Ideas for Future Studies

[abo9]

I just saw the reverse stats on Stats from the reverse end , wondering if there is a website that lists injuries for players and if being hit more often or hitting often correlates to being injured more often / shorter careers.

I expect to see shorter careers or big drop in performances with age for players giving or receiving a lot of hits.

For example I see a guy like Toews being 11th of active players, and with his long contract in place I wonder how receiving so many hits will affect his longevity as a player. I don't want to say un-informed statements, but from the players in range 1-10 you can find Letang (who's been injured lately?), Seidenberg (who got injured last year), Dustin Brown and Seabrook whose numbers have been slowing down (although Brown was pretty good last year).

 

[brachyrynchos]

The consistency aspect is something I'd like to see. While goals/assists/points per game is a good stat, what about % of games with a goal,assist, and point. Or at least fewest consecutive games without a point avg? I mean, a player might have a higher gpg avg but scores in bunches, while another player would have a lower or similar gpg but scores on a more consistent basis and is more dependable.

 

[abo9]

The consistency aspect is something I'd like to see. While goals/assists/points per game is a good stat, what about % of games with a goal,assist, and point. Or at least fewest consecutive games without a point avg? I mean, a player might have a higher gpg avg but scores in bunches, while another player would have a lower or similar gpg but scores on a more consistent basis and is more dependable.

I might give that a shot for a couple stars to begin with, sounds interesting for sure

 

[morehockeystats]

I might give that a shot for a couple stars to begin with, sounds interesting for sure

I think it's more important if the player scores when the game is close, and not in garbage time. Otherwise, as long as the average is the same, there shouldn't be much difference in dependability. And usually you want to remove the EN goals/points from any comparison, especially for the streaky players, because sometimes a team would work together to give a player an EN goal for the hat-trick.

 

[Rakuten06]

Hi, hum, people!

I was wondering if there was stats about the probability to get NHL drafted and the round/rank you are drafted in each of the 3 leagues of the CHL.

I happen to know somebody drafted in the second round of the QMJHL.

Thanks in advance!

I'm new to hockey analytics, but I think this would be a good project for me to work on. I checked EliteProspects, only issue about it is that each leagues (WHL, OHL, QMJHL) have different rounds and some teams tend to pass on their picks, some teams picked players right after the passes and so on. I plan to start with 1990 since that's the first year WHL has a complete draft on EP that continues to today (one before in 1973, that's all).

 

[OvermanKingGainer]

I would like to see a study of recent fathers (0-8month old babies) and if there is any correlation to a difference in home performance vs road. It's a vague idea and I have no clue how to qualtify it, but maybe someone here has a place to start?

 

[morehockeystats]

I would like to see a study of recent fathers (0-8month old babies) and if there is any correlation to a difference in home performance vs road. It's a vague idea and I have no clue how to qualtify it, but maybe someone here has a place to start?

Among chess players there's a well-known semi-joke that getting married is usually followed by a drop of about 50 Elo rating points.

 

[PepsiCenterMagic]

Something I've been wanting to look into for a while is building a good machine learning model for predicting season point totals per player. There are many other things one could also build and analyze, but I wanted to start here first.

I haven't really put any time into ideating a what a good data structure would be for this problem, nor where a good data source lies (thought hockey-reference.com has good downloadable sets), but I think it would be fun to run through a good model tuning and building process.

I'm sure they probably do this and have been doing this for fantasy hockey (though honestly even with high traffic volume, I'm not sure if websites have been as complex as to machine model iterate), but out of curiosity of wanting to compare if I start this project, has anyone already dived into this kind of analysis, here or elsewhere? And would anyone be interested if I undertook this for fun? If there was anyone really interested, it would motivate me more.

 

[goeb]

Not sure if this has been suggested or already completed but....

I am interested in scoring percentages based on shot type (wrist shot, slapshot, backhand, tip-in, wrap shot, etc) split out by team and player. I am wondering if there are any reliability year to year among those for teams or players.

I also would be curious to know the save % for those shots for each goalie.

As I get more conversant with python, I might try to tackle this myself but not sure when that would be or how difficult it will be.

Thanks!

 

[morehockeystats]

Not sure if this has been suggested or already completed but....

I am interested in scoring percentages based on shot type (wrist shot, slapshot, backhand, tip-in, wrap shot, etc) split out by team and player. I am wondering if there are any reliability year to year among those for teams or players.

I also would be curious to know the save % for those shots for each goalie.

As I get more conversant with python, I might try to tackle this myself but not sure when that would be or how difficult it will be.

Thanks!

Should be fairly easy.
Download the live JSON, poll for events of type GOAL, SHOT and possibly MISSED_SHOT. All carry a shot type property, at least since 2007 (sorry, MISSED_SHOT only has the property set in the HTML PBP). All carry the Shooter and the first two carry the Goalie property.

I have all the data and I just deployed an API and need some testers, so you can try to be one and use all data from one place.

 

[goeb]

Should be fairly easy.
Download the live JSON, poll for events of type GOAL, SHOT and possibly MISSED_SHOT. All carry a shot type property, at least since 2007 (sorry, MISSED_SHOT only has the property set in the HTML PBP). All carry the Shooter and the first two carry the Goalie property.

I have all the data and I just deployed an API and need some testers, so you can try to be one and use all data from one place.

