Backcheck
Recap · MTL at PHI

Flyers hold off the Canadiens

The Flyers stole one. They lost the expected-goals battle 2.66–2.03 and won anyway.

By the ModelJul 12, 11:40 AMxG: xgboost-0.758Edited for clarity
Away
MTL
Montreal Canadiens
2
Home
PHI
Philadelphia Flyers
4
FINAL
Process vs result

The underlying story

2.66Expected goals2.03
57% share for the better side
29Shot attempts (SOG)25
3High-danger chances4
59Corsi (all attempts)36
Score-adjusted: 67% / 33%
Source · NHL play-by-play · XGBoost xG

The Flyers won 4-2 despite controlling only 43% of expected goals. The Canadiens dominated play at 5v5 with 57% of xG, yet failed to convert. Philadelphia's finish was elite; Montreal's goaltending faltered.

The gap is substantial: a 2.66-to-2.03 xG split should produce a Canadiens win, not a comfortable Flyers victory. But shot quality tells the story. Philadelphia converted 75% of their high-danger chances (3 of 4), while Montreal managed 3 high-danger looks and scored on only one. The Flyers were lethal on the rush, tallying 4 goals on 21 chances—a finishing rate unsustainable over time.

Goaltending amplified the disconnect. Samuel Ersson stopped 27 of 29 shots (0.931 SV%), outperforming expectations by 0.66 goals. Jakub Dobes allowed four on 25 (0.840 SV%), underperforming expectations by 1.97 goals.

The turning point came in period three at 12:00 when Alex Bump fired a snap shot from the high slot.

How the game was won

The danger ladder

shotsgoalsexpected
High dangerslot & crease
MTL0 / 3 · 0.66 xG
PHI3 / 4 · 0.97 xG
Medium dangermid-range
MTL2 / 13 · 1.26 xG
PHI0 / 4 · 0.58 xG
Low dangerperimeter & point
MTL0 / 43 · 0.74 xG
PHI1 / 28 · 0.49 xG

PHI went 3-for-4 from high danger on 0.97 expected goals — 2.0 above expected.

Source · Backcheck xG model · NHL play-by-play
The luck layer

Finishing vs expected

expectedactual
MTL
2 G · 2.66 xG · -0.7
PHI
4 G · 2.03 xG · +2.0
012345
Goalies · GSAx (goals saved above expected)
J. DobesMTL
-2.0
S. ErssonPHI
+0.7

MTL finished -0.7 against expected, PHI +2.0 — J. Dobes (MTL) allowed 2.0 more goals than expected.

Source · Backcheck xG model
What they shot — and what went in

Shot diet

goalsshots
MTL
WRIST0/25
SNAP0/7
SLAP0/5
BACKHAND1/4
TIP IN1/1
OTHER0/17
PHI
WRIST2/15
SNAP1/7
SLAP0/3
BACKHAND0/2
TIP IN1/2
OTHER0/7
MTL
PHI
Off the rush
1 G on 28 · 1.80 xG
4 G on 21 · 1.96 xG
Off rebounds
1 G on 3 · 0.30 xG
0 G on 0 · 0.00 xG

Wrist shots carried the volume: MTL went 0-for-25, PHI 2-for-15. PHI scored 4 of theirs off the rush.

Source · Backcheck xG model · NHL shot types
Three stars · computed, not voted
M. Michkov
#39 · R · PHI
1 G · 2 A · 8 SOG · 16:18 TOI · +2
P. Martone
#94 · R · PHI
1 G · 1 A · 5 SOG · 17:37 TOI · +1
O. Bonk
#59 · D · PHI
1 G · 1 A · 1 SOG · 16:28 TOI · +1
Ranked from the box score — points first · Backcheck
What the model flagged
  • PHI won despite being outplayed by xG (43%). Goaltending or finishing carried them.
  • J. Dobes (MTL) gave up 2.0 goals more than expected — well below their workload.
  • PHI was clinical in the high-danger zone — 3/4 on HD chances (75%).
  • PHI was lethal off the rush — 4 goals on 21 rush chances.
  • MTL created 3 rebound opportunities (1 converted) — crashing the net effectively.
  • PHI outperformed their expected goals by 2.0 — elite finishing or lucky bounces.
Before the game — the model's pre-game read
Preview · MTL at PHI · 7:00 PM ET

