In the NHL, where even the worst centre still wins 45% of his faceoffs, faceoffs do not matter. End of discussion!

What is next? Y'all want to make a case that Plus/Minus actually is a good stat after all?
- James_Tanner

James, Face-offs do matter. When you lose the draw, you chase. You win the draw you have possession. The only time draws don't matter is at this level...


(but agreed on the plus/minus)

James, Face-offs do matter. When you lose the draw, you chase. You win the draw you have possession. The only time draws don't matter is at this level...


(but agreed on the plus/minus)
- geta02it

generally, i can get on board with the concept of faceoffs not leading to a huge difference in goals. I can see neutral zone faceoffs not really mattering. OZ vs. DZ should seem to make some difference, with that difference amplified on special teams.

for me, the more interesting thing is if coaches can alter deployment (deploy OZ vs. DZ face-off specialists) more efficiently to gain/prevent some extra goals during the year.

but generally speaking, face-offs probably don't matter all that much because coaches don't let them. they probably do a pretty good job with faceoff match ups. anything to gain is likely at the margins.

Well if a player is on the ice for more goals for than against it's very important I think. How can you argue this?
- The-O-G

It's so easy to argue it - goals are a very bad way to track player performance.

This is because you can only put yourself in good position to score and what happens after the puck leaves your stick might as well be random.

If you use shot attempts to make a plus/minus (i.e Corsi) you aren't getting screwed or rewarded based on goalie performance and luck. Also, you get about ten times the sample size making the data much more repeatable.

Goals are a result, not a process and thus any conclusions based on goals are faulty.

James, Face-offs do matter. When you lose the draw, you chase. You win the draw you have possession. The only time draws don't matter is at this level...


(but agreed on the plus/minus)
- geta02it

But you're failing to account for the fact that hockey is a fast game where the puck changes hands every few seconds.

Yes, it's an advantage to win a faceoff and blast a shot at the net. But that happens so rarely that you would be far better off putting the best players on the ice regardless of their skill at faceoffs.

Every time a coach puts out a second centre who is good at faceoffs at the expense of a winger who is better, he is making a mistake.

generally, i can get on board with the concept of faceoffs not leading to a huge difference in goals. I can see neutral zone faceoffs not really mattering. OZ vs. DZ should seem to make some difference, with that difference amplified on special teams.

for me, the more interesting thing is if coaches can alter deployment (deploy OZ vs. DZ face-off specialists) more efficiently to gain/prevent some extra goals during the year.

but generally speaking, face-offs probably don't matter all that much because coaches don't let them. they probably do a pretty good job with faceoff match ups. anything to gain is likely at the margins.
- Tumbleweed

Don't forget, we're talking NHL where the margins between the best and worst players are extremely small. If you could win 80% of your draws that would be significant. The thing is, if you in 55% of your draws, it's only a 5 faceoff swing, because its a binary thing and there are not multiple outcomes available, so 55% SOUNDS way better than it actually is.

Also, because of how rare it is that faceoffs make a big difference, it makes no sense to employ a faceoff specialist. Unless of course he doubles as one of your better players. But employing Stekle or whatever his name was, that was pure mathematical stupidity.

You're looking at faceoffs in all situations, not just OZ and DZ. Also the research includes goals saved rather than just goals gained. So the differential is still created by winning the faceoff whether his team scores 20 seconds after or 3 seconds after, or even if no goal is scored at all. The whole point here is that faceoffs in the OZ and DZ matter a great deal and the research shows this, especially when a faceoffs specialist is utilized.
- sniper11

Presumably, the season-total stat (+3.81 goals) took into account OZ, DZ, and NZ face-offs, and goals gained+goals prevented...so let's say for argument's sake that OZ accounted for 1.5 goals for, DZ accounted for 1.5 goals prevented, and NZ accounted for 0.81 goals overall...that lessens the impact of OZ and DZ face-offs, doesn't it? No matter the situations, he took 1600+ face-offs and gained a net 3.81 goals (~1 win) over the course of the season. I mean that's cool, but it's probably only half a win better than someone who took the same number of draws at 53%. So if the 53% player is better at hockey, why would you ever take the 56% player for an extra half win over 82 games?

