News Article: "Fun With Numbers" - Advanced Stats Talk Here

[Vesa Awesaka]

And yet some people still point to goalkeeping as a reason for Cameron's success...

overall Cameron has had worse save percentage then Maclean. Still both have recieved top ten save percentages. 5th for maclean 8th for Cameron

 

[Sens Rule]

19-1-2. Advanced stats mean crap with Hammond...he just wins!

Go Hamburglar Go!

 

[Vesa Awesaka]

19-1-2. Advanced stats mean crap with Hammond...he just wins!

Go Hamburglar Go!

Well... Unless Hammond gonna score some goals or start passing the puck... lol Seriously the sens have been an above average team without hammond. Hammond just takes that above average team and makes them 19-1-2. Which is godlike. What hammond does is throw out any good advanced stats for the other team i guess lol

 

[Sens Rule]

Well... Unless Hammond gonna score some goals or start passing the puck... lol Seriously the sens have been an above average team without hammond. Hammond just takes that above average team and makes them 19-1-2. Which is godlike. What hammond does is throw out any good advanced stats for the other team i guess lol

Hammond is Thor, Spider-Man and Captain Canada.


Who cares about stats of super heroes?

 

[Vesa Awesaka]

Hammond is Thor, Spider-Man and Captain Canada.


Who cares about stats of super heroes?

Captain Canada never wins anything he just compromises

 

[Sens Rule]

Captain Canada never wins anything he just compromises

He says sorry...after he beats your ass!

Polite ass-kicker. So Canadian.

 

[alfie follower]

overall Cameron has had worse save percentage then Maclean. Still both have recieved top ten save percentages. 5th for maclean 8th for Cameron

hammonds sv% is .939, so I guess, the other 2 guys are what's brought it down ? or was Anderson higher than .939 under maclean ?

 

[BonkTastic]

hammonds sv% is .939, so I guess, the other 2 guys are what's brought it down ? or was Anderson higher than .939 under maclean ?

Anderson was ridiculously good during Maclean's tenure this year. He was rocking a Sv% of ~.930 until he got hurt.

 

[alfie follower]

Anderson was ridiculously good during Maclean's tenure this year. He was rocking a Sv% of ~.930 until he got hurt.

that's what i've heard, so I'm trying to figure out why the sv% under Cameron could possibly be lower than under Maclean. It doesn't seem to make sense

 

[BonkTastic]

that's what i've heard, so I'm trying to figure out why the sv% under Cameron could possibly be lower than under Maclean. It doesn't seem to make sense

He got hurt, then came back from injury at less than 100% and wasn't as effective.

Also: the team's Sv% under Cameron featured most of Lehner's starts, so Cameron's Sv% record certainly isn't helped out by Lehner's hot garbage 9.05 sv%.

 

[SoCalSens]

Well, it's official: the team has acknowledged the existence of Advanced Stats.

Article can be found here, by Chris Lund:
Fun With Numbers: A Stats Rundown

It's a pretty good read if you're not familiar with the terms, worth the 3 mins it'll take to skim through it.

There's no figures. How can you write an article about stats without figures?

 

[BonkTastic]

There's no figures. How can you write an article about stats without figures?

Heck of a bump there, my friend.



To answer your question - the article ACKNOWLEDGES advanced stats - that they exist. That was a pretty big deal 9 months ago, when the article was released and I wrote that post, haha.

The org had never even so much as brought them up in an interview before that. At the time, it was baby steps.

 

[Micklebot]

Team shot quality against data

Playing around with numbers from WaronIce;

Table shows teams Low quality, medium quality, and high quality chances against per team (data is a bit skewed for teams that traded goalies, I'm looking at you Buffalo, and omits goalies with <240 mins). It also shows % of total shots for each quality. And finally, High quality shots per 60, and high and Medium quality shots combined per 60.

Interesting to note that while we give up a lot of shots, were are keeping them to lower quality chances. Once you look at the per 60, we come out about average despite the high shot volume.

