Your New Favorite Fancy Stat: WTF%

There are certain assumptions that Sports Fans hold true, regardless of game. The first is that your team will try to win the game (likelihood of winning is a different story), the second is that the Refs are being paid off by the other team, and the third is that your team will do better with the home advantage.

As a stat nerd, I can’t speak to the other two, but the third is actually a good rule of thumb. In fact, as of 2/18, only 6 NHL teams have more wins on the road than at home this season. When you add in OT points, only 8 teams have more points away than home.

And you guessed it, Dallas is one of those teams.

If you listened to our most recent podcast, I promised to a reader I would take a closer look at why Dallas is insistent on thumbing their nose at convention. (The Stars are the team with the highest differential of Away wins of the 6 teams. Most of them only have 1 extra Away win. Dallas has 3).

We know a few things about teams that win, in very general terms:

  1. Possession correlates with winning, therefore CF% and CF60 should be higher when winning and CA60 lower
  2. Goals (obviously) correlate with winning, therefore GF% and GF60 should be higher when winning and GA60 lower
  3. Scoring Chances correlate with winning, therefore SCF% and SCF60 should be higher when winning and SCA60 lower
  4. Typically, if you’re winning a game, you’re starting more in the Offensive Zone
  5. If you’re winning a game, you’re scoring more goals so your Shooting % should be higher than when you’re losing
  6. If you’re winning a game, your goalie is saving more shots than when you’re losing a game, so his SV% should be higher
  7. If your Shooting % and your Save % are higher when winning, your PDO will be higher when winning.

Yes, these statements are very “eye test” esque, which we know isn’t always true (see: giveaways/60 for our defense – no, Jordie Benn isn’t the team leader), but they are good places to start.

Now, if we’re looking at Home vs Away stats, we probably want to look at the differences between the two (mathematically: Home #s – Away #s= Delta) just to see if anything stands out about a Home vs Away record. Usually, the bigger and more significant the Delta is, the better or worse a team is doing in that category.

When we look at these Deltas, we can assume that if a team has a more winning Home record, certain Deltas like CF%, CF60, GF%, GF60, SCF%, SCF60, ZSO%, PDO, Sh%, and Sv% will be positive, indicating the Home numbers are higher, and CA60, GA60, and SCA60 will be negative indicating the team is worse in those metrics Away. When the team has a better Away record, you can expect those values to be flipped.

2.18 Home vs Away - Away teams

This chart is all 6 of the teams with better Away records (all data is all man situations, all score situations, and is the season through 2/16). Per usual, I have this chart color coded to hell and back. The important thing is to look at the color of the numbers. A green number indicates the Delta is what would be expected to follow the logic outlined above, a red number indicates it’s unusual. Black means the Delta is so small it’s neutral.

It might stand out to you that both Dallas and Washington have an awful lot of red on that chart. It certainly stood out to me. With how much red the other teams have too, it made me reconsider my assumptions. So while I’m not going to show you the whole chart because it’s giant, here’s a random clip of my extended chart of all 30 teams, color coded in this manner. The things I do for y’all.

2.18 Home vs Away Random

Don’t try to make sense of it, it’s painful, I promise. But you can see that even just in this section, there’s a lot more green than we are seeing from the Stars and Caps. So all that’s left is to measure how unusual the Stars are in relation to the rest of the league, hence the birth of what I’m terming What the Fuck Percent, or WTF%.

To put it in “official” terms, WTF% measures the percent of stats that don’t fit conventional expectations. Mathematically, it looks like this:

“Count of Instances that Don’t Fit Assumptions”/(“Total Times Logic Was Applied”-“Instances of Neutrality”)

In our Away Teams example, each team was measured on 13 categories. Dallas has 10 instances where the numbers don’t fit the assumptions and no Neutral numbers, therefore their WTF% looks like 10/(13-0) or 76.9%. Which looks high at first glance, but again, we don’t really know what to expect.

2.18 WTF% Team Chart

No. It’s just high. And it’s nearly two standard deviations off of average. So. Yeah. It’s pretty significantly unusual.

The cool thing about WTF% is that it doesn’t only apply to the teams. For instance, CF% Delta has 17 teams where the number doesn’t fit our assumption that it will be positive for a Home record and negative with an Away record, and has no Neutral numbers. So the WTF% for CF% Delta is 56.7%. Interestingly, all of the Corsi categories had over a 50% WTF%, which leads me to believe that it’s not a great indicator of “winningness” in general (not just when the Stars are involved). SCF% Delta had a 46.7% WTF%, and has already proven a stronger correlation to winning by other people, and WTF% seems to support this conclusion.

Of the 13 categories, there are two that have a very low WTF%, PDO Delta (26.7%) and GF% Delta (14.3%). Guess which two categories the Dallas Stars have significantly large Deltas in.


In fact, Dallas has a 10% better GF% at home than away.

But they’ve won more games on the road.

The only stat for the Stars that makes sense when you look at WTF%, knowing that Corsi is a bad indicator, is Scoring Chances Against per 60 – and that Delta is so low it means almost nothing (0.124 deviations from the average). Even the Caps, who have the same WTF%, don’t have the same kind of deviations we do in the actual stats.

The other thing I take away from all this data is that it looks like goalies do better at Home, no matter what. When you average the SV% Delta of all the Away Winning Teams, you still come out with a positive number, indicating they’re still better at stopping the puck at Home.

All of this is to say: I’m sorry, readers. I have no idea why the Stars are the way they are. They are the chaotic good in a lawful neutral world, and I will just have to leave it at that.


2 thoughts on “Your New Favorite Fancy Stat: WTF%

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s