The Avs Advanced Stats Thread [Now with a Fresh, "New Title" look!]

[henchman21]

So throughout the year a few different stats are going to be tracked, and at the end of the year a simple regression analysis will be done to see if there is anything to these stats. I will update this post as the year goes on to attempt to keep every thing organized.

If you have an idea for a stat, or want to help gather the data please post up. I might add a signup list if we get enough.

Possible stats being tracked:

Puck Possession Time by Zone: Exactly what it says. This will be timed from the time puck possession is clearly won, to the time it is clearly lost (IE board battles can be included in the time). I want to track both teams as there will be time where possession is given up and it hasn't been recovered yet.

Controlled Zone Entries: Simply when a player enters the offensive zone with puck control. At the start it probably won't be broke out by player, but may evolve.

Shots broke out by area (basic shot chart): Tracking the origin of shots (tips don't count) into offensive zone areas... probably something similar to this:

Shot attempts broke out by area (shot attempt chart): Same as above, but the shot doesn't have to be on goal.

Scoring chances: Difficult metric to quantify, but would be very useful info if a method can be found. Stats compiled by the different networks and arena to arena are subjective and tend to be biased.

I will also be putting the Corsi and Fenwick in here as well as a comparison, but not actually tracking those as the data is out there already.

If you have an idea for a stat, post it up. If you want to help track, post up what you can do. Even if it is only a few games this season, it will be helpful.

Cumulative ES scoring chances (Avs-Opp): 27-42

Click to read more...

 

[cgf]

I'd be curious to see both puck possession time by zone and just a simple time spent in each zone total cause I think that using those two and corsi combined would give you a good idea of how the team plays relatively to the average bear.

And since %of shots blocked is becoming a thing, I'd be very curious to see a percentage shots where the goalie screened be tracked. See the difference between the teams throwing everybody between the goalie and the puck, and the teams who ask their goalie to make a lot of easy and clean saves. I suspect some goalies will do better in one scenario than the other, but also wouldn't be surprised to find goalies who see a lot more of the shots they face have much higher save percentages.

 

[Gigantor The Goalie]

I have a different shot chart for something I'm currently fleshing out regarding goalies. I just need the time that school isn't giving me to work on it right now.

 

[henchman21]

I'd be curious to see both puck possession time by zone and just a simple time spent in each zone total cause I think that using those two and corsi combined would give you a good idea of how the team plays relatively to the average bear.

And since %of shots blocked is becoming a thing, I'd be very curious to see a percentage shots where the goalie screened be tracked. See the difference between the teams throwing everybody between the goalie and the puck, and the teams who ask their goalie to make a lot of easy and clean saves. I suspect some goalies will do better in one scenario than the other, but also wouldn't be surprised to find goalies who see a lot more of the shots they face have much higher save percentages.

That is kind of my thought with this. I'm not totally sold that corsi is the best indicator for actual possession.

% of shots where goalie is screened is probably a difficult one to track on the broadcast. Too many weird angles and cuts between them. It would be interesting to see though.

I have a different shot chart for something I'm currently fleshing out regarding goalies. I just need the time that school isn't giving me to work on it right now.

School is for foolz

 

[cgf]

For me it's not even about corsi's issues, it's that most advanced stats people seem to be making the assumption that all teams have the same strategies and goals and so play the same. What I'd like to see is a combination of numbers that can illustrate some of the tactical differences in approach for the numbers guys who don't watch enough of the team they're analyzing to properly understand the context of their numbers.

My thinking being that if you compared a team's corsi, to their zone time to their possession in the offensive zone time, you'd start to get a numerical representation of which teams are setting up and looking for a great shot, and which teams are just dumping it on net/deep every chance they get.

 

[Lonewolfe2015]

I really want to track average shot/goal distance closely. Might be a good one to add.

 

[henchman21]

For me it's not even about corsi's issues, it's that most advanced stats people seem to be making the assumption that all teams have the same strategies and goals and so play the same. What I'd like to see is a combination of numbers that can illustrate some of the tactical differences in approach for the numbers guys who don't watch enough of the team they're analyzing to properly understand the context of their numbers.

