Fancy Stats

[hobarth]

TO has hired fancy stat experts, guys who have come up with Corsi, Fenwick and other such stats and so has other teams hired such people. Why hire them when the info they published was commonly available?

Who are the people publishing Corsi stats and other such stats, how do they have the time to view and evaluate 2,460 games plus playoff games? As far as I know these stats are provided by volunteers who probably are like Leaf fans viewing Leaf games, or Ottawa fans viewing Ottawa games and on and on. Leaf fans have a Leaf bias, Ottawa fans an Ottawa bias so how accurate are these commonly available stats?

Plus and Minus is a stat that the NHL actually pays someone to ascertain, it's a professionally provided and derived result, it might not be perfect but I think it's the only reliable stat that's available to us, the fans.

TO, Ottawa and other teams have analytics departments, I'm positive teams that have analytics dept. don't trust or rely on what many of we fans consider to be the bible.

An example of an amateur or biased results might be hitting, the Islanders have tradionally lead the NHL in the number of hits for the last 10 years(I'm guessing here) because they pad their home stats.

Click to read more...

 

[axlrose87]

TO has hired fancy stat experts, guys who have come up with Corsi, Fenwick and other such stats and so has other teams hired such people. Why hire them when the info they published was commonly available?

Who are the people publishing Corsi stats and other such stats, how do they have the time to view and evaluate 2,460 games plus playoff games? As far as I know these stats are provided by volunteers who probably are like Leaf fans viewing Leaf games, or Ottawa fans viewing Ottawa games and on and on. Leaf fans have a Leaf bias, Ottawa fans an Ottawa bias so how accurate are these commonly available stats?

Plus and Minus is a stat that the NHL actually pays someone to ascertain, it's a professionally provided and derived result, it might not be perfect but I think it's the only reliable stat that's available to us, the fans.

TO, Ottawa and other teams have analytics departments, I'm positive teams that have analytics dept. don't trust or rely on what many of we fans consider to be the bible.

An example of an amateur or biased results might be hitting, the Islanders have tradionally lead the NHL in the number of hits for the last 10 years(I'm guessing here) because they pad their home stats.

Our team has definitely gone all in with this crap and I hate it. We need more guys who play good old fashioned hockey instead of having a bunch of pencil pushers pick players for the team.
Dubas seems to have acknowledged that a bit this offseason.
Fancy stats don’t seem to apply in the playoffs anyhow.
Some players grow a foot in the playoffs and some shrink in the playoffs. They should come up with a fancy stat to help determine who those guys are.

 

[BoredBrandonPridham]

TO has hired fancy stat experts, guys who have come up with Corsi, Fenwick and other such stats and so has other teams hired such people. Why hire them when the info they published was commonly available?

teams buy the stats from agencies that tabulate them. they don't get them from the same sources fans get it from, and they don't hire people to sit there and count stats.

the league now tracks corsi and fenwick. no reason to think it's any different than hits, +/-, etc.. afaik. but wouldn't be surprised if teams still buy that from agencies that probably have more stringent quality review processes.

fun fact: for real-time stats, babcock had the back-up goalie on the bench track a couple things during the game. probably little things like face-off and scoring chance metrics.

Plus and Minus is a stat that the NHL actually pays someone to ascertain, it's a professionally provided and derived result, it might not be perfect but I think it's the only reliable stat that's available to us, the fans.

a useless stat tracked accurately is still a useless stat.

 

[LeafsNation75]

Our team has definitely gone all in with this crap and I hate it. We need more guys who play good old fashioned hockey instead of having a bunch of pencil pushers pick players for the team.
Dubas seems to have acknowledged that a bit this offseason.
Fancy stats don’t seem to apply in the playoffs anyhow.
Some players grow a foot in the playoffs and some shrink in the playoffs. They should come up with a fancy stat to help determine who those guys are.

Was Mark Hunter someone who cares about analytics like the way Kyle Dubas does? He played in the NHL from 1981-1993 and when he was in charge of who the Leafs drafted in 2015, 2016, and 2017. So far the only player who's worked out full time was Mitch Marner.

 

[X66]

I know it’s scary to some to have to learn something new, but it’s okay.

And if you don’t understand it, that’s okay too!

 

[Morgs]

I honestly thought the first two posts were straight up trolls.

It's one thing to attempt to understand and not be okay with it, but it's another thing to outright be against something that all it does is give you more information, a greater understanding, and helps give you more than evaluating players/teams with points, wins, hits, and +/-.

If you want to know more there are plenty of people on this website or twitter that would gladly explain.

