2026–27 Fantasy Hockey Top 100: September 3 Baseline

September 3 baseline: this top 100 retains the format framework and editorial scenarios established on September 3, 2026. The September 28 availability and price corrections and published rankings are preserved.

The rankings are preserved. Role assumptions, selection bands and floor/base/ceiling outcomes are editorial scenarios at that date, not confirmed current assignments or measured forecasts. This labelling correction is not a new statistical review or a post-preseason update.

Data checked: September 28, 2026 at 3:01 p.m. ET. Team assignments and historical totals were cross-checked with NHL records; availability and draft prices were checked against dated sources below. Roles, workload shares and output ranges are editorial scenarios, not confirmed opening assignments or verified forecasts. An absent injury report is not medical clearance.

This is a redraft-first decision board, not a prediction contest. The floor, base and ceiling are planning bands rather than promises. Market price comes from the public Daily Faceoff consensus/ADP snapshot and the NHL.com board available at the baseline date, then PoolForge adjusts for role, scarcity and scoring fit. A good selection can fail; every row therefore names the downside and the event that should change the recommendation.

Read the four format lenses before using a rank

• P — points-only: goals, assists and power-play production receive most of the weight.
• B — multi-category/bangers: shots, hits and blocks can protect a player when scoring cools.
• G — goalie-enabled: goalie workload, wins and ratios change roster construction and the opportunity cost of a skater.
• C — salary-cap efficiency: expected production is judged against cap cost and replacement value, not reputation.

PoolForge scoring-settings callout

PoolForge's default scoring gives goalies zero fantasy value until goalie scoring is enabled. The default skater weights are one point per goal, 0.5 per assist and 0.1 for plus/minus, while wins, shutouts, goals against, save percentage and saves begin at zero. Do not draft from the G lens unless your commissioner has deliberately enabled and reviewed goalie scoring. Confirm all settings in the scoring guide before applying any board.

The compact grades run from A+ to D and are format-specific, not talent grades. “Floor / base / ceiling” describes a reasonable range if the stated role holds. Confidence measures the evidence behind the role, not confidence that the ceiling will occur.

How the baseline board differs from a consensus list

NHL.com and Daily Faceoff provide two independent market anchors. PoolForge then grades each player through four scoring lenses, applies defense and goalie scarcity only when the roster settings create it, and discounts unresolved health or role. Selection bands deliberately overlap: roster construction can justify taking rank 28 before rank 24 without claiming the player is universally better.

Before locking the board, revisit the specialist guides for late-round sleepers, rookie tiers, goalie workload, defensemen roles and breakout and bounce-back prices. Those guides carry the longer evidence trail behind the compact codes below.

Top 100 selection board

The price column is the acceptable overall-pick band; grades are ordered P/B/G/C; outcome numbers are floor/base/ceiling overall finishes. Role codes are T6 or T4, PP1, G1 or 1B. Confidence is H, M+ or M. Risks are workload/ratios (W/R), power-play/minutes (PP/M), or health/role/shooting (H/R/S). Triggers use INJ for official injury and delta-10% for a ten-point start-share move. These codes describe September 3 editorial assumptions; confirm current deployment separately.

What changed on September 28

Moved Hellebuyck and Snuggerud down for availability and acquisition price; Faber’s price is conditional on clearance. The board is an editorial order, not current ADP. Its role and rank ranges remain scenarios, not measured forecasts.

Baseline date and editorial correction

Recheck official news before drafting. The information-verification date records the existing September 28 availability and price review. Roles and outcomes remain September 3 editorial scenarios. This labelling correction is not a new statistical review.

Ten largest baseline risers and fallers

Baseline risers are Celebrini, Bouchard, Dahlin, Thompson, Raddysh, Wallstedt, Snuggerud, Martone, LaCombe and Dobes because market discount, role or scarcity supports the move. Baseline fallers are Suzuki, Hutson, Wedgewood, Schaefer, Bussi, Bedard while injured, Markstrom, Bobrovsky, Stone and Stamkos because price, health or workload narrows the margin. These are relative-to-market moves, not predictions of good or bad seasons, and describe the September 3 baseline only.

How to use the compact board at the table

The board is a sequence of decision bands, not a command to select the next printed name. Begin each turn by crossing out unavailable players, then compare only the remaining players in the current band. The four grades appear in points-only, multi-category, goalie-enabled and salary-cap order. A slash-separated outcome band is floor, base and ceiling; role codes identify top-six or top-four work, PP1 access, or a lead/1B crease. Confidence falls as the board moves toward camp-dependent jobs. Risk codes flag health, role, shooting, minutes, power-play competition, wins or ratios, while the final code states the specific event that reopens the rank.

Use position scarcity only when the roster can exploit it. If several similar forwards remain but one stable PP1 defenseman sits alone in the band, the defenseman can be the rational pick even with a slightly lower raw points projection. The same is not true for a goalie in PoolForge default scoring, where goalie events are worth zero until enabled. In a goalie-enabled format, compare the marginal start and ratio value with the skater being passed over; do not treat the G label as automatic scarcity.

Construct the first half of the roster around dependable minutes and category coverage, then use later bands for discounted volatility. Two fragile goalies, multiple unresolved PP1 bets and several injured fallers can all be defensible individually but form a weak portfolio together. Mark the largest exposure on the roster before every pick and use the next selection to balance it. If live ADP pushes a target one full band earlier, accept that the edge has changed instead of chasing the original list.

The riser/faller section is the audit trail between this board and the public consensus. To reassess a baseline move, require a dated reason tied to official availability, observed deployment, a material projection change or a meaningful price gap. Preseason goals alone do not qualify. When evidence conflicts, widen the outcome band and lower confidence rather than pretending the uncertainty has disappeared. Keep a printed fallback in every band so a late scratch, unexpected run or platform-position difference does not force an improvised reach while the draft clock is moving. Record the reason for every manual override.

A repeatable selection protocol

1. Write your scoring weights, active slots, bench size, transaction cadence and goalie requirements beside the board.
2. Compare the recommendation with the live room price. A player is not a sleeper, breakout or waiver win after the discount disappears.
3. At every refresh gate, check official health and transactions first, then line, power-play and goalie usage. Treat practice reports as evidence of probability, not certainty.
4. Choose the player who improves the roster's weakest scarce category without creating an unmanageable risk cluster.
5. Record the change trigger. Acting on a prewritten threshold is more reliable than reacting to one goal, one save percentage or one viral highlight.

Projection disagreement is information. When independent boards differ by several rounds, identify the hidden assumption—games played, PP1 access, tandem share or category mix—and decide whether your pool rewards it. The purpose of these tables is to make that assumption visible before the pick.

What to verify next

Use the draft-day runbook to turn ranges into a queue, then compare players with the PoolForge player explorer. The explorer is descriptive; the recommendation remains conditional on the cited timestamp and your scoring contract.