NewThe banger score’s weights are now measured instead of hand-set — the 7 multipliers behind every score used to be judgment calls. They are now priced by a rule: a weight is a price per unit, and units are cheap in a category where everyone piles up a lot of them. Hits run from about 20 to 300 across rosterable skaters, so one hit is a sliver of that range; goals run from about 10 to 50, so one goal is a big slice of it. The rule reproduces the old hand-set weights with an r² of 0.968, which is what earned it the job, and it prices shots and blocks higher than the old numbers did while easing off assists and power-play points. Scores sit on the same scale as before and the top of the board barely moves, but shot-blocking defencemen climb: Jacob Trouba, Darnell Nurse and MacKenzie Weegar all gain 25 spots or more.
NewMetrics worth highlighting — every card now opens with the 4 charts that actually explain its projection, picked automatically from the 40 or so further down the page. They surface where a player broke from his own 3-year rate, where his season split in half, where a rate sits at the edge of the league, and where we’re taking the number somewhere other than last year. Each caption is measured off the chart drawn beside it, so nothing gets claimed that the picture doesn’t show. Players with almost no ice time last season get no section at all, which is the honest answer for them.
NewEvery stat name on a card is a link now — click G, Hits, PPP or PP TOI in the projection table, or any label on the banner, and the card jumps to the section that explains that number. The table also shows H1 and H2 columns for 2025-26, so you can see whether a line was carried by one half of the season or held all year.
ImprovedCards read better on a phone — trend charts now sit 2 across instead of stretching one chart to the full width, which was scaling every label and line to about twice its intended size. The 3-year baseline label also no longer disappears under the line it labels.
ImprovedMinnesota reviewed — the Wild are now checked line by line against a real depth chart and a second outside projection set. Matt Boldy gains 8 points, the biggest move on the roster, once his goals and his assists get pulled apart: his even-strength finishing holds up, his power-play finishing was lucky, and his assist rate is held down by Quinn Hughes only arriving at midseason. Yakov Trenin’s hits climb to 370 on a hit rate that has risen 4 straight years. Quinn Hughes takes over the top power-play unit and pushes Brock Faber down to the second one, while Bobby Brink and Danila Yurov both pick up power-play minutes they weren’t projected to get. Blake Coleman and Jared Spurgeon both come down on age. See the Lines page →
ImprovedLos Angeles reviewed — the Kings are now checked line by line against a real depth chart and a second outside projection set. Byfield inherits the power-play slot Kopitar and Kuzmenko left behind, and the hit projections across the roster were rebuilt player by player. See the Lines page →
ImprovedBlocked-shot data repaired for 4 teams — the 2022-23 source export recorded almost no blocks for Los Angeles, New Jersey, San Jose and Tampa Bay, which dragged down the block projections of 74 players who are still on a roster. Those rows are now excluded, so the block totals for Karlsson, Hedman, Sergachev, Raddysh and both Hugheses all move.
ImprovedEdmonton and Florida reviewed — both clubs have now been checked line by line against a real depth chart and a second outside projection set, with every disagreement chased down rather than split. Detroit is deliberately held back and still shows the automatic model, so it stays marked Auto until its roster settles. See the Lines page →
NewPower-play points split into goals and assists — the power-play box on every card now shows what a player’s power-play points are actually made of. A net-front finisher and a quarterback can post the same total and be worth very different things to your lineup.
New5v5 alongside the totals on every counting chart — each goals, assists, points and shots chart now draws a second, muted line for even strength alone. The gap between the 2 lines is the player’s special-teams production, so you can see at a glance who scores on his own and who needs the power play.
NewNext season’s ice time on the ice-time charts — the year-over-year deployment charts now end with our 2026-27 projection as a blue point, the same way the scoring charts already did.
FixThe menu on the Report Card — the Report Card is built separately from the rest of the site and shipped without a header, so the menu vanished the moment you opened the page and the only way back was the browser button. It now carries the same brand and menu as every other page, on desktop and on phones.
FixThe menu on phones — the header links ran off the right edge of a phone screen and got cut off instead of wrapping, so Report Card couldn’t be reached on mobile at all, with no scroll or tap to reveal it. The menu now wraps onto a second line and every link is reachable down to the smallest phones.
FixA power-play cliff that decided points by a rounding error — when we project a player’s power-play scoring rate we normally pull it partway back toward his last 3 years, because one season on the power play is a small, noisy sample. We used to switch that off completely for anyone whose projected power-play minutes rose more than 15%, which fits a player promoted from nothing to the top unit but not an established one whose minutes merely edged up. The cutoff also landed mid-range: Mitch Marner came out at 15.26% and lost 6 points for it, with 29 other skaters sitting within a whisker of the same line. That pullback now fades in gradually instead of snapping off, so a genuine promotion is still respected and a small bump no longer throws away 3 years of history. 60 players moved, up and down, most by a point or so.