 

[CantLoseWithMatthews]

I only read the first sentence, but the Leafs didn't just hire analytics people who came up with corsi/fenwick and then released it to the public

 

[axlrose87]

Was Mark Hunter someone who cares about analytics like the way Kyle Dubas does? He played in the NHL from 1981-1993 and when he was in charge of who the Leafs drafted in 2015, 2016, and 2017. So far the only player who's worked out full time was Mitch Marner.

I don’t remember Mark Hunter ever being the gm of the leafs but I would gladly take his approach over Dubas any day. I imagine you were referring to lamorielo. Yes. Lamorielo would be a much better gm for this team then Dubas. He wouldn’t try to be the players friends and use fancy stats to justify stupid contracts and only acquiring players who are under 5’10.

 

[axlrose87]

I honestly thought the first two posts were straight up trolls.

It's one thing to attempt to understand and not be okay with it, but it's another thing to outright be against something that all it does is give you more information, a greater understanding, and helps give you more than evaluating players/teams with points, wins, hits, and +/-.

If you want to know more there are plenty of people on this website or twitter that would gladly explain.

This approach has led to roster without a heartbeat.
Finally, this offseason, analytics were put aside and we went and got players who actually care.
Guys like bogosian, simmonds and Thornton do not produce good fancy stats but find me a leafs fan who isn’t happy we got them.
We actually went against the grain this offseason and the general consensus is that we had a good offseason.

 

[LeafsNation75]

I don’t remember Mark Hunter ever being the gm of the leafs but I would gladly take his approach over Dubas any day. I imagine you were referring to lamorielo. Yes. Lamorielo would be a much better gm for this team then Dubas. He wouldn’t try to be the players friends and use fancy stats to justify stupid contracts and only acquiring players who are under 5’10.

Besides Hunter being their Director of Player Personnel when Brendan Shanahan hired him, he was also named an Assistant GM by Lou Lamoriello. So I brought him up because you mentioned people who previously played and he fits that criteria.

Hell we could even include Brendan Shanahan since he played in the NHL from 1987-2009 and today he's currently the President of the Maple Leafs. After all he did hire Lamoriello and Dubas as GM's.

 

[LeafsNation75]

Yes. Lamorielo would be a much better gm for this team then Dubas. He wouldn’t try to be the players friends and use fancy stats to justify stupid contracts and only acquiring players who are under 5’10.

I guess you forgot Dubas just signed Joe Thornton and last time I checked his height is 6' 4."

John Tavares is 6' 1"

TJ Brodie is 6' 1"

Wayne Simmonds is 6' 2"

Jason Spezza is 6' 3"

Zach Bogosian is 6' 2"

Jake Muzzin is 6' 3"

Justin Holl is 6' 4"

Jimmy Vesey is 6' 3"

Zach Hyamn is 6' 1"

Those are the players Dubas has signed and traded for, who are way over 5' 10."

 

[Morgs]

This approach has led to roster without a heartbeat.
Finally, this offseason, analytics were put aside and we went and got players who actually care.
Guys like bogosian, simmonds and Thornton do not produce good fancy stats but find me a leafs fan who isn’t happy we got them.
We actually went against the grain this offseason and the general consensus is that we had a good offseason.

I stopped reading after "heartbeat"

 

[meefer]

I don’t remember Mark Hunter ever being the gm of the leafs but I would gladly take his approach over Dubas any day. I imagine you were referring to lamorielo. Yes. Lamorielo would be a much better gm for this team then Dubas. He wouldn’t try to be the players friends and use fancy stats to justify stupid contracts and only acquiring players who are under 5’10.

This approach has led to roster without a heartbeat.
Finally, this offseason, analytics were put aside and we went and got players who actually care.
Guys like bogosian, simmonds and Thornton do not produce good fancy stats but find me a leafs fan who isn’t happy we got them.
We actually went against the grain this offseason and the general consensus is that we had a good offseason.

*For the sake of brevity, I'll assume 'stupid contracts' refer to: Tavares, Matthews, Marner, Nylander and Kerfoot. Only one of whom is 5'10" or less. How much do you think Lou could have saved when signing these players? Again, I'll assume you would have preferred keeping Kadri v Kerfoot?
*It only took you until your next post to contradict yourself by mentioning the acquisitions of Bogo, Simmonds and Thornton. You appear to be happy with those acquisitions, should Dubas not receive some praise?
*Some other players that have been added to the team during Dubas' tenure: Anderson, Boyd, Brodie, Engvall, Holl, Muzzin, Spezza, Vesey. Which one of these players is a smurf and on a stupid contract?

I appreciate what Lou did for the team. His contract was for 3 years. Shanny took the view that Kyle would lead the team forward. So here we are. We would all like to see the team succeed, but let's not falsely color what Dubas has done as GM.

