Jackson Blake scored twice in Carolina's 3-2 overtime victory Friday at Wells Fargo Center, including the unassisted winner 5:28 into the extra frame. The two goals pushed his season total to 24 across 85 games - 22 in the regular season, two more in the playoffs - on 28.4 expected goals according to MoneyPuck's model. That's a -6.4 goals versus expected differential that ranks among the worst conversion rates in the NHL.
Then he ended a playoff game anyway.
Blake's production sits at the intersection of statistical absurdity and playoff necessity. His 11.3% shooting percentage across 81 regular-season games suggests finishing ability well below league average for a player whose shot locations and volume predict 28-29 goals, not 22. Travis Konecny runs +5.1 goals above expected this season and gets celebrated as an elite finisher. Blake sits 11.5 goals below that standard and just beat Philadelphia with a three-point night built on two goals and one assist.
The numbers don't reconcile. They might not need to.
Blake's overtime winner came unassisted, meaning no setup from Taylor Hall or Logan Stankoven to manufacture a higher-danger look. He created the chance himself and buried it past Daniel Vladar, who stopped 37 of 40 Carolina shots and gave Philadelphia every opportunity to force a shootout. The goal registered 0.31 expected goals on MoneyPuck's shot quality model - a legitimate scoring chance from a dangerous area, but not automatic. Blake converted it in the game's most critical moment while carrying a season-long conversion rate that suggests he shouldn't be.
His first goal, assisted by K'Andre Miller and Hall at 12:35 of the second period, tied the game at 1-1. He also assisted on Stankoven's goal at 4:13 of the third that gave Carolina the lead to 2-1 before Philadelphia's Alex Bump tied it at 5:52. Between those three points, Blake logged 17:29 of ice time in a game Carolina dominated territorially - the Hurricanes generated 3.42 expected goals at all situations while Philadelphia managed just 1.28, per MoneyPuck's game data.
Carolina outshot Philadelphia 40-17 across 65 minutes and controlled 78% of 5-on-5 shot attempts in regulation. Blake's two goals came on a fraction of that volume, but they came when the game hung in balance.
The traditional explanation for a player this far below expected goals is simple: he gets chances but can't finish. That's the story MoneyPuck's model tells about Christian Dvorak, whose -5.2 differential across 80 games fits a center who generates decent looks without the touch to capitalize. Blake's underlying numbers suggest similar problems. His 61.0% on-ice expected goals percentage during the regular season ranked among Carolina's best forwards, meaning he contributed to dominant possession and chance creation. He just didn't convert at the rate shot locations predicted.
Until the playoffs demanded it.
Friday's analytics paint Carolina's shot-share advantage in sharper relief than the 3-2 scoreline suggests. The Hurricanes generated 14 high-danger chances to Philadelphia's six, per Natural Stat Trick. They controlled 62.4% of shot attempts at 5-on-5 and 70.3% of expected goals. Frederik Andersen faced just 1.28 expected goals across 65 minutes - Philadelphia managed 15 shots in regulation, the lowest output Carolina allowed all series.
Blake accounted for two actual goals on minimal shot volume in a game where Shayne Gostisbehere fired six attempts without scoring and Seth Jarvis logged four shots with nothing to show. Carolina's 40 regulation shots produced 3.0 expected goals at 5-on-5. Blake converted twice while operating at a season-long efficiency rate that suggests he shouldn't have.
The contrarian explanation is that Blake exploits blind spots in how expected goals models weight danger. If he takes unconventional chances from areas the model rates as low-percentage but executes them at higher-than-average rates, the differential makes sense. He's not defying physics - he's taking shots the model undervalues and converting them at sustainable clips the numbers don't capture.
That theory requires believing Friday's performance isn't variance but signal. It requires trusting that a player 6.4 goals below expected across 81 regular-season games is actually skilled, not lucky, and that Carolina's coaching staff is correct to deploy him 17:29 per night in playoff games. That's third among forwards behind Jarvis and Jordan Staal - not depth-role minutes. That's top-six trust in a player whose underlying finishing metrics suggest he belongs on a checking line.
The alternative is simpler: Blake ran hot for 81 games, stayed hot for four more, and will regress hard in future playoff rounds. If that's true, the Hurricanes (53-22-7, 113 points, first in the Metropolitan Division) are leaning on production built on variance rather than skill. The stakes make the question urgent. Carolina's 113-point regular season was built on depth scoring - Blake's 22 goals were meaningful contributions to that total. If those goals borrowed from future production, the Hurricanes are thinner up front than their record suggests.
If they're sustainable, Carolina possesses a playoff weapon whose value doesn't translate into the models other teams use to evaluate talent.
Philadelphia managed 17 shots and 1.28 expected goals across 65 minutes Friday. Vladar posted a .925 save percentage while facing 3.42 expected goals, per MoneyPuck's goalie data - he allowed 0.58 more goals than the model predicted but kept the Flyers within striking distance until Blake's overtime finish. Andersen stopped 15 of 17 for an .882 save percentage against 1.28 expected, roughly break-even performance that required Carolina's offense to generate volume Philadelphia couldn't match.
The Hurricanes did exactly that. Their 40 shots and 3.42 expected goals represented total territorial control - they controlled 67% of zone entries at 5-on-5, with 28 controlled carry-ins against Philadelphia's 12, per MoneyPuck's tracking data. The Flyers (43-27-12, 98 points) relied on dump-ins for 59% of their entries and generated 0.72 expected goals per 60 minutes off those plays. Carolina's carry-ins produced 1.89 expected goals per 60.
Blake's overtime winner came off that kind of controlled entry. He carried the puck into Philadelphia's zone, created separation, and beat Vladar high. The goal carried 0.31 expected value - not a tap-in, not a desperation heave. A legitimate scoring chance that required execution. Blake provided it while carrying a season-long conversion deficit that suggests he shouldn't have.
Future playoff rounds will test whether Friday's performance represents genuine skill or borrowed production. Carolina has advanced through the playoffs this spring, and Blake has contributed throughout. His two Friday goals pushed his playoff total to four across 10 games, a 10% shooting percentage that aligns with his regular-season rate.
But his regular season included 28.4 expected goals on 195 shots across 81 games. He finished with 22. That's a conversion rate so far below model predictions that it typically signals a player revealing his actual finishing skill rather than running cold. Expected goals models work at scale because most players regress toward predicted totals over sufficient sample size. A forward 6.4 goals below expected across a full season isn't unlucky - he's showing you who he is.
Blake keeps scoring anyway. His three-point night Friday extended Carolina's playoff run while operating at efficiency levels that don't match his underlying shot quality. The question isn't whether this production is sustainable - the models say it isn't. The question is whether it matters if Carolina keeps winning games like Friday's, where territorial dominance and Blake's timely finishing proved enough to overcome Philadelphia's defensive structure and Vladar's strong performance in net.