Has anyone ever checked the Deserve-to-win-o-meter against actual results?

[seventieslord]

I understand the DTWOM is based on expected goals by each team. But I'm not clear on whether it takes score effects into account. A team that goes down by 2 early and spends the whole game throwing everything at the net could end up 80% on the DTWOM. But I'm not sure a team that does that actually wins 80% of the time.

So I'm wondering, has this model been checked against real results? Surely there have been hundreds of games, if not thousands, where the DTWOM said one team was up at 80%. That's in theory. What about in practice. Do those teams actually win 80% of the time? If not, wouldn't it be a good idea to calibrate these likelihoods to reflect reality?

Click to read more...

 

[sr edler]

If not, wouldn't it be a good idea to calibrate these likelihoods to reflect reality?

Reality isn't always the most lucrative alternative.

 

[CharasLazyWrister]

I understand the DTWOM is based on expected goals by each team. But I'm not clear on whether it takes score effects into account. A team that goes down by 2 early and spends the whole game throwing everything at the net could end up 80% on the DTWOM. But I'm not sure a team that does that actually wins 80% of the time.

So I'm wondering, has this model been checked against real results? Surely there have been hundreds of games, if not thousands, where the DTWOM said one team was up at 80%. That's in theory. What about in practice. Do those teams actually win 80% of the time? If not, wouldn't it be a good idea to calibrate these likelihoods to reflect reality?

I think you’re sort of asking the deserve to win o‘meter to be something different than what it is. The keyword is “deserve”, in this case over a full 60 minutes.

If your team/goalie sucks for the first ten minutes of the game and you give up 4 goals before the halfway mark, you very well may dominate the next 55 minutes and technically “deserve” to win, but those first ten minutes did you in and you will almost always lose because of it despite the rest of the game.

In other words, you’re talking about that silly meter as if it should reflect something akin to betting odds which is not at all the target of that site.

 

[thaman8765678]

Deserve to win o meter is pretty useless.

I saw the Kings "deserved" to beat the Oilers. Yet every shot I saw was an unscreened perimeter muffin to the chest.

A team doesn't deserve to win unless they score goals.

 

[Mickey Marner]

Yeah, I snickered at the Leafs-Bruins game one results, too.

From a Reddit AMA four years ago.

 

[A1LeafNation]

I think you’re sort of asking the deserve to win o‘meter to be something different than what it is. The keyword is “deserve”, in this case over a full 60 minutes.

If your team/goalie sucks for the first ten minutes of the game and you give up 4 goals before the halfway mark, you very well may dominate the next 55 minutes and technically “deserve” to win, but those first ten minutes did you in and you will almost always lose because of it despite the rest of the game.

In other words, you’re talking about that silly meter as if it should reflect something akin to betting odds which is not at all the target of that site.

We call this score effects.

 

[Stephen]

Reality isn't always the most lucrative alternative.

Reality is not a big enough sample size.

 

[BB79]

What does "deserving" to win matter? The real deserve-to-win-o-meter is the scoreboard.

Thing is useless.

 

[Beau Knows]

What does "deserving" to win matter? The real deserve-to-win-o-meter is the scoreboard.

Thing is useless.

It could be a useful tool for predicting things like which teams are going to turn their seasons around and which teams are going to regress - if the algorithm is actually accurate enough to proclaim which team "deserved" to win most games.

That's a tall task, and from the sounds of it, this algorithm probably can't do it. For instance it apparently uses expected goals, which from what I understand don't take into account the strength of goaltending you are facing (someone correct me if I'm wrong here). So if you have Hasek in net, you're probably winning games you didn't "deserve" to, but that's not luck.

 

[Crosby2010]

Ray Bourque deserved to win a Cup from a sentimental standpoint.

As for teams, I figure the team that works the hardest deserves it the most. There are those games where a team is completely outworking the other team and loses 2-1 because of a bank shot on a clearing attempt goes in the net. Despite outshooting the other team 42-18. To me, that team would have deserved it, but still didn't win. Not sure how else to measure this.

 

[Thallis]

What does "deserving" to win matter? The real deserve-to-win-o-meter is the scoreboard.

Thing is useless.

The regular season is more about having the correct process rather than individual results, so theoretically an accurate measure could be a useful tool.

As it stands it's a score effects measure.