Toronto Maple Leafs Hockey Roster | TSN (height and weight, 2 players listed as under 5'10")

 

[meefer]

I guess you forgot Dubas just signed Joe Thornton and last time I checked his height is 6' 4."

John Tavares is 6' 1"

TJ Brodie is 6' 1"

Wayne Simmonds is 6' 2"

Jason Spezza is 6' 3"

Zach Bogosian is 6' 2"

Jake Muzzin is 6' 3"

Justin Holl is 6' 4"

Jimmy Vesey is 6' 3"

Zach Hyamn is 6' 1"

Those are the players Dubas has signed and traded for, who are way over 5' 10."

I take to long to respond to posts, we're essentially saying the same thing.

 

[LeafsOHLRangers98]

Every time I hear the words "fancy stats" together I think of the time Mike Johnson just finished talking about how Nylander was a leader in net front goals last year and then Dave Poulin (who most know used to work for the Leafs) goes off on a tangent saying that he's a perimiter player that needs to start going to the dirty areas to score.

You can't make people understand information if they don't want to hear it. Analytics aren't the be all/ end all but having more information before making a decision is always helpful. If you think that's all the Leafs look at when evaluating players you're dead wrong.

 

[hamzarocks]

The season hasnt even started yet. Can we wait to discuss advance stats and their predictive/feedback value once we get to see how our roster moves impact the team

Every team utilize analytics team. Analytics are growing more common in sports and in regular operations for large scale firms. Need to have balance between analytics and general traditional management techniques.

So far dubas+staff seem to be more aligned and went to make moves which addressed actual roster construction needs. Dubas offseason was the first of his 3 where he got to fix our glaring flaws.

If the results end up speaking for themselve I think the first 2 offseasons can be attributed to nylander/marner/matthews(kind of) limiting his flexibility in addressing needs due to their large contracts demands and high scale negotiations.

We need individuals who can assess talent and understand what is good versus poor performance. Scouting, video scouting, analytic models, off ice training are all part of the assessment process which helps Dubas and staff do their job in maximizing our players performance and working towards the ultimate goal of the cup..

 

[Prominence]

I will be disappointed if the OP is a troll.

TO has hired fancy stat experts, guys who have come up with Corsi, Fenwick and other such stats and so has other teams hired such people. Why hire them when the info they published was commonly available?

It has been addressed by someone else in the thread, but I can repeat it. A company collects data and then sell its to teams who are willing to pay lots of money for it. Data can be interpreted in many ways. The guys who dubas brought in, their job is to analyze it properly and look for inefficiencies (who's an undervalued player, which areas can a team improve, when to rest a player, should we call up Rasmus Sandin, or even which upgrades should the stadium have and etc). We are not hiring them to gather data. We are hiring them to interpret the data and create statistical models.

Also keep in mind that not all NHL teams use the same metrics. Columbus, on Tortorella's request, collects a stat that measures how many passes that defensemen make between each other before they break out. Tortorella does this because he has a theory that if dmen pass the puck too much, it leads to bad things. The point I'm trying to make is that the analytics program for each team is different from other teams. They track different stats because they play differently.

Who are the people publishing Corsi stats and other such stats, how do they have the time to view and evaluate 2,460 games plus playoff games? As far as I know these stats are provided by volunteers who probably are like Leaf fans viewing Leaf games, or Ottawa fans viewing Ottawa games and on and on. Leaf fans have a Leaf bias, Ottawa fans an Ottawa bias so how accurate are these commonly available stats?

They are usually companies who make money off it in some form. John Chayka, former Arizona GM, was co-owner of Stathletes, which is a company that provides hockey analytics. That is one example and to access their data, you need a lot of money. They don't provide this info to fans. Sites like Evolving wild, hockey-reference are great for fans, but most likely useless for other teams. What these companies most likely do is develop and run algorithmns that scan various games and the programs most likely collect the data automatically. Then these companies sell this data for lots of money. It's kind of similar to when you go on facebook, enter your personal information, join a bunch of groups and then facebook sells that info for lots of money to a company like cambridge analytica who then processes this info and sells it to the trump campaign who then uses it to win the election. Same concept. I think.

No, these stats are not provided by leaf volunteers and definitely not from Ottawa fans because they don't exist. Where did you get this idea?

Plus and Minus is a stat that the NHL actually pays someone to ascertain, it's a professionally provided and derived result, it might not be perfect but I think it's the only reliable stat that's available to us, the fans.

Even non-analytics folks don't trust plus/minus. The problem with plus/minus is many. Teams who score a lot more than their opponent will have players with a good plus/minus. They don't even have to be good players to have a good +/-. Brett lebda was a +16 in the 06/07 season. Connor brown was a -70 in his draft year because his team sucked. You can easily have a bad plus minus if your goalie lets in a lot of goals or if the coach sucks or if the d sucks. A lot of players with bad plus/minus may be ones with the toughest defensive assignments. Plus/minus tells us nothing.

