Felipe

Around hockey by day. Overthinking hockey by night. Numbers, rabbit holes & dad life.

[Felipe @philsbad]

@OTRHockey For goalies who switch teams, prior WAR is basically useless (r=-.04),
but cap hit still has a little signal (R²=.04).

Curious if salary still predicts Year 1 WAR after controlling for prior WAR + age.

If so, teams may be pricing something the public model can’t see.

[Felipe @philsbad]

https://t.co/u7ovEAncjs

[Felipe @philsbad]

I’ve always known the second period “long change” matters, but I didn’t realize how much.

In 2023-24, after removing empty-net goals, the NHL had 2,349 goals in the 1st period and exactly 2,349 in the 3rd.

The 2nd period? 2,752. That’s +17.2%.

[Felipe @philsbad]

That’s what I’d keep an eye on this season.

Not just whether 2nd-period scoring stays elevated, but which teams are best at exploiting the long change — and which get trapped by it.

There may be real value in treating 2nd-period shift management as its own tactical problem.

[Felipe @philsbad]

https://t.co/P0hPYQLZ2f

[Felipe @philsbad]

I’ve been looking at when NHL teams pull the goalie, and there’s a pretty interesting gap between what teams do and what the numbers suggest they should do.

Teams have adjusted over time. The problem is the “right” answer may have moved even faster. 🧵

[Felipe @philsbad]

The interesting part to me isn’t just “pull the goalie earlier.”

It’s that teams adopted an analytics lesson, but the game changed enough that the optimal answer appears to have moved again.

Makes you wonder how many other “solved” hockey decisions have already moved too.

[Felipe @philsbad]

None of this means teams should stop resting goalies. Health and injury prevention still matter.

But from a performance standpoint, it raises an interesting question:

Are NHL teams managing fatigue most aggressively at the position where the on-ice effect is actually smallest?

[Felipe @philsbad]

I’ve been looking at how NHL teams manage fatigue, and one thing keeps standing out:

We’ve gotten much more careful about resting goalies over the years.

But the stronger performance drop may actually be happening with the skaters in front of them. 🧵

[Felipe @philsbad]

Very small sample, but 2026-27 has started in the same direction.

Through tonight, I count 12 games where one team was on a true B2B against a rested opponent.

The tired teams are 4-8 and have been outscored 35-23.