My thinking being that if you compared a team's corsi, to their zone time to their possession in the offensive zone time, you'd start to get a numerical representation of which teams are setting up and looking for a great shot, and which teams are just dumping it on net/deep every chance they get.

I agree. I've stated in a couple threads that more 'creative' teams might hold the puck more and still shoot less (also probably have a higher shooting percentage)... and other teams may just throw everything at the net, but from low percentage areas and may not possess the puck as long.

I really want to track average shot/goal distance closely. Might be a good one to add.

That one shouldn't be difficult if a good shot chart can be found or created.

 

[cgf]

Soccer has some similar problems. Possession time has become the big stat like corsi/fenwick, but there is no break down of which part of the pitch most of this possession occurs in nor have people come up with a good solution for teams who control the movement of the ball without firm possession like Dortmund who use their forecheck, speed and aggressive leading passes to push the ball into dangerous areas via a series of a puck battles basically. So their possession stats are often misleadingly low despite them controlling where the ball is going.

Though in soccer visual representations are used to augment that missing context. You'll find heat maps that track individual players positioning during the match that show where players are operating and you'll find passing maps that show every pass a player has made, but neither really can show the team wide data that can not be acquired in any way other than watching the match.

 

[Lonewolfe2015]

That one shouldn't be difficult if a good shot chart can be found or created.

You can actually track the shots against distance via the real time stats sheet at the end of each game.

 

[henchman21]

You can actually track the shots against distance via the real time stats sheet at the end of each game.

The question there is are they reliable? I honestly don't know.

I know the NHL shot charts are extremely unreliable.

 

[Lonewolfe2015]

The question there is are they reliable? I honestly don't know.

I know the NHL shot charts are extremely unreliable.

Sounds like some investigation is due then.

 

[Gigantor The Goalie]

I use http://www.sportingcharts.com/nhl/ for my shot chart information. What I'm focused on is the difficulty of shots that goalies face and how they control the rebound or if the defense has to clear it away.

 

[henchman21]

I have seen that site before, but have never looked into where they got their data... anybody know?

 

[Freudian]

I have seen that site before, but have never looked into where they got their data... anybody know?

It's the NHL recorded data. I can't imagine anyone else tracking it.

 

[Pokecheque]

Cool. I'll be frequenting this thread aplenty.

 

[cgf]

Now the really important question...can we get this thread stickied? :-D

 

[Freudian]

Here are three interesting shot charts for Avs players. Barrie, Holden and O'Reilly were 'accused' of having career high shooting percentages last season.

O'Reilly switched from center to wing. If you compare 11/12 with 13/14 (both full seasons) you can see not only has shot location changed for him but also shot type. He shoots more wrist shots and shoots closer to the net. It's not that strange his shooting percentage has gone up and his goal production on the PP has gone up.

I compared Barrie to Yandle, who is also a puck rushing defenseman but also is a very successful PP quarterback. Also we see a completely different pattern. Barrie doesn't shoot a lot from the point but gravitates towards the slot where he uses wrist shots. He shoots from much closer to goal than Yandle does.

We all know how Holden scored last year and it's no surprise his shot chart looks like this:

Without proper context, claiming that someone's shooting percentage is too high is a meaningless statement but I've seen this claim made by fans of other teams. It's possible these three guys will have their shooting percentage go down, but they also play in such a way that enables a high likelihood of scoring when they shoot. Much like Alex Tanguay.

 

[tigervixxxen]

I have seen those charts before, always worth a look. Great comparison with Yandle. That right there is how a team inflates their Corsi.

Shooting percentages can't be compared player to player. If someone is shooting higher than their career percentage then there might be some correction. In young guys it's hard to say what their career percentage will be set at. For a guy like O'Reilly who has changed his game a lot from when he came into the league I can see it's even tougher to pinpoint for him. I believe Stastny shot 16% last year and I'm confident he won't replicate that, for example. It's the same thing with Varly people arguing he performed much better than his career average, which is stupid to say because of his age and a lot of those years he was playing for awful teams.