Team | TOI | L | M | H | %L | %M | %H | H/60 | HM/60
FLA | 3963.7 | 942 | 476 | 435 | 50.84% | 25.69% | 23.48% | 6.58 | 13.7901455711
NSH | 3874 | 909 | 453 | 418 | 51.07% | 25.45% | 23.48% | 6.47 | 13.4899328859
OTT | 3851.1 | 995 | 538 | 494 | 49.09% | 26.54% | 24.37% | 7.70 | 16.0785230194
BUF/DAL | 1140.5 | 291 | 179 | 156 | 46.49% | 28.59% | 24.92% | 8.21 | 17.623849189
BOS | 4008.2 | 891 | 559 | 489 | 45.95% | 28.83% | 25.22% | 7.32 | 15.6878399281
MIN | 2010.2 | 448 | 241 | 235 | 48.48% | 26.08% | 25.43% | 7.01 | 14.2075415382
ANA | 3805.7 | 855 | 455 | 453 | 48.50% | 25.81% | 25.69% | 7.14 | 14.3153690517
L.A | 3977.8 | 762 | 501 | 440 | 44.74% | 29.42% | 25.84% | 6.64 | 14.1937754538
NYR | 3915.8 | 960 | 477 | 509 | 49.33% | 24.51% | 26.16% | 7.80 | 15.1080239032
DET | 3768.8 | 790 | 449 | 439 | 47.08% | 26.76% | 26.16% | 6.99 | 14.1371258756
MTL | 3938.7 | 912 | 502 | 502 | 47.60% | 26.20% | 26.20% | 7.65 | 15.2943864727
STL | 3948.1 | 760 | 524 | 457 | 43.65% | 30.10% | 26.25% | 6.95 | 14.9084369697
CBJ | 3674.8 | 907 | 514 | 510 | 46.97% | 26.62% | 26.41% | 8.33 | 16.7192772396
N.J | 3893.2 | 842 | 530 | 501 | 44.95% | 28.30% | 26.75% | 7.72 | 15.8892427823
WPG | 3825.8 | 779 | 471 | 460 | 45.56% | 27.54% | 26.90% | 7.21 | 14.6008677924
WSH | 3873.7 | 836 | 442 | 482 | 47.50% | 25.11% | 27.39% | 7.47 | 14.3118981852
PHI | 3937.3 | 933 | 447 | 536 | 48.70% | 23.33% | 27.97% | 8.17 | 14.9798084982
S.J | 3904.9 | 851 | 501 | 527 | 45.29% | 26.66% | 28.05% | 8.10 | 15.7955389383
DAL | 2884.7 | 622 | 409 | 403 | 43.38% | 28.52% | 28.10% | 8.38 | 16.8891045863
COL | 3962.1 | 962 | 559 | 597 | 45.42% | 26.39% | 28.19% | 9.04 | 17.5058681002
ARI | 3094.9 | 709 | 441 | 457 | 44.12% | 27.44% | 28.44% | 8.86 | 17.4092862451
VAN | 3861.6 | 833 | 493 | 538 | 44.69% | 26.45% | 28.86% | 8.36 | 16.0192666252
CAR | 4051.7 | 781 | 508 | 523 | 43.10% | 28.04% | 28.86% | 7.74 | 15.2676654244
CHI | 4002.2 | 880 | 529 | 580 | 44.24% | 26.60% | 29.16% | 8.70 | 16.6258557793
CGY | 4023.9 | 868 | 530 | 579 | 43.90% | 26.81% | 29.29% | 8.63 | 16.5361962275
EDM | 4007.3 | 827 | 549 | 572 | 42.45% | 28.18% | 29.36% | 8.56 | 16.7843685274
PIT | 3838.2 | 773 | 454 | 516 | 44.35% | 26.05% | 29.60% | 8.07 | 15.163357824
TOR | 3860.2 | 873 | 595 | 626 | 41.69% | 28.41% | 29.89% | 9.73 | 18.9782912802
T.B | 3872.5 | 713 | 462 | 512 | 42.26% | 27.39% | 30.35% | 7.93 | 15.0910264687
NYI | 3733.9 | 699 | 480 | 541 | 40.64% | 27.91% | 31.45% | 8.69 | 16.4064383085

 

[Vesa Awesaka]

Hammond has like a 999 save percentage on low quality shots. Similar to what anderson had in the shortened season.