ImprovedSecond pass on the first six teams — Anaheim, Boston, Buffalo, Calgary, Chicago and Columbus have now been re-checked category by category against two independent outside projection sets, not just our own model. Where all three disagreed we went digging rather than splitting the difference. See the Lines page →
ImprovedWe tested one of our own rules — when a player’s shot or block rate has moved the same direction for several years, how much should we trust the trend versus his longer track record? We checked both against what actually happened last season. Blocks genuinely bounce around, so leaning on the longer record is right there — only about a quarter of block-rate trends continued. Shots are far more repeatable and want the most recent season, which is already how our model treats them. Worth saying plainly: one of our own Columbus corrections was on the wrong side of this, and we reverted it the same day.
ImprovedHits and blocks for young players — for a prospect we blend a scouting projection with whatever he did in the NHL. We were applying that blend to hits and blocks too, which quietly halved them. Those are the steadiest things a young player does and scouting reports say little about them, so they now lean on what the player actually did.
ImprovedA few good games no longer set a player’s physical stats — a hit or block rate from a short stint used to carry across a full season at full strength. It now gets weighted by how much of the season there actually was. A player whose rate is steady year after year is left alone; only the ones built on a handful of games move.
FixOffseason moves that our roster feed missed — a player who signs or gets traded after the roster data is pulled can stay listed with his old club, which quietly keeps him off his new team’s depth chart. We now cross-check every roster against outside sources and can pin a correction so it survives the next refresh. See the Lines page →
FixPlayers who share a name both get projected — Vancouver has two Elias Petterssons, a forward and a defenseman. Only the forward existed on our board; the defenseman played 70 games and had no projection at all. Both are now listed separately.
ImprovedSmoother response to ice-time changes — projections used to round at several steps in a row, so a few seconds of projected ice time could swing a player by two assists, and some totals were simply unreachable. The math now carries full precision and rounds once, at the end.
ImprovedPower-play production now has to match power-play minutes — a new check flags anyone given first-unit minutes while still scoring at his old second-unit rate, which is what happens when a player changes teams into a bigger role.
NewProjection Report Card — we graded our 2025-26 preseason projections head-to-head against 10 other public projectors on draft-day usefulness and category accuracy, then put our own misses on the same table as everyone else’s — including the deeper players below our published board, so nothing is cherry-picked. Open the Report Card →
ImprovedSmarter top-role projections — an elite, stable role (a No. 1 power-play quarterback, a 25-minute No. 1 defenseman) now holds high but no longer just repeats last year’s ice time to the decimal — each carries a small, realistic year-over-year drift instead of an exact copy.
NewHow the model works — a full, plain-language breakdown of how every projection is built, from ice time to finishing regression, with a pipeline diagram. Read the Method page →
NewPlayer bios on every card — birthplace country, height and weight, and full draft info (year, overall pick, and team) now sit under each player’s name.
NewHand-review status on the Lines page — each club’s depth chart is now marked Reviewed or Auto as we work through all 32 teams alphabetically, checking lines, deployment, and player projections against real usage. Open the Lines page →
FixShared-name player cards — cards for players who share a name (e.g. Sebastian Aho) showed an error instead of the projection; now they resolve to the right skater.
NewTeam Lines & Power Plays — projected 2026-27 depth charts and PP1/PP2 units for all 32 clubs, with fantasy ranks, category leaders, and power-play time by quarter. Open the Lines page →
NewRisers & Downers — a buy-low / sell-high board that screens every projection against shot-quality and luck signals. See the Movers →
NewCompare players — line up any set of skaters side by side on their projected line and rank. Open Compare →
NewPower play & deployment on every card — projected ice time and power-play time, plus which way the role is trending.
NewCoaching context — who’s behind each bench for 2026-27 and how a new voice or an in-season change moves a player’s minutes.
NewRegression reads & upside — every stat shows the process behind it — shot quality, on-ice percentages, finishing luck — and how the model moves.
NewQuarter-by-quarter & year-over-year trends — every category charted so you can see a role or a rate rising (or fading) inside the season, not just the total.
NewContext-brain reads — deployment, role, coaching and injury notes on each card explaining why a projection should move.
ImprovedComponent projection model — every stat is now built from role × rate across even strength, power play, and penalty kill — not one blended number.
NewTeam depth-chart snapshot — each card shows where the player sits among his own teammates at his position.
NewShareable cards — one-click share images and copy-ready blurbs on the Risers & Downers board.
NewFull player cards — per-82 projections vs last season vs a 3-year baseline, with Fantrax-aligned teams and positions.
NewLeague-percentile shading — a card’s bars are shaded against the whole league, so “good” is calibrated, not guessed.
