Grading Our Own Preseason

The season opens tomorrow night. Over the last four weeks this site published twelve pieces, and most of them said some version of "the board has this player wrong." That is easy to write and harder to be held to, so before the games start, here is every call we can still check, set against where the market finished.

One caveat covers every row. Only the season can say whether these calls were right. A player the market came around on can still have a bad year, and a player it ignored can still score. The board can only show whether the reasoning survived a month of camp news, and whether anyone else ended up in the same place.

The scorecard

The misses come first, because they cost the most.

What we got wrong

Hellebuyck

On 2 September we argued that the board had knocked Connor Hellebuyck from the first goalie drafted to the fourth for a risk that "mostly runs one direction." The worst case we could see was that he stayed in Winnipeg and stayed excellent, so we said take him at 61.6.

The worst case turned out to be one we never listed. He did not report to camp, and the Jets suspended him on 16 September. As of this weekend there is still no trade. Kevin Cheveldayoff says "no doors are closed," and Pierre LeBrun reports that the offers have been "far from fitting." On Sunday's board Hellebuyck went at 82.9, seventh among goalies, and anyone who took our advice in early September paid twenty-one picks more than the market now asks.

The flaw was in the section we titled "What would make me wrong." It covered a bad destination, no trade at all, a rate-based scoring format and the board running ahead of us. It did not cover the player refusing to play, which is the only outcome that turns a Vezina goalie into zero points. A trade request that has gone public can end in a holdout, and we treated it as if it could only end in a trade.

The cliff at pick 148

On 4 September we wrote that the board holds together for 148 picks and then falls apart, so three of your eighteen rounds are not really yours. On that day's file it was true. Jordan Kyrou at 146.9 to Oliver Bonk at 152.8 was a gap of 5.9 picks, 7.6 times the average.

It never showed up in the same place twice. Four days later, on the 8 September file, the biggest gap was Morgan Geekie at 138.8 to Yaroslav Askarov at 143.5. On 20 September it had jumped to pick 26. On Friday's board it sat between Jakub Dobeš and Porter Martone at pick 96. On Sunday the biggest hole was Noah Dobson at 87.9 to Lucas Raymond at 91.7, 3.8 picks and 4.8 times the average. A board that is still filling in has gaps, and they move every time somebody drafts, so what we called a structure belonged to one day's file. The finding that has lasted is the one from our 21 September update: past pick 160 the board stops ranking anyone. Sixty ranked players now sit between 161 and 180, about three per pick.

No goalie in the first fifty

We published this on Saturday, and by Sunday it was false by seven tenths of a pick. Andrei Vasilevskiy moved from 50.2 to 49.5, which makes him the 50th player on the board, and the first fifty now reads 34 forwards, 15 defencemen and one goalie. The piece was accurate to its file, but a headline that depends on one player sitting one spot outside a round number can flip overnight. The rest of the argument still stands: the market waits on goalies because best ball forgives a bad start, and a weekly lineup does not.

What held

Start with the positional mix. Our worst moment of the preseason was pulling the ten-forwards piece after a CSV bug reversed its thesis, and rewriting it to say the market's roster shape was right. On Sunday 125 forwards, 63 defencemen and 35 goalies were ranked. Across an eighteen-man roster that is 10.1 forwards, 5.1 defencemen and 2.8 goalies, and after four weeks it has barely changed.

We said Darren Raddysh at 33.3 was priced right and Morgan Rielly at 152.9 was a declining player at a declining player's price. Both have moved less than four picks since. We did not claim an edge on either, and the board never offered one.

On 16 September we showed that a player coming off a shooting-percentage spike keeps about thirty percent of it, and that the board does not discount for it. Of the twenty hot shooters we listed, eleven have fallen since 8 September and nine have risen. The two big fallers are Tom Wilson and Alex Tuch, and both went down because Washington's forward group got crowded, which has nothing to do with how they shot. Leave those two out and the other eighteen moved about 0.1 picks on average, the same as the board as a whole. The market still treats a 21% shooting season as the new normal.

Where the board came to us

A price moving toward a call is not proof the call was right, since the market can be wrong in the same direction we were. Each of these did move on news, though, and the news is below.

Hedman has gone from 109.9 to 94.3. It is the strangest row on the card, because camp went against us. Tampa's first power-play unit this preseason has been John Carlson with Brayden Point, Nikita Kucherov, Brandon Hagel and Jake Guentzel, with Hedman on the second. The board bought Hedman anyway. In August we wrote that camp units are guesses and the first two weeks of real games are the depth chart, so this row stays open until mid-October.

Snuggerud has gone from 119.5 to 112.2 and was on St. Louis's first power-play unit on 21 September, alongside Robert Thomas, Mason McTavish, Dylan Holloway and Philip Broberg. Nine power-play points last season was the room to grow we pointed to.

Michkov has gone from 167 to 151.4, the largest move among our five contract-year players. He scored in the Flyers' preseason win over Boston, and Rick Tocchet said he "felt it with the puck." Will Smith has gone from 94 to 84.1.

Kreider was unranked on every board we checked for three weeks, and was the subject of a whole piece about why a 50-point winger was going to nobody. He opened Montreal's camp on the top line with Nick Suzuki and Cole Caufield and is now at 166.5.

Gridin joined the board this weekend at 173.1, the 220th player of 223. He was one of the three unranked players our breakout piece said cost nothing, and for Saturday's preseason game in Edmonton he was projected on Calgary's top line with Morgan Frost and Matt Coronato.

Still open

Cole Hutson was 86.4 when we wrote about him and is 86.0 now. We called him "either the buy of this draft or its trap," on fourteen games of evidence, and the board has not picked a side either.

Gabe Perreault is still unranked. The Rangers opened the preseason with him on the Mika Zibanejad and Alexis Lafrenière line that finished last season together, and Lafrenière has risen fifteen picks since 8 September on that same line. Perreault still costs nothing.

Marco Kasper is still unranked, and his case has moved further than any other on the card. We said his job depended on Dylan Larkin leaving, and that only a Larkin trade would open it. Larkin has not been traded. He hurt his upper body before camp, and while he was out Kasper centred Viktor Arvidsson and Emmitt Finnie and picked up two assists against Columbus on Saturday. Larkin has fallen from 77.7 to 97.6. The door we described is partly open for a reason we did not predict.

Evangelista and Marchenko have barely moved in either direction.

What we are taking into the season

Each miss leaves something to change. When a player asks out, we will price the holdout along with the trade. A gap in a board that is still filling in is noise until it turns up in the same place twice; the cliff had moved four days after we published it, and we kept measuring it as though it were fixed. And a headline should not hang on one player sitting a single spot out of place.

Every call on the "Open" and "Board came to us" lines gets graded properly from tomorrow, in real games. We will come back to them once there is enough season to say something.