Oilers analytic/advanced stat thread

[Aceboogie]

You're using objective data (not even good data necessarily) to draw subjective conclusions about how well a player is/isn't playing.

If you dont like the use of data and shot metrics and the conclusions draw than you never had to click in here to begin with

The stats objectively show RNH is getting off more shots than ever before and also more high quality shots then ever before. My subjective conclusion based on years of playing hockey and watching hockey and following the numbers and using past comparables is that in order to get a high number of shots away you have to be playing well. Not many players can stumble in to be top of league for shots with poor play

There is a number of players who have had similiar attributes to RNHs cold streka right now (high number of shots, high number of high quality chances, history of being a good player with above average shooting %) and all broke out of it

Like I said before, I wish I could buy stocks in players like I could with corporations. I was actually Kadris biggest defender last year. Even making this thread on main boards with the EXACT premise: http://hfboards.mandatory.com/showthread.php?t=1985381&highlight=kadri

Not even 2 months later he started to heat up, then continued it to this year. 10 goals in 24 goals, seen as one of leagues better 2 way players. I wish I could have bought shares in Kadri at that time. Could have been rich.

read that thread, all your arguments and answers regarding RNH will be in there. Just change Kadri to RNH

 

[McDraekke]

I think the OPs use of analytics is faulty and biased.

I'm allowed to say that. You don't like it, feel free to put me on ignore. Neither you nor the OP get to dictate who can/cannot respond to this thread.

You're allowed to express your opinion, but it makes someone kind of an ******* if you feel the need to continue to post your opinion to the detriment of the thread.

Never once did I say you can't post in here. I said you shouldn't. Very different.

 

[Aceboogie]

I think the OPs use of analytics is faulty and biased.

I'm allowed to say that. You don't like it, feel free to put me on ignore. Neither you nor the OP get to dictate who can/cannot respond to this thread.

Feel free to post your own use of analytics to argue against it. Dont just scream wrong and then make people cater to you.

How are the stats wrong exactly? How are they biased?

I used the EXACT stats to defend a friggin Maple Leafs player whom I despise and low and behold Kadri turns it around big time (not saying I was the cause, just that shooting % corrected itself). Read Leafs opinions on Kadri in that thread, read them now. He was dirt last year up till then, hes a god now. Analytics are the same,Shots are the same, play is largely the same. Shooting % is only deifference

You can post in any thread you want, this is the internet and I am not the internet police. This thread was intended to be a space to discuss analytics and disagreements are encouraged. But to come in here just to slag on analytics is the definition of a debbie downer. Either post why you think they are wrong or dont click in here if your going to be triggered. You are also free not to click into a thread...

 

[Spawn]

You're allowed to express your opinion, but it makes someone kind of an ******* if you feel the need to continue to post your opinion to the detriment of the thread.

Never once did I say you can't post in here. I said you shouldn't. Very different.

Lol

 

[Aerrol]

I don't have much to add as I only loosely follow hockey analytics, but I just want to say that I appreciate the thread idea greatly. It'll be a good reference tool for later discussions as well.

 

[McGoMcD]

what is the best site for these stats? I recall behind the net back in the day was really good.

 

[Aceboogie]

what is the best site for these stats? I recall behind the net back in the day was really good.

Good question

Behind the net really has fallen off. Hasnt really been updated for this year. Was good for corsirel and quality of comp

Hockey Analysis is one of my favourites due to ease of use and easy to navigate

Corsica is the grand daddy of them all but it can get overwhelming. It has a TON of info

Puckalytics is good too

 

[McDraekke]

I don't have much to add as I only loosely follow hockey analytics, but I just want to say that I appreciate the thread idea greatly. It'll be a good reference tool for later discussions as well.

Much agreed.

Good work Aceboogie. I've been thinking about keeping my own set of stats for a while as well, but I've never given myself the time to do it. Unlike a lot of fans still, I believe that stats are a very important part of the game that people have a hard time being able to see with the "eye test" due to the speed of the game, and many other factors.