It's so easy to argue it - goals are a very bad way to track player performance.

This is because you can only put yourself in good position to score and what happens after the puck leaves your stick might as well be random.

If you use shot attempts to make a plus/minus (i.e Corsi) you aren't getting screwed or rewarded based on goalie performance and luck. Also, you get about ten times the sample size making the data much more repeatable.

Goals are a result, not a process and thus any conclusions based on goals are faulty.
- James_Tanner

Ok so if a team has a bunch of players that are +50.....won't their team be really good? I think the results speak for themselves.

Ok so if a team has a bunch of players that are +50.....won't their team be really good? I think the results speak for themselves.
- The-O-G

You have the causation backwards. +50 players don't cause a team to be good...a good team (i.e. one that scores a lot and doesn't allow a lot of goals) will give players high ratings. But a bad goalie will let in lots of goals (and give players minus ratings) no matter who is on that team. A lineup of Karlsson/Doughty/Crosby/Kane/Benn would have a minus rating if I was in net.

Don't forget, we're talking NHL where the margins between the best and worst players are extremely small. If you could win 80% of your draws that would be significant. The thing is, if you in 55% of your draws, it's only a 5 faceoff swing, because its a binary thing and there are not multiple outcomes available, so 55% SOUNDS way better than it actually is.

Also, because of how rare it is that faceoffs make a big difference, it makes no sense to employ a faceoff specialist. Unless of course he doubles as one of your better players. But employing Stekle or whatever his name was, that was pure mathematical stupidity.
- James_Tanner

but finding edges at the margins is part of the reason to gather and use more data to me.

but really, the problem i have with looking at face-offs stats is looking at them in aggregate and saying they don't matter. it's drilling down into the data and finding the margins.

boyd gordon and gaustad have to give your pk a better chance by winning 90% of draws in the DZ.

crosby/tavares/datsyuk/toews all dominate the OZ circle (+60% last 5 years) and struggle in their own end. So to me, it's pretty easy to say deploy these guys in the OZ for faceoffs. Your top players can't play every shift, so you have to try and deploy them as efficiently as possible imo.

but finding edges at the margins is part of the reason to gather and use more data to me.

but really, the problem i have is looking at face-offs stats is looking at them in aggregate and saying they don't matter. it's drilling down into the data and finding the margins.

boyd gordon and gaustad have to give your pk a better chance by winning 90% of draws in the DZ.

crosby/tavares/datsyuk/toews all dominate the OZ circle (+60% last 5 years) and struggle in their own end. So to me, it's pretty easy to say deploy these guys in the OZ for faceoffs. Your top players can't play every shift, so you have to try and deploy them as efficiently as possible imo.
- Tumbleweed

but finding edges at the margins is part of the reason to gather and use more data to me.

but really, the problem i have with looking at face-offs stats is looking at them in aggregate and saying they don't matter. it's drilling down into the data and finding the margins.

boyd gordon and gaustad have to give your pk a better chance by winning 90% of draws in the DZ.

crosby/tavares/datsyuk/toews all dominate the OZ circle (+60% last 5 years) and struggle in their own end. So to me, it's pretty easy to say deploy these guys in the OZ for faceoffs. Your top players can't play every shift, so you have to try and deploy them as efficiently as possible imo.
- Tumbleweed

and i know you think gordon and gaustad probably shouldn't be in nhl .... but coaches still need to play the hand that they are dealt. they need to utilize their players efficiently.

these are 2 players that as far as i can see only do one thing well - win DZ faceoffs.

when i look i gordon and gaustad's short-handed FO winning %, i see that both are well below their DZ winning %. I can't remember where gordon is taking all his short handed draws from watching games (a detail i don't pay attention to), but sure makes me think that both of these guys aren't being deployed efficiently.