TO, Ottawa and other teams have analytics departments, I'm positive teams that have analytics dept. don't trust or rely on what many of we fans consider to be the bible.

I don't know what the hell you mean by the bible. People who look at analytics don't follow the same stuff. Lot of people still follow corsi and fenwick. Lots like WAR/GAR. I actually don't like WAR/GAR. I personally like RAPM even though it overvalues zone starts. I also look at zone data, but that forms a small portion of my analysis. There are a lot of micro-stats ranging from pass data, zone data, there's even some on 50/50 battles. We all like different things.

An example of an amateur or biased results might be hitting, the Islanders have tradionally lead the NHL in the number of hits for the last 10 years(I'm guessing here) because they pad their home stats.

No one who follows analytics looks at hits. Also, I believe hits is collected by the home team. The companies who collect data strive for objectivity as they have algorithms that do it automatically. Where there might be subjectivity is for example, what constitutes a hit, but the data is usually accurate. It's what you do with the data that matters the most.

You seem to think that analytics creates more bias. It's actually the opposite because we can watch the same player but have different opinions. The right analytics and the right analysis cuts through this and provides objectivity.

There seems to also be a misconception that just because a person likes advanced stats that they don't watch the games. What we do is we first watch the players. Then we form some opinions, look at the advanced stuff. We look at where players are proficient and deficient and then we watch the players to see why they produce well/poorly.

Typically good players have good results.

 

[LeafsNation75]

*For the sake of brevity, I'll assume 'stupid contracts' refer to: Tavares, Matthews, Marner, Nylander and Kerfoot. Only one of whom is 5'10" or less. How much do you think Lou could have saved when signing these players? Again, I'll assume you would have preferred keeping Kadri v Kerfoot?
*It only took you until your next post to contradict yourself by mentioning the acquisitions of Bogo, Simmonds and Thornton. You appear to be happy with those acquisitions, should Dubas not receive some praise?
*Some other players that have been added to the team during Dubas' tenure: Anderson, Boyd, Brodie, Engvall, Holl, Muzzin, Spezza, Vesey. Which one of these players is a smurf and on a stupid contract?

I appreciate what Lou did for the team. His contract was for 3 years. Shanny took the view that Kyle would lead the team forward. So here we are. We would all like to see the team succeed, but let's not falsely color what Dubas has done as GM.

Toronto Maple Leafs Hockey Roster | TSN (height and weight, 2 players listed as under 5'10")

When it comes to Lou let's not forget he screwed up not getting William Nylander to sign a contract extension prior to the 2017-18 season or anytime after it started before July 1, 2018 when he became an RFA.

 

[Martin Skoula]

I don’t remember Mark Hunter ever being the gm of the leafs but I would gladly take his approach over Dubas any day. I imagine you were referring to lamorielo. Yes. Lamorielo would be a much better gm for this team then Dubas. He wouldn’t try to be the players friends and use fancy stats to justify stupid contracts and only acquiring players who are under 5’10.

Unfortunately the "Hunter approach" doesn't really work in the NHL since you can't cheat by paying players under the table like in London.

Unless you're talking about his drafting approach, where the best draft pick he made outside of the first round is a #6/7 D at 24. I don't understand how someone who wasn't good enough for EDMONTON's scouting staff is still taken seriously as an NHL scout by anyone. Either one of Dubas' 2018 or 2019 drafts on its own is likely to produce more NHL games than Hunter will draft outside the 1st in his lifetime.

 

[Mickey Marner]

There are 1271 games in an 82 game season, not 2460. Only 868 in a 56 game season. Plus ~82-83 in the playoffs.

Their data collection is proprietary, not the stuff the public has at their disposal. It would be information that provides the 'why' and 'how', not the 'what'.

 

[Joey Hoser]

Yeah, I too hate it when facts get in the way of the opinions I want to have

 

[ACC1224]

Shanahan hiring Dubas shows how keen an eye he has for talent.

He also hired Lou and Hunter.

 

[BertCorbeau]

I like how we live in a world where there are just "hockey guys" and "nerds who use stats to understand hockey"

With nothing in between. No overlap whatsoever.

:eyeball:

 

[Nineteen67]

Watch a game while tracking the fancy stats listed in the OP, there is no way in hades any nhl team is using those for anything.

 

[Fogelhund]

So... I don't understand it, so I don't like it... is that basically the theory here? Flat Earth, Anti-Vax thread essentially? and this 5'10" stuff... y'all crack me up. This is a thread that should just be closed.