People keep bashing a number of our players (particularly RNH and Eberle right now...), but if you break things down to stats you'll see that even though they may not "look" to be playing that well, the truth is not all what you see. People calling for Eberle's head is my biggest pet peeve. We all know he's a defensively lacking skilled winger. He's top 30 in scoring (or was when I made my post in the Eberle Must Go thread a week ago) this year and in his career... what are people expecting of a defensively lacking skilled winger? Yea, he has pretty bad defensive numbers, but that is to be expected, much like a lot of other top line wingers, oddly enough.

 

[Spawn]

K here's a legitimate question rather than a "Nuge sucks posts." I'll admit I have a bias against that player that can colour my perception on him at times.

What are the qualifications of a high danger scoring chance? I question that being able to be an objective stat that can be accurately kept. I imagine it has to do with distance from the net when shot. But that is not really an indication of how dangerous a scoring chance is is it?

Maybe it's more elaborate than that.

 

[Aceboogie]

Much agreed.

Good work Aceboogie. I've been thinking about keeping my own set of stats for a while as well, but I've never given myself the time to do it. Unlike a lot of fans still, I believe that stats are a very important part of the game that people have a hard time being able to see with the "eye test" due to the speed of the game, and many other factors.

I still think a mix of the 2 together is best. But I do personally have some issues with people saying "eye test is all that I need". I have issues with it because:
a) The game moves so fast and with 5 players on the ice (often 3 involved in play at once) its hard to properly judge a single player. People miss a lot of stuff from watching a game. If you isolate one player each shift then its better. But I dont think many people strickly iso one player. So much is missed

b) Even if you try and scout a single player we have so much bias in on our judgement that we dont even realize. A semi big hit will leave fans amazed and I think they overvalue that. People tend to overvalue things and undervalue other aspects of the game. This is why some players have gotten fat contracts from head offices and then completely failed after. Or other players have come out of nowhere

c) I dont think people are all that good of scouts truth be told. I think we are adequate but limited. Myself included, I dont think people have the best trained eye for stuff. Maybe they do but just dont post it. One of rare people this applies to is LMFH, his post game reports are insane and very detailed. Hes got a great eye and id take his scouting report over anything. Even sometimes he make me rethink my own view of a players performance

I equate analytics in hockey to our views self driving cars. I think people overvalue their own ability to judge hockey as we overvalue our ability to drive. We think we are really good. But then self driving cars (analytics) come along and challenges our traditional views and people get upset their word is being challenged. For self driving cars- people saw 1 crash with a self driving car (after hundreds of thousands of miles of safe driving) and then pointed to these cars as being totally wreckless and dangerous. Yet himans cause millins of accidents and thousands of death each year, but because people overestimated their own skills they think they are much better. Even self drivign cars are much safer per mile driven. Thats my feeling on analytics, people become challenged and then say "Oh well analytics says this particular player is great but IMO he sucks bad so all analytics are flawed" and ignore all the misses of "eye tests" in the past

Neither approach on its own is great. Both together is good

 

[Aceboogie]

K here's a legitimate question rather than a "Nuge sucks posts." I'll admit I have a bias against that player that can colour my perception on him at times.

What are the qualifications of a high danger scoring chance? I question that being able to be an objective stat that can be accurately kept. I imagine it has to do with distance from the net when shot. But that is not really an indication of how dangerous a scoring chance is is it?

Maybe it's more elaborate than that.

Yes distance from net. As of right now their is no feasible way to measure the quality of a shot otherwise. When we get tracking chips in pucks we can better measure shot speed/deflections etc. But shot location is used for high danger chances. Shots from in close are more dangerous and also harder to get to in the first place (getting to middle of slot).

Yes there is the times where a puck from the slot is deflected or misshot so its marked as a high danger chance even tho it is really not dangerous. But also there is similar # of occurances where a player gets off a completely clean shot from that area. In the middle is the regular shot, some distortion from defenders but still a strong shot

To prove these areas actually do render more dangerous shots. They used goalies save % from different areas on the ice and found the save % was much lower in these high danger chances than in others.

 

[Speed220DChalavan]

Matt Benning

But Benning right now is one of the best possession D in the NHL. Here is where he ranks amongst regular D in different areas:

CF%: 2nd in NHL
FF%: 6th in NHL
SF%: 14th in NHL
GF%: 9th in NHL
SCF%: 1st in NHL
HDCF%: 5th in NHL
Playing really easy minutes by analytics are Muzzin-esque

He will regress a bit I suspect, but what a find. Moves the puck north quickly and effectively. Local boy too.