What’s new — players
Notable projection calls — search any player up top for the full card.
BumpBrady Tkachuk — we had cut his power-play time to less than he saw in any of his 4 Ottawa seasons, on the theory that Florida’s loaded top unit would leave him splitting time. He is the net-front piece on that unit instead. Minutes restored, and his even-strength time raised too, because he opens the year alongside Barkov and Reinhart.
BumpMatt Savoie — promoted to McDavid’s wing, where he finished last season with 9 goals and 19 points in 24 games after the Olympic break. The bet is on his rates rather than his minutes, since a 22-year-old rarely holds a top-line spot wire to wire and Edmonton has a new coach.
FadeSam Bennett — off the first power-play unit. His minutes there were earned across a season Barkov missed from start to finish, and Barkov is back with Brady Tkachuk arriving alongside him. Both spots Bennett might have claimed are spoken for.
FadeAnton Lundell — corrected for the same vacancy. His power-play time reads 1.65, 1.16 and 1.29 in the 3 years before last season, then 2.68 with Barkov out. Trimmed back toward the role he held before that opening, though still above it.
HedgeBrad Marchand — we had cut his power-play time to less than half of what he has run every year for 4 seasons, which was too steep even with Barkov back. Minutes restored to a second-unit share. His finishing is a separate story and stays faded: he shot nearly 20% at even strength in a 52-game age-37 season, against a career norm near 10%.
FixGustav Forsling — we were projecting power-play time for a defenseman who does not play the power play. He filled in for an injured Seth Jones for one stretch last season, and our model read that fill-in as a standing role.
BumpVasily Podkolzin — real second-unit power-play minutes, which he had not been given. His goal projection is deliberately left alone, because his shooting percentage doubled last season while his shot rate fell, and that part does not repeat.
FixIsac Lundestrom — removed from Columbus’ fourth line — he was reported in early July to have torn an Achilles and is expected out until around December. He was sitting on the depth chart as a healthy regular. Del Bel Belluz takes the spot, which is also the likeliest real outcome.
UpdateKent Johnson — reshaped rather than raised. We had him at 33 assists and only 7 power-play points — the rebound was landing entirely at even strength while his power play stayed stuck at last season’s collapse. He was in on 14% of the goals scored while he was out there on the man advantage, against a league norm near 60%. That does not repeat. Assists down, power play up.
HedgeSean Monahan — trimmed. We were using his season-long power-play time instead of how the season ended — his minutes there fell by more than half after Christmas and stayed down for 41 games.
BumpDmitri Voronkov — power-play minutes restored. He missed nearly all of the final quarter with a hand injury, and the model was reading that absence as a demotion. His scoring rate on the man advantage was left alone — it was already right.
BumpZach Werenski — we were the lowest projection of everyone on a defenseman whose scoring rate has climbed three years running and whose power-play time was still rising at season’s end. Raised to hold last year’s rate, not beat it. His blocks were corrected too — then corrected back, see above.
BumpCole Sillinger — scored 8 goals on chances worth about 15. That is a finishing slump, not a skill change. No power-play bump — he does not play there.
UpdateDamon Severson — moved up to the top pairing next to Werenski. We already projected him for the second-most ice time on the Columbus blue line while listing him on the third pair.
FixPatrick Kane — was missing from Chicago’s depth chart entirely — no line, no pairing, no power-play unit — because our roster feed still had him in Detroit. A 65-point player was invisible on his own team’s page. His projection never changed; it sits between both outside reads.
BumpAnton Frondell — reshaped from playmaker to shooter. We had him at more assists than goals, based on 12 games in which he did not score once on the power play despite shooting constantly. His shot volume is higher than Bedard’s. Still projected under both outside reads.
BumpFrank Nazar — he finished below what his chances were worth at even strength and on the power play in the same season, and he takes over as the top center for roughly the first month while Bedard recovers.
HedgeTeuvo Teravainen — trimmed. There is no first-unit power-play spot left for him in Chicago, and his power-play scoring has run hot for three straight years while his even-strength game fell off across the board at 32.
HedgeBowen Byram — trimmed. We were paying twice for the same promotion — once by giving him first-unit power-play minutes, then again with a raise on top. His even-strength rate has been flat for four straight years across two teams.
BumpConnor Bedard — assists raised, goals deliberately left alone. His shooting cooled in the second half but his playmaking did not, and his chance quality never moved. His goal projection already assumes his best-ever finishing rate.
BumpTyler Bertuzzi — only half his shooting luck was real luck. At even strength he ran hot, but his power-play finishing was actually below what his chances were worth — he scores from the net front, and that part repeats.
BumpArtyom Levshunov — we had him projected for fewer goals than his own chances were worth, after a season where he shot 1.5% at even strength. That corrects upward.