 

[Gigantor The Goalie]

So I've been trying out some new Advanced Stats focused on the play of goalies and team defense. The real time stats I tried out yesterday to success I believe in evaluating Berra's play using the three indexes I've come up with. I just went through the first 11 games (October 2013) of last season using my Difficulty Index to show if the Avs actually kept shots out of the immediate danger zones and were able to get in close for their goals/shots.

Here's a chart created from the data:

Out of the 11 games, only in 3 games did the opposing team have the more dangerous shots then our Avs. In fact if you look at the Boston game you'll just how many low percentage shots they were shooting from most of the time. You can also see that in our only loss in October that the Wings were successful in pushing back our shooters resulting in the loss.

For those familiar with Corsi I'd like to see how we did relative to the Difficulty Index posted above.

The Difficulty index is the offensive/defensive zone split into different area's and labelled with numbers indicating how dangerous the area is for a shot. So the doorstep and in the crease is the most dangerous place and therefore is given a "1" label. Behind the goal line and beyond the blue line are given a "5" as they are the area's least likely to score a goal.

The math is simple from there. If a team takes 2 shots from each zone then you'd end up with:

(zonexshots)
5x2=10
4x2=8
3x2=6
2x2=4
1x2=2

You add the result of the zone x shots equation together then divide that by the number of shots against/for. So the Difficulty Index for this example would be 3. Which you want to keep the other team at because it means their taking more shots to the outside and in low scoring areas.

I have two additional statistics that are real time stats, Rebound Index and Shot Index. I'm not going to go more into it today though.

 

[henchman21]

That is interesting. I would like to see a breakout of the zones and how you came up with the danger. Great work!

 

[tigervixxxen]

I think I saw some of your tweets yesterday, where do you get the data on the goalies? Is it just your observation and where do you get the index numbers? Can you expand on what you saw from Berra?

 

[Pokecheque]

For me it's not even about corsi's issues, it's that most advanced stats people seem to be making the assumption that all teams have the same strategies and goals and so play the same. What I'd like to see is a combination of numbers that can illustrate some of the tactical differences in approach for the numbers guys who don't watch enough of the team they're analyzing to properly understand the context of their numbers.

My thinking being that if you compared a team's corsi, to their zone time to their possession in the offensive zone time, you'd start to get a numerical representation of which teams are setting up and looking for a great shot, and which teams are just dumping it on net/deep every chance they get.

I don't think that's true, even if you're just talking Corsi. Advanced statheads not only acknowledge there are different tactics/strategies, they can offer some great insight as to how each one is effective, or ineffective. Each one has its own strengths and weaknesses. For instance, fancy stats usually tell us that dump-and-chase is a bad idea, which it is...except the Kings have done it for years and are still successful, because they do a lot of other things to make up for it. Fancy stats folks don't usually like conservative trapping styles, but acknowledge that at least in terms of possession, the Devils employ a pretty decent version of it. Usually an aggressive forecheck and speed win the day as far as they're concerned. But possession is only part of the game, they didn't have enough talent to put the puck in the net.

The one universal thing advanced stats preach is that the game is usually won and lost in the neutral zone. Play aggressively in the neutral zone and you gain a huge advantage. That doesn't necessarily mean just sit back and trap though.

The Avs played their best game last season against the Blues in that Shattenkirk-got-cut-so-let's-all-lose-our-minds slugfest. The Avs not only employed an effective forecheck, they were relentless in the neutral zone, disrupting everything the Blues tried to put together. It's those times when the Avs practically backed into their own goalie and allowed the opposition to hit the zone with speed that they ended up giving up so many shots.

I don't care what players say about fancy stats, it does them no good. It's the coaches and management who should pay them some heed. I'm glad Roy at least pays attention to zone time, which is KINDA analytic in nature, but I can't stand this nonsense about "we play a style that gives up a lot of shots," as if analytics teaches a team to play conservatively. Hogwash. The team that wins races to the puck, wins battles for the puck, moves the puck up and out of the zone efficiently, employs effective forechecking, AND stays aggressive in all zones is usually rewarded.