Don't forget, we're talking NHL where the margins between the best and worst players are extremely small. If you could win 80% of your draws that would be significant. The thing is, if you in 55% of your draws, it's only a 5 faceoff swing, because its a binary thing and there are not multiple outcomes available, so 55% SOUNDS way better than it actually is.

Also, because of how rare it is that faceoffs make a big difference, it makes no sense to employ a faceoff specialist. Unless of course he doubles as one of your better players. But employing Stekle or whatever his name was, that was pure mathematical stupidity.
- James_Tanner

Can't score unless you have the puck.
Let me put it this way. When the puck is dropped, one team gains possession. How you adapt to that play on either defense or offense is yet another aspect of the game. The bottom line is in order to gain an advantage albeit brief as you describe it, you must have possession of the puck and that starts (in the game of hockey) with the drop of the puck. Thats the difference between a statistical analysis of a play, shift or game vs the practical. To say it doesn't matter is illogical. If that were the case anyone can take a draw and be damned the result.

You're wrong. This is only for 5v5 neutral zone faceoffs. In a research paper published by St. Lawrence University in 2012, along with Statistical Sports Consulting, LLC, they reported results from two analyses. In the first of these they found that, in terms of goal differential, not all faceoffs are equal. Tanner is going to say he did not read this. Faceoff wins in the offensive and defensive zone as well as those won on special teams yield a goal differential more quickly.

From the second analysis they conclude that faceoff win percent is a metric that does not currently need adjusting since raw faceoff win percentage is very highly correlated with adjusted faceoff win.

Their results suggest that there are strategic advantages to be gained by having the best faceoff players take faceoffs outside the neutral zone and on special teams. He really is going to say he did not read this. For players that win 60% of their faceoffs, taking 20% more outside of the neutal zone can add an additional 5-7 goals or 2-3 wins per season.

There are ~13,000 off/def zone, pp/pk faceoffs per season. 35 wins yields a goal differential, not 35 more than you lose, just 35 wins. For instance, Toews has already taken over 1000 faceoffs for his team, he will easily eclipse 1500. He has a FOW% of 58.2 this season. He has won 239 faceoffs in the OZ this year and 152 in the DZ, which means he has generated, solely by winning faceoffs, a +11.04 goal differential. Previous works by many others have concluded that a goal differential is worth 1/3 standings points. That means that Jonathan Toews, single-handedly, has earned 3.68 wins, or 7 standings points, for the Blackhawks just by winning faceoffs in the offensive/defensive zone and he will earn at least 2 more before the season is over.

Without telling me to "go find it myself," show me a study with sources, as I have done, that proves you are not a charlatan and making stuff up although "exact numbers evade you."

Tanner loses again.
- sniper11

This is a great post!

This is a great post!
- camfor

Well except for it being totally wrong.

but finding edges at the margins is part of the reason to gather and use more data to me.

but really, the problem i have with looking at face-offs stats is looking at them in aggregate and saying they don't matter. it's drilling down into the data and finding the margins.

boyd gordon and gaustad have to give your pk a better chance by winning 90% of draws in the DZ.

crosby/tavares/datsyuk/toews all dominate the OZ circle (+60% last 5 years) and struggle in their own end. So to me, it's pretty easy to say deploy these guys in the OZ for faceoffs. Your top players can't play every shift, so you have to try and deploy them as efficiently as possible imo.
- Tumbleweed

My point is this: if you replaced boyd gordon with a player who was below avg on draws, but who was faster and more skilled, youd be coming out ahead because the odds of recovering the puck/ clearing the puck are way better and would have an impact for an entire shift, not just three seconds

My point is this: if you replaced boyd gordon with a player who was below avg on draws, but who was faster and more skilled, youd be coming out ahead because the odds of recovering the puck/ clearing the puck are way better and would have an impact for an entire shift, not just three seconds
- James_Tanner

well, that's fine at a strategy level.

deploying limited resources in cap dollars to limited resources in talented players may leave you short in that department.

a coach's job is to deploy his players to gain every advantage possible...which leaves you at times trying to to make the best of a bad situation. as opposed to saying the data doesn't matter, i believe you can use that data to help you make better decisions.