He's one of our own,
He's one of our own,
Matthew Benning,
He's one of our own ...
(Repeated...)

 

[Aceboogie]

He will regress a bit I suspect, but what a find. Moves the puck north quickly and effectively. Local boy too.

He's one of our own,
He's one of our own,
Matthew Benning,
He's one of our own ...
(Repeated...)

Oh yeah absolutely. The guy right near him in those rankings? Justin Schultz. Playing well on third pairing in Pitt

 

[Paralyzer008]

Matt Benning

But Benning right now is one of the best possession D in the NHL. Here is where he ranks amongst regular D in different areas:

CF%: 2nd in NHL
FF%: 6th in NHL
SF%: 14th in NHL
GF%: 9th in NHL
SCF%: 1st in NHL
HDCF%: 5th in NHL
Playing really easy minutes by analytics are Muzzin-esque

As much as I don't believe that Benning has much more of a ceiling, when I saw the analytical darling LA Kings going after him, I had a good feeling about him.

He looks pretty solid as far as rookie D go.

 

[duul]

So what you're telling me is that Matt Benning is a really really good defenceman, and Kris Russell is a really really bad defenceman?

 

[Aceboogie]

So what you're telling me is that Matt Benning is a really really good defenceman, and Kris Russell is a really really bad defenceman?

Not at all. Matt Benning is doing well in his role. Russell is doing well in his role which is in harder minutes. The stats point to Russell being a really good 3rd pairing D ideally, the eye says he can handle 2nd pairing minutes

Ive actually really budged on Russell and I enjoy him on this team. But I see where his poor analytics come from. He gives up the blueline almost every rush to avoid getting burnt wide. This prevents those type of breakaways/rushes but then allows the forward time and space to make a play or get a shot off from further. More shots against him are allowed by he makes up for it by blocking a ton of them. He allows a lot of low and mid danger shots but prevents high danger ones on the rush

 

[Aceboogie]

After game ill post WheatnOils game tracking numbers. This twitter user watchs each game pretty closely and manually tracks a ton of data for the D. So everybody should like these figures as it invovles watching the game over and over and manually tracking

Here is an example:
http://www.theoilersrig.com/2016/11/oilers-vs-coyotes-tracking-stats/

Here is one of the types of data they track- zone entries. This game was randoly chosen but oddly aligns with my comment above. Russell was targeted on the rush 11 times (they broke into the zone on his side). 0 times did he prevent them from coming in. 36% of time they had control coming in, the rest was dumps in on his side

There is zone entry/ zoen exit, puck retrievals/ D zone passing

 

[phaedrusDH]

Kris Russell:

1st in Oilers 5v5 for time of possession per game at 1:27 (per Mike Kelly)

At 45.54% Corsi, he's -7.1% below the average of the rest of the team and is clearly in last place among Oiler D. In fact, the next worst is Adam Larsson, who is above 50% and only about -1.8% below the average.

So, the puck is on Kris Russell's stick a lot but the Oilers are getting outshot at a heavy rate on average when he's on the ice.

Russell's Corsi split last year? 44.8% Corsi, -4.75% below average from the rest of the team.

Last year, Russell's PDO was at 100.5, 47.7% GF, goalies had an average of .903% save percentage when Russell was on ice.
This year, Russell's PDO is at 104.1, 62.5% GF, goalies have an average of .949% save percentage when Russell is on the ice.

to me, Russell's high 5v5 possession #s represent his willingness to skate the puck up the ice (a trait lacking in Oilers D for some time), but also his willingness to hang out behind the net and wait for things to develop up-ice.

Re: Russell,
David Staples' feud with the Twitterstats crowd is annoying as hell. Staples can't seem to understand corsi, as he constantly argues it's "tracked wrong," when the whole point of it is that it's an objective stat. meanwhile the statisticians can't seem to see past their numbers and their acknowledgements about the limitations of their stats appear to be only lip service.

