You're looking at a 12-game sample and drawing conclusions about efficiency that ignores regression. Edmonton's PP was 19.4% for the season before McDavid's return - that 28.7% clip is unsustainable variance, not a new baseline. Books aren't stupid.

Colorado's penalty kill numbers are skewed by playing Vegas twice and Dallas once in that 10-game stretch. Against average offensive teams, they're allowing 1.8 PP goals per game, not the 2.4 your math suggests. The +140 pricing accounts for natural regression to Edmonton's season mean, which makes the under the sharp play here.

Had a similar read on Edmonton's PP efficiency last month when they hosted Vegas. Loaded up on the over at 2.5 goals thinking their 31% conversion rate over 8 games was sustainable. McDavid looked unstoppable on the man advantage, Draisaitl was finding the back door consistently, and Vegas was taking 5+ penalties per game on their road trip.

Game started perfectly - Edmonton drew two penalties in the first 12 minutes and converted both. I was already counting my winnings. Then reality hit. They went 0-for-4 on their next four powerplays, with McDavid hitting two posts and Draisaitl getting stonewalled on a breakaway off a 2-on-1 shorthanded rush. Final tally: 2 PP goals exactly, pushing my bet.

Point is, even when the underlying numbers scream value, hockey variance can kill you on single-game props. That 28.7% efficiency you're citing will regress - question is whether it happens tonight or over the next 10 games. I've learned to fade these hot streaks rather than chase them, especially when books are basically begging you to take the over at +140.

Ran the numbers on Edmonton's PP deployment since McDavid returned, and there's a hidden factor here: ice time distribution. McDavid's averaging 3:47 per game on PP1, but Draisaitl's PP2 time dropped from 2:31 to 1:52 because they're running more 5-on-3 situations with both on the same unit.

That 28.7% efficiency is actually driven by 6 goals on 11 two-man advantage opportunities - their 5-on-4 conversion is only 22.9%, which aligns closer to season averages. Colorado's 5-on-3 penalty differential is +0.8 per game (they rarely take overlapping minors), so Edmonton likely sees 3-4 standard powerplays tonight rather than the extended advantages that inflated their recent numbers.

The sharp play here is actually the under 2.5 at -165. Books at 30Bet are offering it at -155, which provides slightly better value on what should hit 65% of the time based on adjusted efficiency rates.

Sample size concerns aside, Edmonton's home PP splits are worth noting. They convert at 24.8% at Rogers Place versus 18.1% on the road this season. Tonight's game factors in crowd noise during Colorado penalties and familiar ice dimensions that benefit McDavid's cross-ice passing lanes.

That said, the juice on the over suggests books expect public money on Edmonton's recent hot streak. I'm staying away from this prop entirely - too much variance in a single game outcome when the long-term edge isn't clear.

Cross-border perspective here - been tracking this same prop on US books, and the pricing tells a different story. DraftKings has the over at +125, FanDuel at +135, but offshore books serving Canadian players are consistently higher at +140 to +155. That's not coincidence.

US books get more sharp action on NHL props because of the betting volume, so their lines move faster toward efficient pricing. The fact that BetOnline is offering +145 on the over suggests they're expecting Canadian public money to hammer Edmonton after seeing those McDavid highlight reels.

When there's a 15-20 cent difference between sharp US pricing and offshore books targeting recreational Canadian action, it usually means the recreational side is overvalued. I'd lean toward the under or pass entirely.

Still learning to read these efficiency stats - when you say 28.7% conversion, does that account for short-handed goals against? And how much does McDavid's individual shooting percentage factor into team PP goals versus his assist rate on Draisaitl scores?

Also seeing different O/U totals across books - some at 2.5, others at 3.5. Which total are you finding the best value on?