Player Discussion Jacob Trouba

[Machinehead]

I'd rather win the most games than have the best analytics. Different models produce different results and none of them are fully accurate representations of reality. That's not to suggest that the statistics don't have value as a tool. They do. But they should never be used as the only tool, and there are times when they should be ignored completely in favor of actual results.

Well, obviously, I get that point, but to suggest that faceoffs are more important or show a strong correlation with winning hockey games is completely off-base. If anything, the correlation with strong underlying numbers is better.

Again, this feels like an apples to oranges argument. I'm positing the point that faceoffs aren't an important driver of analytics and you're saying "well, I'd rather win hockey games than have good analytics." It's a non sequitur.

 

[SA16]

Zone starts make a very big difference in terms of "possession" numbers or do you just think it's a coincidence that the 21 players who played 500+ mins last year with a offensive zone faceoff percent under 30 had a combined CF% of 45.6 and the 41 players over 60 has a CF% of 53.8? There is obviously some player selection bias in there but not to that extreme.

The better way to do it would be to look at individual players are break down their CF% based on where they started a particular shift however I do not have the capability or time to do it.

 

[Mandar]

Yeah this is basically it.

Sticking with Bonino as the example here - it's fair to say that he has value on the PK. He's a good shot blocker, good positionally and his FO ability will give the team the opportunity to get a clear right off of puck drop - these are all good things. This applies late in games when face with a 5v6 too, even with the threat of icings against.

The ability for him to win a FO at ES has far less value because your end goal isn't (or shouldn't be) to fire the puck out of your zone. The drawbacks in his game are glaring here.

Plus the bulk of the teams in the league (not the outlier teams that are either way ahead of the pack or way behind) are within range of one another. I have a hard time saying that winning an extra 3-4 faceoffs for every 100 taken is going to dramatically impact a team's performance vs the majority of the league.

It's a nice to have quality and being at basically 56% as a team is hardly a bad thing. I'd easily sacrifice some of that to strengthen other areas though.

**Checks name of thread***

Ah.....

Carry on....

 

[Machinehead]

Zone starts make a very big difference in terms of "possession" numbers or do you just think it's a coincidence that the 21 players who played 500+ mins last year with a offensive zone faceoff percent under 30 had a combined CF% of 45.6 and the 41 players over 60 has a CF% of 53.8? There is obviously some player selection bias in there but not to that extreme.

The better way to do it would be to look at individual players are break down their CF% based on where they started a particular shift however I do not have the capability or time to do it.

When you look at the 10% of the league getting the most extreme usage, it's going to make usage look more impactful.

 

[GAGLine]

Well, obviously, I get that point, but to suggest that faceoffs are more important or show a strong correlation with winning hockey games is completely off-base. If anything, the correlation with strong underlying numbers is better.

Again, this feels like an apples to oranges argument. I'm positing the point that faceoffs aren't an important driver of analytics and you're saying "well, I'd rather win hockey games than have good analytics." It's a non sequitur.

I don't disagree with the bolded. I just argue against the idea that these things don't matter at all.

 

[Machinehead]

I don't disagree with the bolded. I just argue against the idea that these things don't matter at all.

Nobody thinks that, unless they're just annoyed and shouting it at you to make a point.

 

[80shockeywasbuns]

You gave an isolated example that doesn't prove your case. If Bonino wins the draw, is he the one actually turning it over? One of the other players will be the one actually possessing the puck after the faceoff. Maybe the puck eventually makes its way to Bonino again and he cedes possession, but you're making a false equivalency. Bonino being good or bad doesn't validate or invalidate the benefits of winning faceoffs.

Yes, Bonino being bad doesn’t “invalidate” the value of the faceoff, but it certainly offsets the value of the faceoff. On any given shift the negative value from Boninos lack of actual hockey skills is far greater than the positive value from the faceoff win

 

[SnowblindNYR]

With the amount of arrogance coming out of Zlev you'd think he's looking at PhD level math. Congrats you can do basic arithmetic.

 

[haohmaru]

Here’s the thing.

If you have to argue that advanced stats show Trouba sucks at defense just because he starts in his own zone so much, you’ve already lost the high ground. Ok, so maybe he just sucks a little less because of his usage.

He gets paid $8m.