Sweet. Do I just apply through your site to test the API? How much of a donation would be appropriate in this case?

 

[morehockeystats]

Sweet. Do I just apply through your site to test the API? How much of a donation would be appropriate in this case?

Try to read https://morehockeystats.com/data/api if it makes any sense to you. Email me or message me on Twitter if you're interested.

 

[PepsiCenterMagic]

Something I've been wanting to look into for a while is building a good machine learning model for predicting season point totals per player. There are many other things one could also build and analyze, but I wanted to start here first.

I haven't really put any time into ideating a what a good data structure would be for this problem, nor where a good data source lies (thought hockey-reference.com has good downloadable sets), but I think it would be fun to run through a good model tuning and building process.

I'm sure they probably do this and have been doing this for fantasy hockey (though honestly even with high traffic volume, I'm not sure if websites have been as complex as to machine model iterate), but out of curiosity of wanting to compare if I start this project, has anyone already dived into this kind of analysis, here or elsewhere? And would anyone be interested if I undertook this for fun? If there was anyone really interested, it would motivate me more.

Actually I've realized I just probably don't have the time I need in order to do this.

 

[kxx]

Hey everyone - just a quick question regarding the following article.
https://thehockeywriters.com/future-of-hockey-analytics/
Mainly the line:
"Answering these questions, as Alison Mah points out, will be easier with the advent of new technology. Teams are exploring placing tracking chips into pucks and players’ jerseys to measure aspects of the game that were previously not measurable. Infrared cameras installed around the arena will track everything from skating speed to puck movement."

(I think there was a similar article in the Athletic not too long ago as well)

I'm curious to know if anyone knows the status of these puck and player trackers - this seems like it will completely change the analytics game (or at least open up a new dimension to start exploring) Would love to know more information about it.

 

[Chriscftb97]

I would like to see a study of recent fathers (0-8month old babies) and if there is any correlation to a difference in home performance vs road. It's a vague idea and I have no clue how to qualtify it, but maybe someone here has a place to start?

I actually wrote about this a couple years ago.

Dad Strength: Myth or Miracle?

 

[xtra]

I got a question; has there been a study on the relation between corsi for and injuries?

Are teams with a better corsi less injured or not- is there any correlation at all etc?

 

[shylock104]

I was perusing on natural stattrick trying to figure out if a better predictor for goalie stats would be H/M/LDSV. Could never figure out how to extrapolate the numbers but I think it would be really interesting to see if those are better indicators of overall performance or if certain DSVs are more luck driven than others, how age correlates with it, etc.

 

[Runforthecup]

Question for statisticians.

Until 92-93 the head of the NHL was called the President. In the Spring of 93 the Montreal Canadians won the Stanley Cup. Gil Stein retired on July 1, 1993

Gary Bettman became the first Commissioner in 93-94. After this spring, from that time forward American teams have won 26 consecutive Stanley Cups. Given the varying number of Canadian Teams in the league over the period what is the probability of this happening?

 

[Runforthecup]

Question for statisticians.

Until 92-93 the head of the NHL was called the President. In the Spring of 93 the Montreal Canadians won the Stanley Cup. Gil Stein retired on July 1, 1993

Gary Bettman became the first Commissioner in 93-94. After this spring, from that time forward American teams have won 26 consecutive Stanley Cups. Given the varying number of Canadian Teams in the league over the period what is the probability of this happening?

 

[Runforthecup]

Okay. So answering my own question there is a 0.213% chance this could occur on a random basis. Once every 469 years if you chose teams on a random basis. I will have to examine the possible reasons for this at a later date (less scotch involved).

 

[Buffaloed]

Create a database table of statistics available for every player for every year by coach. From this you can select players who played for the same coach on different teams and compare stats with different coaches. The goal is to develop an algorithm to be able to objectively rate coaches. We'll call this the Howdy Doody Number (named after Phil Housley). Maybe something like this already exists? It's crazy that players have every stat imaginable analyzed to determine their value and a coach can talk his way into a job by being able to give great interviews.

 

[Doctor No]

Create a database table of statistics available for every player for every year by coach. From this you can select players who played for the same coach on different teams and compare stats with different coaches. The goal is to develop an algorithm to be able to objectively rate coaches. We'll call this the Howdy Doody Number (named after Phil Housley). Maybe something like this already exists? It's crazy that players have every stat imaginable analyzed to determine their value and a coach can talk his way into a job by being able to give great interviews.

I've started doing this for goaltenders, although it's currently in the "curiosity" stage. The intention is to include the coach as a predictive variable for future seasons' statistics.

 

[Mashed Potatoes]

Actually I've realized I just probably don't have the time I need in order to do this.

Did you end up doing the ML model in the end?

 

[Uncle Rotter]

Has the effect of (for lack of a better phrase) "equipment malfunction" on scoring or game results has ever been studied? To name just three recent examples, 2019 WCH (GWG resulting from a lost stick), 2019 WJC QF between Canada and Finland (GWG after a broken stick), 2018 Olympic Final (tying goal as a result of a helmet coming off).
Edit: more in the past, the 2006 Olympic Final (GWG scored after a broken stick.)

 

[Meatball]

I had this idea about which teams/organizations get the most production from their drafted/homegrown players and wondered if that could be measured somehow.