 

[phaedrusDH]

Ive actually really budged on Russell and I enjoy him on this team. But I see where his poor analytics come from. He gives up the blueline almost every rush to avoid getting burnt wide. This prevents those type of breakaways/rushes but then allows the forward time and space to make a play or get a shot off from further. More shots against him are allowed by he makes up for it by blocking a ton of them. He allows a lot of low and mid danger shots but prevents high danger ones on the rush

good summary. he also has a short reach and doesn't poke-check a whole lot, which leads to more shots of the rush.

 

[Aceboogie]

to me, Russell's high 5v5 possession #s represent his willingness to skate the puck up the ice (a trait lacking in Oilers D for some time), but also his willingness to hang out behind the net and wait for things to develop up-ice.

Re: Russell,
David Staples' feud with the Twitterstats crowd is annoying as hell. Staples can't seem to understand corsi, as he constantly argues it's "tracked wrong," when the whole point of it is that it's an objective stat. meanwhile the statisticians can't seem to see past their numbers and their acknowledgements about the limitations of their stats appear to be only lip service.

Agree fully. Oilers twitter is terrible. Both sides are out of line. When they discuss their interest seperatly they are reasonable and good to read. But when one side is trying to prove the otherside wrong its painful to read. Analytics community cant give up that sometimes corsi is wrong on a player, or more research is needed. MSM cant give up that analytics is valuable and growing. That relationship started negative from get go and never improved

 

[McDNicks17]

what is the best site for these stats? I recall behind the net back in the day was really good.

Hockey Analysis is probably the easiest to use.

Natural Stat Trick is similar, but also tracks scoring chances and high danger scoring chances. NST also tracks game by game numbers, so you can see who someone played against and their numbers against them.

 

[Aceboogie]

good summary. he also has a short reach and doesn't poke-check a whole lot, which leads to more shots of the rush.

Yup. And doesnt try to poke check that much as it leaves him exposed. He really just plays inside a small box but does it well, knows his limitations.

Youd prefer to have a D that prevents zone entries from the start but unfortunately not every D is Lidstrom. So players have to make up for it in other ways to compensate. Russell went from small offensive D in junior to a 2 way/defensive guy in the NHL by learning to compensate for his short comings (not a great 1 on 1 defender) and blocking shots. Hes played several years in NHL as a result

 

[phaedrusDH]

Agree fully. Oilers twitter is terrible. Both sides are out of line. When they discuss their interest seperatly they are reasonable and good to read. But when one side is trying to prove the otherside wrong its painful to read. Analytics community cant give up that sometimes corsi is wrong on a player, or more research is needed. MSM cant give up that analytics is valuable and growing. That relationship started negative from get go and never improved

i don't know what the consensus is on him, but i think Jonathan Willis may walk that line the best as far as Oilers MSM

 

[phaedrusDH]

Yup. And doesnt try to poke check that much as it leaves him exposed. He really just plays inside a small box but does it well, knows his limitations.

Youd prefer to have a D that prevents zone entries from the start but unfortunately not every D is Lidstrom. So players have to make up for it in other ways to compensate. Russell went from small offensive D in junior to a 2 way/defensive guy in the NHL by learning to compensate for his short comings (not a great 1 on 1 defender) and blocking shots. Hes played several years in NHL as a result

i think the discussion on Russells highlights one of the biggest deficiencies of advanced stats, which is objectively quantifying strong defensive attributes. corsi, zone exits, etc., all provide some measure, but are not nearly as indicative as offensive stats.

maybe i'm biased because i play defense, but it is IMO, a much more cerebral position that makes it much harder to track.

 

[Aceboogie]

i don't know what the consensus is on him, but i think Jonathan Willis may walk that line the best as far as Oilers MSM

I fully agree. He has his foot in both, uses both strengths and recognizes both weaknesses. I enjoy reading him

I enjoy Staples insight as well and I regard him well. But man can be obtuse sometimes. Same goes for the bloggers.

For bloggers its like frig I know Marincin isnt Chara stop trying to say he is because hes got good possession on the bottom pairing. Obviously Babcock thinks otherwise. Marincin/Russell are the hills they are choosing to die on.

MSM also have their guys they dont budge an inch on