Golden Knights deserved to lose despite favorable advanced metrics
The Vegas Golden Knights' performance in Game 4 of the Stanley Cup Final was proof that Corsi - and other advanced metrics - mean nothing in a one-game sample size.
The Golden Knights generated 65.4 percent of the game's shot attempts at even strength (Corsi For percentage), but lost 6-2. They also outchanced the Capitals 23-14, and had 12 high-danger scoring chances compared to Washington's five.
So, what went wrong?
For starters, their defensive zone coverage was abysmal. Six goals against would lead one to believe that Marc-Andre Fleury was the problem, but for the most part, he was hung out to dry. Capitals players were routinely left wide-open in the offensive zone, which is why they were able to capitalize on their chances.
They also lost the special teams battle, which isn't factored into most advanced metrics, since the majority are calculated at even strength. The Capitals went 3-for-5 on the power play - partly due to some of the aforementioned defensive-zone lapses - while Vegas was 0-for-4 with the man advantage.
Moreover, in the third period, when the Capitals were supposed to be weathering a 20-minute storm while preserving the lead, the Golden Knights failed to apply the pressure, generating just four even-strength shots on goal in the final frame.
Here is what makes a stat like Corsi flawed: While shot attempts are a good way to indicate offensive-zone puck possession, it means nothing unless those shot attempts are actually getting through. They had 71 shot attempts in the game, but only 30 on goal.
Corsi and other advanced metrics are great over a full season, where trends can be predicted, and puck luck tends to even out, but in playoff hockey, that all goes out the window. The victory formula in the postseason is about doing the little things right, taking advantage of the chances you do get, and winning the goaltending battle. The Capitals did all of these in Game 4, and are one win away from the Stanley Cup as a result.