Canadiens at Flyers: a genuine coin flip

By the ModelJul 12, 11:39 AMEdited for clarity
Win probability

The model's lean

54%
PHI
Model favorite
46%
54%
MTL50PHI

PHI by 8 points of win probability — a modest lean. Projected score MTL 2.32.6 PHI; 19% chance it's tied after 60.

Source · Backcheck Poisson projection · season scoring rates
Season profile · edge from the spine

Tale of the tape

MTL← edge
edge →PHI
48.0%5v5 xG%51.0%
49.0%Corsi%48.0%
2.41xGF / 602.26
2.62xGA / 602.20
2.24Goals for / gm1.96
1.96Goals against / gm1.82
1.018PDO — luck, not skill1.008

The season numbers split 33 — no clear process edge — and PDO (1.018 vs 1.008) says MTL has run hotter.

Source · MoneyPuck season tables · 5v5

The projection screams equilibrium: Montreal 2.3–2.6 Philadelphia, with the Flyers pegged at 54% win probability. This is a genuine pick'em.

The Flyers own the underlying play. Montreal generates 48.0% 5v5 expected goals; Philadelphia returns 51.0%. That's a meaningful edge—the kind that usually shows up in the score over time. But Montreal's Corsi% of 49.0% against Philadelphia's 48.0% suggests the Canadiens are closer to parity in raw shot volume than in shot quality, a gap that could close.

Montreal's PDO sits at 1.018, running hot. The Flyers register 1.008, essentially neutral. Regression matters here. If Montreal's luck normalizes, their goal differential tightens visibly.

D. Vladar plays on a back-to-back, a detail worth monitoring. Backup duty could materialize, or a tired starter could leak on a weak night.

The Players

Cole Caufield anchors Montreal at 0.97 xG/60 with 33 goals and 16 assists; his 20.0% shooting rate is both an advantage and a target for mean reversion. Oliver Kapanen (0.83 xG/60, 19 goals, 12 assists) adds secondary punch, though his 17.8% shooting rate suggests more regression risk.

Philadelphia counters with Owen Tippett at 0.88 xG/60, 19 goals, and 18 assists. His 12.1% shooting is safer ground. Christian Dvorak (0.81 xG/60, 12 goals, 21 assists) is the set-up man, though his 11.2% shot rate leaves room for downside.

Where It Breaks

Mike Matheson is Montreal's defensive weak link. On Philadelphia, that distinction falls to Christian Dvorak—unlikely for a playmaker, a detail that itself telegraphs roster construction issues.

Both teams sit at 0% on high-danger chances and rebound goals, a statistical curio that won't hold. Variance will bite both sides.

The Flyers' 54% projection is really "pick Philadelphia, but barely." Montreal is live. If Caufield's shooting normalizes down and Vladar plays tired, the Canadiens steal this. If the Flyers capitalize on their xG edge and Tippett's shot selection holds, Philadelphia wins decisively. Neither outcome defies the data.

This is the closest thing hockey offers to a true coin flip—two teams separated by talent but intertwined by luck and circumstance.

Computed danger · scouting

Players to watch

MTL
Cole Caufield33G 16A
0.97 xG/60 · 1 HD · 20.0% sh
Oliver Kapanen19G 12A
0.83 xG/60 · 4 HD · 17.8% sh
PHI
Owen Tippett19G 18A
0.88 xG/60 · 2 HD · 12.1% sh
Christian Dvorak12G 21A
0.81 xG/60 · 3 HD · 11.2% sh
Source · MoneyPuck skaters