For the basic stats, NaturalStatTrick is the best site to use in my opinion -- it's free of charge too. I also use EvolvingHockey, AllThreeZones, and HockeyViz, but those are behind a paywall.
That's why drafting intelligently is so important, right? "Buying" a free agent like Sam Bennett at the top of his market is like trying to buy a bitcoin right now, as opposed to a few years back. Unfortunately, off season bargain hunting is even less effective, and you end up with two or three guys who look like Oilers Gretzky on PTO contracts and turn into Rangers Gretzky, when the season starts.
I'm loving your posts so far. Good luck on getting this substack off the ground.
Yeah, that's a perfect analogy. I've never really been a fan of Fitzgerald's drafting, and he often displays the pick for need instead of pick BPA strategy, which hurts so many teams in the long run. Holtz over Rossi/Perfetti. Nemec over Cooley/Wright. Stillman over Stankoven.
Thank you very much! The support so far has been pretty awesome, and I'm looking forward to what's to come!
this situation could have been avoided had Fitz not tried to galaxy brain his first three drafts and just taken literally any one of Marco Rossi, Logan Cooley or Shane Wright.
"B-B-BUT THEY'LL BE BURIED BEHIND HISCHIER AND HUGHES! IT MAKES NO SENSE TO HAVE THEM AT 3C!"
Hughes and Hischier have one full 82 game season between the two of them, and besides they could have just done the same thing with one of those guys that they did with Dawson Mercer. Instead, this team will pretty much always be looking for center depth and will probably never satisfyingly fill the need the same way ONE DRAFT PICK could have.
I say forget the free agent market for centers, I probably would just move Mercer to 3C, Glass to 4C and try to find decent wingers to put with both.
1. The Devils chose not to use some of their younger players in the middle. Halonin was a center in Utica, correct. Then they moved him to wing. Maybe not the quickest, but a different type of center. That is what they said when they signed him. Then moved him to wing.
2. You rely too much on advanced vs real statistics. If the Expected Goals is high, but the actual isn't close, than, maybe that stat is not really all that good. Meier has a great Expected Goal number, but actual goals are less than Bennett so Meier is better doesn't work for me. You want to not sign Bennett because he is bad defensively fine, overpriced fine. But actual goals count on the scoreboard. Expect goals dont
Halonen in the system is fine in a bottom six capacity. Best player in Utica imo.
As far as the xG/goals debate, expected goals directly lead to actual goals on a grand scale. If a player does not provide their team a boost in offensive zone time, shots taken, chances produced, high-danger chances produced — all of which show up in the xG column — chances are they’re not providing those actual goals, either. It’s a predictive measure of on-ice success, so sure, the actual results get the actual wins, but to say that the player whose xG produced is significantly better than the others’ xG is unimportant because the actual goals “don’t reflect that” is negligent. Actual goals count on the scoreboard, and expected goals count in predicting what players are going to show up there more often.
I am questioning the predictive nature of it. To me this is same argument baseball has with expected batting average based on how hard you hit the ball. (which is wildly inaccurate), and the concept that stat people came up with that that were no such thing as clutch hitting. Predicting goals using wide-scale averages seems counter-intuitive to also tracking shooting percentage. Some guys are just better finishers than others. Bennett, who from my limited exposure to him, seems to be a guy who gets a lot of garbage goals, and has done so for a while, long enough to dismiss the "it is a fluke" theory. Through his career, does Meier exceed or under-achieve his expected goal totals.
(I am not trying to be difficult, for the record. As a financial professional, I have worked with numbers for years, and see how they can be misused in doing analysis. Which is why I am skeptical of anything "expected")
In fact, an interesting statistical analysis would be a comparision of expected stats vs actual stats, and a subjective review as to why the differences exist.
You are correct in saying that some guys are just better finishers than others, but that is also reflected in xG. Players with high shooting %s also typically score more goals than they were “expected” to. In both regards, you can label a player as a plus finisher.
I can tell you right now that there is a direct correlation between expected goals and actual goals in that the player/team with the higher xG output also generally has the higher actual G output.
The same goes for xGF% — the teams with the highest xGF% almost always have the highest GF%, too, and are almost always the playoff teams come the end of the regular season. There is obviously variance in specific cases and will always be outliers, but generally speaking, it’s absolutely true.
What sites do you use, can I find those stats?