No single stat will tell you the whole story. But everyone knows that.

 

[Gigantor The Goalie]

Here's the basic chart that I use. I screen shot the ice tracker from sportingcharts.com and import to photoshop where I match it up with the black lines I created. From there I just count the shots and through them into an Excel spreadsheet that helps me with the math and such.

Last game I pretty much had to use my better judgement as to where the shots were taken. The Rebound Index was just simply seeing where Berra sent the puck and assigning a number to it.

1= Berra got possession
2= Berra sent the puck to the corner
3= The rebound ended up right in front of him
4= The rebound ended up going to Berra's weak side
5= Goal

 

[Freudian]

It's not like Avs plan to allow a lot of shots. What they are saying is if the shots they allow are shots taken from the outside, they feel comfortable that the goalies have a good chance of saving the puck. That's not a strange idea. All 30 teams in the league are trying to protect the slot.

The main reason Roy wanted Avs to get bigger is to win more puck battles. Winning puck battles is a key component in puck possession.

I think the reason the analytics crowd are so provoked by Sakic is because he says he don't need analytics to figure out what Avs needed to fix because he could see it with his own eyes so he doesn't see a big need to use them.

Avs track zone time and scoring chances in real time during games. Roy often reference it during intermissions. Those are at least as useful for tracking possession as the shot attempt based stats are, so what's the problem?

 

[cgf]

I don't think that's true, even if you're just talking Corsi. Advanced statheads not only acknowledge there are different tactics/strategies, they can offer some great insight as to how each one is effective, or ineffective. Each one has its own strengths and weaknesses. For instance, fancy stats usually tell us that dump-and-chase is a bad idea, which it is...except the Kings have done it for years and are still successful, because they do a lot of other things to make up for it. Fancy stats folks don't usually like conservative trapping styles, but acknowledge that at least in terms of possession, the Devils employ a pretty decent version of it. Usually an aggressive forecheck and speed win the day as far as they're concerned. But possession is only part of the game, they didn't have enough talent to put the puck in the net.

The one universal thing advanced stats preach is that the game is usually won and lost in the neutral zone. Play aggressively in the neutral zone and you gain a huge advantage. That doesn't necessarily mean just sit back and trap though.

The Avs played their best game last season against the Blues in that Shattenkirk-got-cut-so-let's-all-lose-our-minds slugfest. The Avs not only employed an effective forecheck, they were relentless in the neutral zone, disrupting everything the Blues tried to put together. It's those times when the Avs practically backed into their own goalie and allowed the opposition to hit the zone with speed that they ended up giving up so many shots.

I don't care what players say about fancy stats, it does them no good. It's the coaches and management who should pay them some heed. I'm glad Roy at least pays attention to zone time, which is KINDA analytic in nature, but I can't stand this nonsense about "we play a style that gives up a lot of shots," as if analytics teaches a team to play conservatively. Hogwash. The team that wins races to the puck, wins battles for the puck, moves the puck up and out of the zone efficiently, employs effective forechecking, AND stays aggressive in all zones is usually rewarded.

No single stat will tell you the whole story. But everyone knows that.

Your last sentence is what I was getting at. That the over emphasis on shot based stats leave out a lot of that contextual information that I think matching them with zone times and possession times would put into the mix.

Sure they recognize different tactics exist, but the most popular stats don't do a very good job of telling us whether a team is playing a specific system and so getting those shot discrepancies, or just getting out played.

Like the kings, they'd probably do very well in zone time and shots, relatively worse on possession time, so we could see that they were still getting in through the neutral zone a lot and getting plenty of shots off of it, but weren't passing their way in while looking for the ideal shot. Or the Avs, I bet that you'd find we spent a lot of our o zone time with the puck, but I bet you'd also find that we had relatively less zone time because of how much we based our attack on the rush.

That's what I want to see from the popular stats, actually let me get some context for the shot based stats, that doesn't come exclusively from watching the games.