Injury report
McDavid's return and what it does to your top line
Six weeks out with a high ankle sprain is not the kind of absence a centre skates off in one shift. McDavid came back to nineteen minutes on Tuesday, three under his season average, and the coaching staff kept him off the second power-play unit entirely. That is the usage pattern to watch, not the highlight reel.
Our projection drops him to 1.4 for the next three fixtures before climbing back toward 1.8. Drafting him tonight is not wrong — he is still the best centre on the board — but paying full freight for a player on a minutes ramp is how solid squads end up mid-table.
The secondary effect matters more. With McDavid back, the winger who inherited his power-play slot loses roughly four minutes of prime ice a night. If that winger is on your season-long roster, this is the week to see what he fetches in a trade.
Market watch
Sleeper goalies trending high for week twelve
Goaltending is the most mispriced position in daily fantasy hockey, and it is mispriced in a predictable direction. Managers chase netminders on winning teams, when the scoring table actually rewards volume: a point per ten saves means a goalie facing forty shots behind a leaky blue line can outscore a shutout on a quiet night.
Hellebuyck is the obvious example. The Jets concede the third-most shots in the league, and he has faced thirty-five or more in six of his last eight starts. His save percentage sits at .921 through that run, which is remarkable given the traffic in front of him.
Two others worth a look this week face similar workloads on the road. Neither will win you a board on their own, but both give you a floor that a low-volume starter simply cannot, and a floor is what lets you take a risk with your third forward slot.
Prospect watch
The rookie minutes curve nobody is pricing in
Every season a handful of first-year skaters get promoted to a scoring line in December and stay there. The market takes about three weeks to notice, and those three weeks are the cheapest points available all year.
The tell is not production, it is deployment. Watch for a rookie whose even-strength ice time crosses fifteen minutes and who starts appearing on the second power-play unit. Points follow deployment with a lag, almost every time.
We publish a deployment tracker each Monday for exactly this reason. It is not glamorous reading, but it is the single most useful column on the site for anyone playing season-long leagues.
Under the hood
How our projection model actually works
The model retrains every morning on the previous night's event log. It weighs recent usage heavily, career baselines lightly, and opponent quality somewhere in between. There is no black box scraping social media for sentiment — it is a regression on ice time, shot share and quality of competition.
It is right often enough to be useful and wrong often enough to keep the game interesting. Over a full season the projections explain roughly half the variance in nightly scoring, which sounds modest until you remember the other half is a puck bouncing off a shin pad.
We publish the model's error bars alongside every projection because a number without a spread encourages false confidence. A skater projected at 1.6 with a narrow band is a very different pick from one projected at 1.6 who swings between zero and four.
