Armchair Hockey Scout

Ask the book

2025-26 edition Ask a direct question about this edition of the book — who, how many, how common, which group more — and it is answered with a table computed on this page from the edition’s index: the season line, height and age, the on-ice percentiles, the grades and the traits read out of every report. Ask for a view on named players — what a comparison comes down to, what makes a man — and you get the book’s reading of their own pages, and only those. The model turns your question into a query and phrases the result, or reads the pages it is handed; it never originates a figure, and an answer carrying a figure that is on neither the table nor the pages is withheld. It does not forecast, and it says so when a question is outside what the book holds.

Tables are computed on this page from the 2025-26 edition’s index; readings rest on the players’ own pages; the model never counts. Twenty questions an hour.
← 2025-26 edition

This is the 2025-26 edition — an archived read. Read the 2026-27 report →

51

Collin Graf

San Jose Sharks Right Wing

Undrafted out of Quinnipiac, and 44 per cent of his shot attempts from inside 20 feet in a 33-game rookie season.

SeasonTeamGPGAPTS+/-PIMPPSHGWSS%TOI
2023-24SJS7022-4000070.016:36
2024-25SJS335611-481114012.515:52
2025-26 so far—————————————
Career—405813-881114710.615:60

This Season So Far

The 2025-26 season has not started. The empty row above is his, and both it and this paragraph fill in on their own from the morning after he first plays: what he has done, what it is worth against his own last season, and the run he is on when the two disagree.

Last Season

5 goals and 6 assists in 33 games — 11 points at 15:52 a night, minus-4, with a shorthanded goal and only 8 penalty minutes.

43.2 per cent of the expected goals at five-on-five, the 4th percentile, across 415 minutes. His ice went 15-17 against an expectation of 12.52-16.48, which on a sample that small says almost nothing.

The Finesse Game

40 attempts at a median of 24 feet with 44 per cent from inside 20 and only 6 per cent from beyond 40 — the least perimeter profile of any forward on this roster. He goes to the middle and he shoots from there.

6 assists in 33 games. He was a point-a-game college player and the hands are the reason San Jose signed him out of a bidding war; the pace of the NHL game is the thing he has yet to solve.

The Physical Game

6-1 and 190 and honest without being heavy. He forechecks properly and he is not going to move a defenceman off a puck.

8 penalty minutes in 33 games. He is disciplined and uninvolved in the physical half of the game, which for a rookie winger is neutral rather than good.

The Intangibles

Undrafted, and he signed with San Jose in 2024 after three seasons at Quinnipiac, where he won a national championship and finished as a Hobey Baker finalist.

40 NHL games and 13 points. He is 23 and on a club that will give any young forward as much rope as he can use.

What kind of player

The searchable half of the report. The first group is computed from the season's own data; the rest are read out of the prose above, and each one carries the sentence that produced it. A subject the report never raised is simply absent.

  • a strengthread out of the prose, and claimed
  • ordinaryraised, and called ordinary
  • he lacks itraised, and he does not have it

Each computed tag also says how well its measure holds up from one season to the next, because a tag describes a season that has finished and a reader will assume it predicts the one coming. Some do and some do not, and the gap is wide: five-on-five workload repeats well, scoring late does not repeat at all.

  • undrafteda strength· arrived without a pick attached

    “**Undrafted**, and he signed with San Jose in 2024 after three seasons at Quinnipiac, where he won a national championship and finished as a Hobey Baker finalist.”

  • net-fronta strength· lives in the blue paint; the screens and the tips

    “He goes to the middle and he shoots from there.”

  • forechecks-aggressivelypresent and ordinary· first man in, and on the puck

    “He forechecks properly and he is not going to move a defenceman off a puck.”

  • hard-to-movehe lacks it· cannot be shifted from where he decides to stand

    “He forechecks properly and he is not going to move a defenceman off a puck.”

The numbers underneath

On the ice

Five-on-five, 2024-25 · 415 minutes

Expected goals
43.2%
10th pct

share of the chances while he was out there

Shot attempts
44.2%
9th pct

share of the attempts while he was out there

Goals while he was on the ice: 15 for, 17 against. Expected: 12.5 and 16.5.

The eye and the numbers

Hindsight
Report
36th
Numbers
10th

The report ranks him 26 points above where the numbers do. It was written for a back edition, with the seasons that followed already known. Neither the agreement nor the gap says anything about the eye, and this report is excluded from the club-level comparison for the same reason. Why the numbers land there sets out the club he did it on, the minutes he played and the men he played them with.

Both bars are percentiles among the 637 skaters in this edition with a report and enough ice time. The report bar ranks the mean of the grades above — 70 on their own scale — against every other report's mean. The numbers bar ranks his share of expected goals. A gap is a disagreement about where he belongs, not two scales talking past each other.

Tracked

NHL Edge, 2024-25 · 33 games

Top speed
22.6 mph
67th pct

fastest he was clocked all season; the league's average best is 22.2

Bursts over 20 mph
1.94 a game

how often he reaches that speed, not how fast he goes

Distance
2.55 miles a game

skated, all strengths

Hardest shot
83.6 mph
36th pct

one shot; the league's average best is 83.9

A top speed is a season's best, and a man who takes more shifts gets more tries at one; the percentile is the feed's own, against every skater, and says less between a top-pair defenceman and a fourth-line wing than it does between two men in the same job. The counts are shown per game for the same reason.

Usual linemates

2024-25 · 415 minutes at five-on-five

Share of his five-on-five ice time spent alongside each player, measured from shift charts. Special teams are excluded: a power-play unit is not a line.

Why the numbers land there

Five-on-five, 2024-25

The club
SJS took 44.2% of the expected goals as a club. His own share is 43.2% — 1.0 below the club he did it on.
The role
415 minutes at five-on-five — 12th percentile in ice time.

With and without

TeammateTogetherHim apartMate apart
Mario Ferraro47.3%39.7%46.4%
Timothy Liljegren41.7%44.2%48.3%
Alexander Wennberg54.6%36.6%42.6%

47.3% together, 39.7% for him apart and 46.4% for Ferraro apart. He was worse away from Ferraro, who was no better away from him — he needed that pairing more than Ferraro did.

Expected-goal share while both were on the ice, and for each man in the minutes he played without the other. Apart is the pair subtracted from his own totals. It does not settle who was carrying whom — apart from each other, both were still playing with somebody — but it does say whether the share travelled when the pairing broke up.

Where he shoots from

goal on goal missed

54 unblocked shots in 2024-25, 5 of them goals. His median goal came from 25 feet. 44% of his shots came from inside twenty feet.