 

[Micklebot]

Hammond has like a 999 save percentage on low quality shots. Similar to what anderson had in the shortened season.

He's actually around 97.7, Anderson is 97.4. Lots of goalies are higher, including Rask, Price Niemi, Dubnyk, Schnieder, Rinne, Pavalec, Quick, Fleury, Crawford, and Mason.

What set's Hammond apart is his High Quality sv% which is tops among goalies with 1000 mins played.

 

[Vesa Awesaka]

He's actually around 97.7, Anderson is 97.4. Lots of goalies are higher, including Rask, Price Niemi, Dubnyk, Schnieder, Rinne, Pavalec, Quick, Fleury, Crawford, and Mason.

What set's Hammond apart is his High Quality sv% which is tops among goalies with 1000 mins played.

my mistake. but all situations hammond is 7th for goalies on low quality shots. I guess i used only used a small range when i first did a look at the goalies.

anderson is actually better on medium quality shots which is kinda interesting

 

[StefanW]

Raw save % is, IMO, alongside plus minus as the worst stat in hockey. I like some of the things that are being done right now. The danger zones at War on Ice have been an excellent addition, but they cannot track the motion of the puck. So a fast pass across the slot followed by a one-timer is much more dangerous that a shot against a goalie who is set and square to the puck.

There is an ongoing data tracking project that tracks passes and how they influence shots and saves, and a presentation by the lead was made at the DC analytics conference this past weekend. Here is the talk in case anyone else is interest. I find this stuff amazing, so it was my favourite of the conference.

 

[Micklebot]

Raw save % is, IMO, alongside plus minus as the worst stat in hockey. I like some of the things that are being done right now. The danger zones at War on Ice have been an excellent addition, but they cannot track the motion of the puck. So a fast pass across the slot followed by a one-timer is much more dangerous that a shot against a goalie who is set and square to the puck.

There is an ongoing data tracking project that tracks passes and how they influence shots and saves, and a presentation by the lead was made at the DC analytics conference this past weekend. Here is the talk in case anyone else is interest. I find this stuff amazing, so it was my favourite of the conference.

Not sure if it's incorperated into the High Medium and low danger zone data or just adjust afterwards, but at least for their scoring chance date, they incorperate whether the shot is off a rebound (any shots within a certain timeline of a save, block or miss) or off the rush (shot within 10 seconds of: a shot attempt by the other team on the other net; a face off at the other end or in the neutral zone; or a hit, giveaway or takeaway in the other end or the neutral zone.)

 

[SoCalSens]

Heck of a bump there, my friend.



To answer your question - the article ACKNOWLEDGES advanced stats - that they exist. That was a pretty big deal 9 months ago, when the article was released and I wrote that post, haha.

The org had never even so much as brought them up in an interview before that. At the time, it was baby steps.

Oops, just looked at the article and didn't see any pretty pictures. Didn't notice the length of the thread or how old it was. Sorry!

 

[Vesa Awesaka]

Just curious if anyone knows anything about point share. Is it accurate? What is your opinion on it?

http://www.hockey-reference.com/teams/OTT/2015.html

http://www.hockey-reference.com/about/point_shares.html

 

[Micklebot]

Just curious if anyone knows anything about point share. Is it accurate? What is your opinion on it?

http://www.hockey-reference.com/teams/OTT/2015.html

http://www.hockey-reference.com/about/point_shares.html

What exactly do you mean by is it accurate? Point shares are very strongly correlated to team strength, in that if you add up all the point shares of a team's players, you'll be very close to the number of points the team got. This makes it hard for a good player on a poor team to look good relative to an equally good player on a great team.