You can get someone to take all the defensive zone assignments and suck at it for far less.

This is only a debate because the fan base has developed an unhealthy attachment to him because ooooooh big hits.

According to Natural Stat Trick, Trouba and Miller have posted the fifth lowest goals-against per 60 rate at 2.06 among D pairs with at least 350 minutes logged.

 

[haohmaru]

Well, obviously, I get that point, but to suggest that faceoffs are more important or show a strong correlation with winning hockey games is completely off-base. If anything, the correlation with strong underlying numbers is better.

 

[Machinehead]

View attachment 782194

That's admittedly a lot stronger than I thought it would be. The CF% and xGF% numbers would be right up there.

That's not really the point, anyway, or not the point I wanted to make before the whole conversation went on a tangent.

Having good analytics is good. Winning faceoffs is good. Both of these things, on the surface, without needed further analysis, are good.

The point is, one thing is not driving the other. It's like, having good goaltending and and having a good powerplay are both good, but there's no relationship between them in terms of performance.

There is an assumed connection between faceoffs and possession-based metrics but it doesn't bear out.

 

[Hunter Gathers]

According to Natural Stat Trick, Trouba and Miller have posted the fifth lowest goals-against per 60 rate at 2.06 among D pairs with at least 350 minutes logged.

Yeah but they suck, duh.

 

[NickyFotiu]

Love this guy. He isn’t just a leader. He makes the game entertaining and gets fans into the game with huge hits and clutch plays. Can’t overstate the importance of that.

The effect he has on his teammates can't be measured but I know how I would feel if he was on my team.

The only good defense is having the puck at all times, or so I've been told.

LOL I do not even want to know who told you that. I really don't

 

[Machinehead]

According to Natural Stat Trick, Trouba and Miller have posted the fifth lowest goals-against per 60 rate at 2.06 among D pairs with at least 350 minutes logged.

Also according to Natural Stat Trick, they have the fifth highest expected goals-against per 60 rate at 2.59 and their on-ice save percentage is .930.

It's not unheard of for guys to break the system and raise save percentage, but 9 times out of 10, it's just the goalie. Quick has been phenomenal.

Over the last three seasons, their on-ice save percentage is only .917.

 

[SnowblindNYR]

That's admittedly a lot stronger than I thought it would be. The CF% and xGF% numbers would be right up there.

That's not really the point, anyway, or not the point I wanted to make before the whole conversation went on a tangent.

Having good analytics is good. Winning faceoffs is good. Both of these things, on the surface, without needed further analysis, are good.

The point is, one thing is not driving the other. It's like, having good goaltending and and having a good powerplay are both good, but there's no relationship between them in terms of performance.

There is an assumed connection between faceoffs and possession-based metrics but it doesn't bear out.

I have thought a lot about this. I don't think a good way of doing the analysis is comparing season statistics for faceoffs and possessions between teams, it should be done within a schedule for each individual team because there are too many other factors between teams. This would be more apples to apples. Has that type of analysis been done?
team.

 

[zlev]

If neither matter, then using one to prove that the other doesn't matter is a poor argument by the other poster.

Both faceoffs and zone starts make a difference. We can argue about how much of a difference, but regardless, the context shouldn't be ignored.

it's not a poor argument at all. the fact that we're well above 50% on faceoffs just makes which zone the play started in even more meaningless than usual.

 

[zlev]

With the amount of arrogance coming out of Zlev you'd think he's looking at PhD level math. Congrats you can do basic arithmetic.

yes, it's very easy to understand, yet people still don't take the 3 minutes it takes to understand what they're talking about. Machinehead is basically saying the exact same things I was last week, are you going to call him arrogant?

 

[haohmaru]

That's admittedly a lot stronger than I thought it would be. The CF% and xGF% numbers would be right up there.

That's not really the point, anyway, or not the point I wanted to make before the whole conversation went on a tangent.

Having good analytics is good. Winning faceoffs is good. Both of these things, on the surface, without needed further analysis, are good.

The point is, one thing is not driving the other. It's like, having good goaltending and and having a good powerplay are both good, but there's no relationship between them in terms of performance.

There is an assumed connection between faceoffs and possession-based metrics but it doesn't bear out.

Oh I agree. I also think there's a huge grey area on a ton of face-offs figuring out who actually won.

