Get the fantasy hockey advantage.

We project Connor McDavid at 136 points, eight clear of Kucherov. The top three are close enough that the order between them is preference rather than projection.

Celebrini is the one worth arguing about. His range runs 106 to 140, more than twice as wide as anyone else in the top five, because a 20-year-old still climbing has far less settled history to project from. That is the upside and the risk in the same number.

So: the top three are close to interchangeable, and McDavid's floor of 129 is higher than Draisaitl's ceiling of 124. Celebrini is where you take the swing. Jack Hughes at 114 is the next name out.

Thompson faced 1,587 shots that our model priced at 172.7 expected goals. He allowed 140. That is 32.7 goals saved above expectation, 5.3 clear of the next goalie.

His .912 save percentage does not stand out, which is the point: save percentage treats every shot the same, and the shots he faced were harder than average. Wedgewood posted a better .921 and still saved eleven fewer goals than Thompson.

So: the raw save percentage understates him. Workload difficulty is the whole gap between 11th and 1st.

McDavid finished with 138 points to MacKinnon's 127, and more per game, 1.68 to 1.59. But 54 of McDavid's points came on the power play against MacKinnon's 30.

Strip the power play out and it flips hard: 75 points at 5v5 for MacKinnon, 57 for McDavid. MacKinnon's line also controlled more of the expected goals at 5v5, 59.2% against 56.5%.

One caution on MacKinnon. He scored 53 goals against 38.5 expected, so 14.6 of them came from finishing above his chances. McDavid was almost exactly on his number, plus 2.9. Finishing that far above expectation is the part least likely to repeat.

So: if your league counts power-play points, McDavid. If you want the player driving play at even strength, MacKinnon, with the caveat that some of his goal total was finishing luck.

At entry level it is Macklin Celebrini, $0.97M against a grade of 73.2 where that cap hit normally buys 39.5. But an entry-level deal is underpaid by construction, so that says more about the CBA than about San Jose.

At real money it is Brandon Hagel, $6.5M in Tampa. He lands in the top five under 74% of the 102 ways we tested weighting offence, defence and intangibles against each other, more than any other skater.

Nobody clears 75%, and that is the honest part. Change what you think defence is worth and the order changes. Read these five as one tier rather than a ranking, and note Matt Boldy is the next name out, close enough that the fifth spot is a coin flip.

So: if you want one name, Hagel. If you want the truth, these five and Boldy are a single tier, and where anyone sits inside it depends on what you think a defenceman is worth.