So, if by accurate, you mean how much did that player contribute to his team's success, sure, it's probably pretty decent. But if you want to compare players across teams, or across years, I'm not so sure. It'll probably get you in the right ballpark, but I would go comparing Norris winners to see which winner had the best season by a defense ever (well, that might be a poor example, as it would probably come up Orr and still be right).

 

[Vesa Awesaka]

What exactly do you mean by is it accurate? Point shares are very strongly correlated to team strength, in that if you add up all the point shares of a team's players, you'll be very close to the number of points the team got. This makes it hard for a good player on a poor team to look good relative to an equally good player on a great team.

So, if by accurate, you mean how much did that player contribute to his team's success, sure, it's probably pretty decent. But if you want to compare players across teams, or across years, I'm not so sure. It'll probably get you in the right ballpark, but I would go comparing Norris winners to see which winner had the best season by a defense ever (well, that might be a poor example, as it would probably come up Orr and still be right).

Accurate in determining player value to that particular season. Is defensive point share a proper measure or defensive contribution? Is offensive point share a proper measure of offensive contribution?

The first player i was looking was James Wisneski. Anaheim is healthy scratching him in the playoffs. Last year was one of his best point share wise this year is one of his worst defensive point share years.

 

[Busboy]

This isn't technically "Advanced stats," but I'm developing an infographic exploring "sophomore slumps." Essentially I'm compiling data from 2005 onwards to see if there is any statistical merit to the idea that "sophomore slumps" are common or if there is a consistent trend. I'm curious to see what we can learn from the data. I have some suspicions but I'm holding off until I can start playing around with the full set of data. From there I will develop a design that articulates the findings with some visual flair.

I've finished collecting data for the first round of rookies to compare against their sophomore seasons and now I have another 8 seasons to go. I'm wondering if there's anyone who would like to help out with data wrangling/analysis? I'm working through it all but it's a bit tedious and time consuming at this point.

The project is simply for fun. Anyone who would like to help will get full credit in the final piece. I'm also open to sharing raw data (excel sheets) for others to play with and offer input/analysis. If you'd like to help out then send me a PM and I can send you what I've got so far!

 

[StefanW]

This isn't technically "Advanced stats," but I'm developing an infographic exploring "sophomore slumps." Essentially I'm compiling data from 2005 onwards to see if there is any statistical merit to the idea that "sophomore slumps" are common or if there is a consistent trend. I'm curious to see what we can learn from the data. I have some suspicions but I'm holding off until I can start playing around with the full set of data. From there I will develop a design that articulates the findings with some visual flair.

I've finished collecting data for the first round of rookies to compare against their sophomore seasons and now I have another 8 seasons to go. I'm wondering if there's anyone who would like to help out with data wrangling/analysis? I'm working through it all but it's a bit tedious and time consuming at this point.

The project is simply for fun. Anyone who would like to help will get full credit in the final piece. I'm also open to sharing raw data (excel sheets) for others to play with and offer input/analysis. If you'd like to help out then send me a PM and I can send you what I've got so far!

Interesting. How are you dealing with survivorship bias? A lot of players only get one season in their career.

 

[Busboy]

Interesting. How are you dealing with survivorship bias? A lot of players only get one season in their career.

I intend not to include players who play only one season. Since the premise is "sophomore slump" I don't believe they will be relevant to the topic.

I will also use a cutoff (probably around 25 points scored in rookie season, less for defenseman) since the premise of a "sophomore slump" is mostly predicated on a decrease in production. If a rookie is scoring less than that then a decrease in points is likely to be insignificant. Obviously an increase in points would be relevant since the more players performing above their rookie season there is the more the idea of a "sophomore slump" starts to crumble but there's only so much time I can devote to compiling data.

Currently I'm including GP, G, A, P, PP/G, +\-, TOI only. The main trends I will be looking at are differences in PP/G and TOI. There is obviously more I could track but I'd like to get a firm grasp on these trends before digging deeper into the stats.