I just thought it pretty telling that very, very few teams that are shit on draws win Cups.

 

[SnowblindNYR]

yet people still don't take the 3 minutes it takes to understand what they're talking about. Machinehead is basically saying the exact same things I was last week, are you going to call him arrogant?

I'm not calling Machinehead arrogant because as annoying as him being a contrarian 100% of the time is
he's always humble about it doesn't put anyone who doesn't use or understand advanced stats down.

 

[haohmaru]

Also according to Natural Stat Trick, they have the fifth highest expected goals-against per 60 rate at 2.59 and their on-ice save percentage is .930.

It's not unheard of for guys to break the system and raise save percentage, but 9 times out of 10, it's just the goalie. Quick has been phenomenal.

Over the last three seasons, their on-ice save percentage is only .917.

If that's the case, why are they ahead of Fox/Lindgren who everybody thinks is lights out better defensively? They're playing in front of the same goalies.

And, honestly, I'll weigh actual results a little over expected results.

 

[Machinehead]

If that's the case, why are they ahead of Fox/Lindgren who everybody thinks is lights out better defensively? They're playing in front of the same goalies.

And, honestly, I'll weigh actual results a little over expected results.

Lindgren and Fox were lights out better defensively. They've been pretty widely panned this year.

"Actual results" is meaningless. It implies that expected goals aren't events that occurred on the ice, which is obviously a ridiculous assertion.

 

[JCProdigy]

According to Natural Stat Trick, Trouba and Miller have posted the fifth lowest goals-against per 60 rate at 2.06 among D pairs with at least 350 minutes logged.

Also according to Natural Stat Trick, they have the fifth highest expected goals-against per 60 rate at 2.59 and their on-ice save percentage is .930.

It's not unheard of for guys to break the system and raise save percentage, but 9 times out of 10, it's just the goalie. Quick has been phenomenal.

Over the last three seasons, their on-ice save percentage is only .917.

My question to both of you is why are you using rankings after creating a subset that results in only 14 qualifying pairs?

MH your on-ice SV% do you mean 21-22 thru this year? Because breaking it down per year you have a .930 this year an .892!!! last year and then back up to .933 the year before and then a .937 for that psuedo UCL season that was NHL 20-21.

Good news is their xGF is 2.7 with on on-ice SH% of 10 so we'll see how it plays out the rest of the year.


Digressing back to faceoffs, there are benefits there that can be leveraged to greater effect by smart teams who are likely using the more granular data that AFs are providing them and putting their personnel in the best situational positions to succeed.

https://arxiv.org/ftp/arxiv/papers/1902/1902.02397.pdf

 

[Machinehead]

My question to both of you is why are you using rankings after creating a subset that results in only 14 qualifying pairs?

MH your on-ice SV% do you mean 21-22 thru this year? Because breaking it down per year you have a .930 this year an .892!!! last year and then back up to .933 the year before and then a .937 for that psuedo UCL season that was NHL 20-21.

Good news is their xGF is 2.7 with on on-ice SH% of 10 so we'll see how it plays out the rest of the year.


Digressing back to faceoffs, there are benefits there that can be leveraged to greater effect by smart teams who are likely using the more granular data that AFs are providing them and putting their personnel in the best situational positions to succeed.

https://arxiv.org/ftp/arxiv/papers/1902/1902.02397.pdf

Yeah, it was all of the seasons collectively.

That the same pair is having their on-ice sv% fluctuate 40 points every year is a good demonstration of the validity of "actual results" as they relate to goaltending.

 

[zlev]

I'm not calling Machinehead arrogant because as annoying as him being a contrarian 100% of the time is
he's always humble about it doesn't put anyone who doesn't use or understand advanced stats down.

i'm sorry, i didn't realize people had such thin skin here. as if people aren't constantly arrogantly putting advanced stats down on this forum, either. it's never about not using or not understanding, it's that most people just don't want to understand them. they think it's nerd shit.

 

[JCProdigy]

Lindgren and Fox were lights out better defensively. They've been pretty widely panned this year.

Agreed. Let's not go comparing Trouba/Miller to Fox/Lindgren before this year. The Fox/Lindgren pair were top 10 in the league in %xGA both years. Not so much for the other duo.