more shots, more assists, a bigger role. So I looked at what else changed, per eighty-two games:

The hot shooters took the same number of shots the next year. Their assists did not budge. The entire loss is goals, and it is almost exactly the size of the goals they "found" the year before.

That is what luck looks like in a stat line. The player did not change. The puck went in, and then it went in less.

The rule: about thirty percent

Put all 2,433 seasons through one regression and the answer is tidy. A player keeps about 30% of the gap between last season's shooting percentage and his career number.

That number barely moves when you cut the data: forwards keep 30%, defencemen 25%. Players 24 and under keep 33%, players 25 to 29 keep 31%, players 30 and over keep 25%.

But age does matter to the points, and more than the retention rate suggests:

The young hot shooters lose goals like everyone else, but their shots and assists keep growing and cover most of it: they lose about two points. Thirty-year-olds lose the goals and start aging, and give back twelve and a half. A hot season from a veteran is the one to be most careful of.

The market is not discounting it

If the draft market understood any of this, the hot shooters would be going later than their raw numbers suggest and the cold ones earlier.

They are not. Across 143 ranked skaters with enough history to test, the correlation between draft position and how far last season's shooting strayed from career is −0.04, in other words nothing. The thirty hot shooters average a draft position of 98.8. Players who shot right on their career number average 91.0. Ten of the forty-two testable players in the top fifty picks are running hot. One is running cold.

Here are all thirty hot ones, four or more points above career. The last column is how many goals each would have scored on exactly the same shots at his career rate plus thirty percent of last season's gap (the rule above, applied to the same season).

And the cold ones, three or more points below career, where the same rule works in the player's favour:

Four worth talking about

Cole Caufield is the headline. Fifty-one goals is a number that gets a player drafted in the second round, and he shot 19.8% to get there against 12.7% before. On the same 258 shots, the rule gives him 38. That is still a very good goal scorer, and he is young, which is the group that covers the loss best. But the thirteen goals between 51 and 38 are exactly the part of his season that is least likely to come back, and they are the part the price is built on.

Auston Matthews is the one cold name near the top, and the most interesting line on either table. He scored 27 goals in 60 games, but he took 227 shots doing it: nearly four a night, which is not the shot volume of a player in decline. He shot 11.9% against a career 15.8% built on more than 2,500 shots. On the same shots and the same sixty games, the rule gives him 33, which is a 45-goal pace over a full season. I cannot tell you why he played sixty games, and a season that short is its own risk. The finishing is not the thing I would worry about.

Nathan MacKinnon and Nikita Kucherov are the first- and third-ranked players in the pool and both ran hot. I am not telling you to pass on them, since MacKinnon's 127 points do not depend on eleven goals. But MacKinnon's career rate is built on nearly 3,500 shots, the steadiest baseline of anyone on either list, and he still beat it by 4.6 points. Budget for 42 goals, not 53, and be pleasantly surprised.

Morgan Geekie is the reason I would not treat any of this as a law. He shot 22.0% the season before last, which made him the fourth-biggest spike in the entire twenty-year study. The rule said he would come back down. He shot 21.5% and scored 39. Two consecutive seasons above 21% is starting to look like a player who changed, not a player who got lucky. His "career" number above now includes the first of those two years, so some of that change is already built into his projection. He is the row that disagrees, and I would rather show it to you than leave it out.

What would make me wrong

• Players do genuinely change. Geekie is the proof. A new role in front of the net, a new linemate who makes better passes, or a player simply getting better at finishing will all look like luck to this method for a season or two. Thirty percent is an average; the players who keep more of it exist.
• Young career baselines are thin. I required 300 career shots, which Logan Cooley cleared by nine. For players that early in their careers, the "career number" is itself a small sample, and the young group already keeps more of the gap than anyone else.
• This is individual shooting only. I have each player's own shots and goals, not the shooting percentage of his linemates while he is on the ice. I found no sign that assists regress after a hot year, but I cannot see the finer version of that question.
• The goal columns assume the same shots and games. Nobody takes exactly the same shots twice. Those columns show what last season looks like with luck taken out, not a projection that accounts for injuries, ice time or a trade.
• The study requires a next season. Players who shot hot and then got hurt, demoted or released drop out, which probably makes the decline look slightly smaller than it really is.
• Some players could not be tested. Of 188 ranked skaters, 28 have fewer than 300 career shots, twelve took fewer than eighty shots last season, and four have nothing to test: Aleksander Barkov has no 2025-26 season in my data, and Gavin McKenna, Ivar Stenberg and Roman Kantserov have no NHL history. Names in the export do not always match the NHL's, so I matched on position as well as name. That still failed for Elias Pettersson, because my database holds two players of that name; I placed the Vancouver centre by hand, and he is not part of the 143 behind the correlation.
• Goalies are excluded. Save percentage has its own version of this problem and this piece makes no claim about it.

The short version

When a skater shoots well above his career rate, he keeps about thirty percent of the gap the following season. He takes the same shots, collects the same assists, and scores fewer goals, as more than four in five of the hottest shooters do. Young players cover most of the loss through growth; players over thirty do not.

The draft market prices none of it. Caufield, Necas, Johnston, Boldy, Nylander, Kucherov and MacKinnon all go in the first twenty-six picks on shooting seasons the history says will not repeat, while Matthews takes 227 shots in 60 games and goes as if the missing goals were real.

Ottawa paid for eight years of Drake Batherson. For one of those years, the next one, expect about twenty-five goals.