-0.05: Offense slot-share, repeatability. 0.85: Defense slot-share, repeatability. 7 vs 13: Team spread, offense vs defense (pts). OTT: Bleeds the most slot There is a comfortable idea in hockey analytics that two teams with identical expected goals can be built differently: one lives in the slot, another peppers from the outside, and that shot diet is part of who they are. It is a good story. For the PWHL, the data mostly does not support it, at least not where people expect. We split every 5-on-5 shot this season into slot chances and outside chances, on both ends, and then asked the only question that matters for a team identity: does the pattern repeat? On offense, it does not. Every team generates roughly the same share of its expected goals from the slot, around 53%, in a tight 7-point band. And the small differences are noise: a team's slot share in odd-numbered games barely predicts its share in even-numbered games (repeatability of -0.05 , essentially zero). Nobody is secretly a perimeter team. Everyone is trying to get to the same dangerous ice, and who succeeds in a given month is mostly luck and matchups, not a style. On defense, it is a completely different story. How much slot a team concedes ranges across a wide 13 points, and it repeats hard: the odd-game and even-game numbers line up at 0.85 . Keeping opponents to the outside, or failing to, is a real and stable team trait. Shot location, in other words, is a defensive identity, not an offensive one. At the top of the chart sit the teams that wall off the slot: BOS allows only 48% of its expected goals against from in tight, the best defensive structure in the league. At the bottom is OTT , conceding 61%, far more slot than anyone else. If that team still posted good results, it is worth asking how, and the answer is usually the goalie: a team that bleeds the slot and survives is leaning on its netminder, exactly the kind of arrangement our regression read flagged as living on borrowed time. The defensive structure ranking Sorted by slot suppression, best to worst. Read it as a structure ranking, not a full defense ranking; it is only where shots come from, not whether they go in. BOS (allows 48% of its expected goals against from the slot; shoots 54% of its own from there). MIN (allows 48% of its expected goals against from the slot; shoots 54% of its own from there). TOR (allows 50% of its expected goals against from the slot; shoots 48% of its own from there). MTL (allows 52% of its expected goals against from the slot; shoots 54% of its own from there). NY (allows 52% of its expected goals against from the slot; shoots 54% of its own from there). SEA (allows 54% of its expected goals against from the slot; shoots 52% of its own from there). VAN (allows 55% of its expected goals against from the slot; shoots 50% of its own from there). OTT (allows 61% of its expected goals against from the slot; shoots 55% of its own from there) What this read isn't Eight teams, wide error. A repeatability measured across eight teams has a wide confidence interval, so trust the direction (offense uniform, defense varies and sticks) more than the exact order of the middle of the defensive ranking. The extremes are safe; the cluster is not.. Location, not goaltending. Slot allowed is about where shots come from, which is team structure. Whether those shots go in is the goalie, a separate and noisier question we handled in the goaltending read. A team can concede the slot and still win if its goalie is elite, and that is the interesting tension.. High-vs-rest, our xG. We use a high-danger (slot) cutoff rather than three buckets, a single clean danger line; the model is well calibrated across low, medium, and high (each within a couple of points of perfect, pooled over all seasons). 5-on-5 only, and the values are our model's.. Not rush vs cycle. We can see where a shot came from, not how the play got there. Two teams can concede the same slot share off very different breakdowns; untangling that needs entry tracking we do not have. Methodology Slot share. The fraction of a team's 5-on-5 expected goals (for, and against) that comes from high-danger shots, defined as 0.10+ expected goals per shot. High-vs-rest keeps the danger line clean and is where the slot signal lives.. Repeatability. Each team's slot share computed on odd-numbered games vs even-numbered games, correlated across teams and Spearman-Brown corrected. Offense came to -0.05, defense to 0.85.. Numbers regenerate on every build. A play-by-play re-sync or an xG recalibration reshapes the shares. This pairs with the goaltending read (who stops the slot chances) and the regression read (who is living on their goalie). Where the shots come from, and who stops them, are two different questions.