well, that's fine at a strategy level.

deploying limited resources in cap dollars to limited resources in talented players may leave you short in that department.

a coach's job is to deploy his players to gain every advantage possible...which leaves you at times trying to to make the best of a bad situation. as opposed to saying the data doesn't matter, i believe you can use that data to help you make better decisions.
- Tumbleweed

Well yes, I agree about that. However, don't get me wrong. I do not think that faceoffs 'don't matter' in the context of reality. What my point is, and has always been, is that you they don't matter as much as the frequency of their being brought up would lead you to believe.

So when someone says, "You conveniently ignore the fact that McDavid is bad at faceoffs" I say "That doesn't matter" Because it doesn't. McDavid could lose all his faceoffs and he would still be among the best in the world, because the things he can do that others can't have a far greater impact on the game than faceoffs do.

Well yes, I agree about that. However, don't get me wrong. I do not think that faceoffs 'don't matter' in the context of reality. What my point is, and has always been, is that you they don't matter as much as the frequency of their being brought up would lead you to believe.

So when someone says, "You conveniently ignore the fact that McDavid is bad at faceoffs" I say "That doesn't matter" Because it doesn't. McDavid could lose all his faceoffs and he would still be among the best in the world, because the things he can do that others can't have a far greater impact on the game than faceoffs do.
- James_Tanner

fine then.

he's still not good enough to get the into the playoffs.

That whole "best hockey player in the world" thing may have been a tad premature.

How is a bloated government more of a problem than a guy who denies the two undebatably most important issues of our time? Thats just negligence.

I am extremely well versed on what both sides and all major candidates are offering. And its a matter of fact that Ted Cruz has blatantly lied over 100 times in public.

Plus, all of his policies are outdated 20 years ago. The guy is literally nuts.
- James_Tanner

You don't even know what the word "literally" means.

Faceoffs don't matter? Panarin says otherwise http://youtube.com/watch?v=olybN0l30qo

In the NHL, where even the worst centre still wins 45% of his faceoffs, faceoffs do not matter. End of discussion!

What is next? Y'all want to make a case that Plus/Minus actually is a good stat after all?
- James_Tanner

You are always boasting about empirical evidence you occasionally use and throwing insults at people who usually respond the exact same way you just did here. Faceoffs are very important when it comes to creating significant strategic advantage. The fact is that for a player that wins 60% of their 1200 faceoffs, taking 20% more outside of the neutral zone can add an additional 3 goals or one win per season. End of discussion!

Well yes, I agree about that. However, don't get me wrong. I do not think that faceoffs 'don't matter' in the context of reality. What my point is, and has always been, is that you they don't matter as much as the frequency of their being brought up would lead you to believe.

So when someone says, "You conveniently ignore the fact that McDavid is bad at faceoffs" I say "That doesn't matter" Because it doesn't. McDavid could lose all his faceoffs and he would still be among the best in the world, because the things he can do that others can't have a far greater impact on the game than faceoffs do.
- James_Tanner

Like what? And how is it better than a faceoff win? Remember, empirical data only. No eye test. I'm very interested in what skills are more meaningful when related to faceoffs, especially when they are coming after a faceoff loss. I suspect you are just blowing smoke and have no idea what you are talking about, factually.

You are always boasting about empirical evidence you occasionally use and throwing insults at people who usually respond the exact same way you just did here. Faceoffs are very important when it comes to creating significant strategic advantage. The fact is that for a player that wins 60% of their 1200 faceoffs, taking 20% more outside of the neutral zone can add an additional 3 goals or one win per season. End of discussion!
- sniper11

Another "Great post".

Like what? And how is it better than a faceoff win? Remember, empirical data only. No eye test. I'm very interested in what skills are more meaningful when related to faceoffs, especially when they are coming after a faceoff loss. I suspect you are just blowing smoke and have no idea what you are talking about, factually.
- sniper11

And. Another "Great post".
You win! Tanner loses(again)