Absolutely, all of these are publicly available!
For the basic stats, NaturalStatTrick is the best site to use in my opinion -- it's free of charge too. I also use EvolvingHockey, AllThreeZones, and HockeyViz, but those are behind a paywall.
That's why drafting intelligently is so important, right? "Buying" a free agent like Sam Bennett at the top of his market is like trying to buy a bitcoin right now, as opposed to a few years back. Unfortunately, off season bargain hunting is even less effective, and you end up with two or three guys who look like Oilers Gretzky on PTO contracts and turn into Rangers Gretzky, when the season starts.
I'm loving your posts so far. Good luck on getting this substack off the ground.
Yeah, that's a perfect analogy. I've never really been a fan of Fitzgerald's drafting, and he often displays the pick for need instead of pick BPA strategy, which hurts so many teams in the long run. Holtz over Rossi/Perfetti. Nemec over Cooley/Wright. Stillman over Stankoven.
Thank you very much! The support so far has been pretty awesome, and I'm looking forward to what's to come!
this situation could have been avoided had Fitz not tried to galaxy brain his first three drafts and just taken literally any one of Marco Rossi, Logan Cooley or Shane Wright.
"B-B-BUT THEY'LL BE BURIED BEHIND HISCHIER AND HUGHES! IT MAKES NO SENSE TO HAVE THEM AT 3C!"
Hughes and Hischier have one full 82 game season between the two of them, and besides they could have just done the same thing with one of those guys that they did with Dawson Mercer. Instead, this team will pretty much always be looking for center depth and will probably never satisfyingly fill the need the same way ONE DRAFT PICK could have.
I say forget the free agent market for centers, I probably would just move Mercer to 3C, Glass to 4C and try to find decent wingers to put with both.
1. The Devils chose not to use some of their younger players in the middle. Halonin was a center in Utica, correct. Then they moved him to wing. Maybe not the quickest, but a different type of center. That is what they said when they signed him. Then moved him to wing.
2. You rely too much on advanced vs real statistics. If the Expected Goals is high, but the actual isn't close, than, maybe that stat is not really all that good. Meier has a great Expected Goal number, but actual goals are less than Bennett so Meier is better doesn't work for me. You want to not sign Bennett because he is bad defensively fine, overpriced fine. But actual goals count on the scoreboard. Expect goals dont
Halonen in the system is fine in a bottom six capacity. Best player in Utica imo.
As far as the xG/goals debate, expected goals directly lead to actual goals on a grand scale. If a player does not provide their team a boost in offensive zone time, shots taken, chances produced, high-danger chances produced — all of which show up in the xG column — chances are they’re not providing those actual goals, either. It’s a predictive measure of on-ice success, so sure, the actual results get the actual wins, but to say that the player whose xG produced is significantly better than the others’ xG is unimportant because the actual goals “don’t reflect that” is negligent. Actual goals count on the scoreboard, and expected goals count in predicting what players are going to show up there more often.
I am questioning the predictive nature of it. To me this is same argument baseball has with expected batting average based on how hard you hit the ball. (which is wildly inaccurate), and the concept that stat people came up with that that were no such thing as clutch hitting. Predicting goals using wide-scale averages seems counter-intuitive to also tracking shooting percentage. Some guys are just better finishers than others. Bennett, who from my limited exposure to him, seems to be a guy who gets a lot of garbage goals, and has done so for a while, long enough to dismiss the "it is a fluke" theory. Through his career, does Meier exceed or under-achieve his expected goal totals.
(I am not trying to be difficult, for the record. As a financial professional, I have worked with numbers for years, and see how they can be misused in doing analysis. Which is why I am skeptical of anything "expected")
In fact, an interesting statistical analysis would be a comparision of expected stats vs actual stats, and a subjective review as to why the differences exist.
You are correct in saying that some guys are just better finishers than others, but that is also reflected in xG. Players with high shooting %s also typically score more goals than they were “expected” to. In both regards, you can label a player as a plus finisher.
I can tell you right now that there is a direct correlation between expected goals and actual goals in that the player/team with the higher xG output also generally has the higher actual G output.
The same goes for xGF% — the teams with the highest xGF% almost always have the highest GF%, too, and are almost always the playoff teams come the end of the regular season. There is obviously variance in specific cases and will always be outliers, but generally speaking, it’s absolutely true.