UpdateNick Lardis — hits raised sharply. His points were right, but we had him at little more than half his own hit rate — and in this format the hits are most of his value.
HedgeEthan Del Mastro — cut hard. We had him at nine goals; he has never scored one in 45 NHL games, and his chance quality says he is not about to.
HedgeConnor Mackey — cut hard. He was ranked inside the top 210 on hits and penalty minutes from a season three years ago, having played three NHL games last year.
BumpSimon Nemec — Calgary traded two first-round picks for him and lost both of its veteran defensemen, but we were still projecting the third-pair role he had in New Jersey. Given real first-unit power-play time — hedged, because he hasn’t produced on the man advantage yet.
BumpZayne Parekh — raised, but from even strength rather than more power play. He started nearly every shift in the offensive zone and was on the ice for almost none of his team’s scoring, which corrects hard.
BumpJoel Farabee — moved up to Backlund’s line after the Kadri trade and scored 18 points in his last 29 games. We had him below every version of himself.
HedgeJonathan Huberdeau — trimmed. He’s 33 and coming off hip surgery, and we had him repeating his three-year average exactly, with three separate boosts stacked on top of each other.
HedgeMaxim Tsyplakov — trimmed. We were applying his rookie-season scoring rate to fourth-line minutes.
BumpMatt Coronato — assists raised. Calgary had the league’s worst even-strength shooting percentage, and the assist column is where that hurts most — you can’t set up goals nobody finishes.
FixAnthony Mantha — was still listed in Pittsburgh after signing with New Jersey, which kept him off the Devils’ depth chart entirely.
FixNicholas Robertson — was still listed in Toronto after his trade to Pittsburgh.
BumpJosh Norris — his assist rate doubled last season and we were projecting it back at his old level — the step held all year, so we credited most of it.
HedgeDavid Pastrnak — trimmed. His shot volume has fallen four straight years and we had him above every outside projection.
BumpCharlie McAvoy — we were fading him twice for the same reason — once by the model, once by hand. Corrected.
FixJosh Doan — his goals and assists were flipped: the point total was right, but seven goals were sitting in the assist column.
BumpCutter Gauthier — the whole industry had him at 23 goals last season and he scored 44. His shot rate is still climbing.
BumpKiefer Sherwood — hits restored — we had cut them for a smaller role, but his ice time actually went up.
BumpFrank Vatrano — buried in the lineup and hitting more because of it — his hit rate rises as his minutes fall.
BumpTristan Luneau — Anaheim gutted its blue line and he was still being projected as a call-up.
BumpBrad Marchand — the model already regressed a 52-game hot streak — the manual cut on top was double-counting it.
BumpFilip Hronek — we were regressing a spike he never had; his scoring rate has been flat for four years.
NewElias Pettersson (D) — the other Elias Pettersson — 70 games, roughly 160 hits and 100 blocks, and no projection until now.
NewGavin McKenna — first overall pick with a first-unit power-play role, held below the outside read as a no-sample rookie.
NewTyson Hinds — a 125-block defenseman who was missing from the board entirely.
UpdateRasmus Dahlin — hits raised and blocks lowered — both were sitting on the wrong side of what he actually did.
BumpMartin Necas — the power play, not even strength, is what broke last year — his scoring rate on the man advantage fell by a third while his power-play ice time hit a career high. Betting on a partial rebound.
BumpSam Malinski — his ice time climbed all season and finished in a top-four role — the projection was fading a 40-point pace back toward a poor prior year.
UpdateNazem Kadri — re-anchored on his Colorado role after the trade — fewer even-strength minutes on a deeper team, but a real power-play spot. Assists nudged up; the line was repeating the worst season of his last four.
BumpGabriel Landeskog — back in a genuine top-6 role after the long layoff — production scaled up to match, hedged for the time away. Power-play points pinned up too: he held first-unit minutes at the lowest power-play rate on the roster.
HedgeNick Blankenburg — power-play time trimmed to spot duty — he sits 6th on Colorado’s blue line and the carried-over minutes were producing points a real unit would get.
HedgeFabian Lysell — trimmed to a conservative depth line — he still has to crack a deep lineup, and there’s no NHL sample behind the number.
BumpBowen Byram — takes over Chicago’s PP1 quarterback job — the big new contract says he’s the guy.
HedgeArtyom Levshunov — cedes the PP1 quarterback role to Byram — still a real power-play piece, just on the second unit.
BumpNikolaj Ehlers — Carolina’s power-play focal point — nudged into the PP1 minutes band.
HedgeJackson Blake — power-play minutes capped to his real second-unit role, but even-strength time bumped for his rising usage.
FixK'Andre Miller — power-play time corrected — last year’s was a Q1 mirage; his value is 5v5 minutes and blocked shots.
BumpFraser Minten — sophomore step — growing role and rising chances, not a flat rookie repeat.