Zone breakouts numbers

[Number 57]

Hi, reading up on analytics, I thought for sure I would see zone breakouts numbers, but I can't seem to find anything on the subject. When analyzing stats for defensemen, the good old stats did not provide much aside from +/-, which is basically useless. Blocked shots and hits are fine and all, but don't tell much of the story.

I feel like the big stat that's missing is breakouts. When eveluating defensemen, what they do once they get the puck in the defensive zone is a big deal. Can he make a clean breakout pass? Can he skate the puck out of the zone? Is he committing turnovers? Isn't he shooting that thing off the glass every single time?

I was thinking of dividing the stats into street lights i.e. Red, Yellow, Green.

Red = a breakout attempt that leads to a turnover (giving up the puck inside the zone after gaining possession)

Yellow = zone cleared but breakout aborted (defenseman sent the puck out but his teammates couldnt do anything with it, other team regains possession outside of the zone) include icings here.

Green - successful breakouts (the defenseman made a nice pass to a forward that started a breakout OR the defenseman skated the puck out of his zone himself)

The stats would read something like this for a single game:

BREAKOUT ATTEMPT STATS -- RED -- YELLOW -- GREEN
Shayne Gostisbehere -- 3 -- 2 -- 4
etc

Now it may be complicated to actually track this, but it isnt rocket science either. If the NHL can have someone sitting out there counting every single shots taken and by whom, I'm sure they can have someone who judges every single breakout attempt as well.

It seems to me like how a dman breaks out of his zone is one of the most important thing in hockey, so it seems weird that there are no stats for this. While Corsi, etc is great and all, it's not the best way to evaluate a blueliner.

Click to read more...

 

[Michael Farkas]

In very short, that requires video, not already-counted stats like shots, AND an understanding of the game and multiple people per game to track.

That's quite a chore.

It is valuable data, and as such, it is proprietary and not public. Data that's no longer overly useful is made public (shots (player and team), plus/minus, efficiency/save pct. over the years have set this precedent to a degree)...the good stuff is kept close to the vest.

 

[morehockeystats]

In very short, that requires video, not already-counted stats like shots, AND an understanding of the game and multiple people per game to track.

That's quite a chore.

It is valuable data, and as such, it is proprietary and not public. Data that's no longer overly useful is made public (shots (player and team), plus/minus, efficiency/save pct. over the years have set this precedent to a degree)...the good stuff is kept close to the vest.

Not only it's a chore, but it involves a human evaluation and a possibility of a bias.

 

[Michael Farkas]

The only way to get meaningful data is to have "bias". People think bias is a negative term, but it can be very much positive (or, of course, negative if you used in such a way).

Bias prevents Maxime Boisclair from going in the 1st round of a draft, for instance.

The current stats that you guys play with and hold on a pedestal (shot-counting stuff, save pct., etc.) are also biased in their own way. They are a quantity stat, so the hope is that they lessen bias, but they're still very much biased against certain factions of players. Save pct is biased against goaltenders in certain systems.

It's just a matter of what bias the community feels is permissible and what bias they intend to introduce and, further, stand behind. But let's not pretend that counting shots is unbiased, certainly...

 

[Canadiens1958]

Bias

Not only it's a chore, but it involves a human evaluation and a possibility of a bias.

Bias is a necessity for certain stats. Zone breakout numbers being an example. To do the numbers properly you have to know the team game plan, objectives depending on match-ups, game circumstances and other internal data that will never be released publically.

 

[morehockeystats]

The only way to get meaningful data is to have "bias". People think bias is a negative term, but it can be very much positive (or, of course, negative if you used in such a way).

Bias prevents Maxime Boisclair from going in the 1st round of a draft, for instance.

The current stats that you guys play with and hold on a pedestal (shot-counting stuff, save pct., etc.) are also biased in their own way. They are a quantity stat, so the hope is that they lessen bias, but they're still very much biased against certain factions of players. Save pct is biased against goaltenders in certain systems.

It's just a matter of what bias the community feels is permissible and what bias they intend to introduce and, further, stand behind. But let's not pretend that counting shots is unbiased, certainly...

That's not the bias I an talking about.
I am talking about the mere bias in the definition of terms, what a "clean breakout pass" is, for example.

There is a similar bias, for example, of hit counting, depending on the perception of a hit by the scorekeeper.

The bias in shot counting and in SVP counting is smaller. Much smaller.

 

[Canadiens1958]

Bias Part 2

That's not the bias I an talking about.
I am talking about the mere bias in the definition of terms, what a "clean breakout pass" is, for example.

There is a similar bias, for example, of hit counting, depending on the perception of a hit by the scorekeeper.

The bias in shot counting and in SVP counting is smaller. Much smaller.

You are the only one in this thread or anywhere else raising the issue of a "clean breakout pass". Not even on or close to the table.

Just as no one differentiates between a clean goal and a dirty goal at any level of data gathering be it for scoring, SV%, winning, etc.

 

[Doctor No]

You are the only one in this thread or anywhere else raising the issue of a "clean breakout pass". Not even on or close to the table.

I'm happy to jump in.

Are you claiming that the term "clean breakout pass" has no degree of subjectivity to it, and that if you show 100 hockey fans a video of 100 breakout passes, each of the fans will correctly identify the "clean" ones from the bunch?

There's a degree of subjectivity on what constitutes a "shot on goal", on what constitutes a "hit", and what constitutes a "takeaway", so why wouldn't there be subjectivity with the term "clean breakout pass"?

 

[Number 57]

I'm happy to jump in.

Are you claiming that the term "clean breakout pass" has no degree of subjectivity to it, and that if you show 100 hockey fans a video of 100 breakout passes, each of the fans will correctly identify the "clean" ones from the bunch?

There's a degree of subjectivity on what constitutes a "shot on goal", on what constitutes a "hit", and what constitutes a "takeaway", so why wouldn't there be subjectivity with the term "clean breakout pass"?

Well, in my proposal at least, it would be pretty straight-forward.

Green - successful breakout (retained possession)

Yellow - puck goes out but team doesnt retain possession

Red - turnover inside the defensive zone

 

[Canadiens1958]

No

I'm happy to jump in.

Are you claiming that the term "clean breakout pass" has no degree of subjectivity to it, and that if you show 100 hockey fans a video of 100 breakout passes, each of the fans will correctly identify the "clean" ones from the bunch?

There's a degree of subjectivity on what constitutes a "shot on goal", on what constitutes a "hit", and what constitutes a "takeaway", so why wouldn't there be subjectivity with the term "clean breakout pass"?

Just saying that it does not matter. A goal is a goal whether clean or not. Likewise for a break-out pass.

 

[morehockeystats]

Well, in my proposal at least, it would be pretty straight-forward.

Green - successful breakout (retained possession)

Yellow - puck goes out but team doesnt retain possession

Red - turnover inside the defensive zone

retained possession because the opponent botched the intercept?

It's still liquid. Yes, can be quantified. Involves a lot of work and still doesn't preserve the definition bias.

 

[morehockeystats]

Just saying that it does not matter. A goal is a goal whether clean or not. Likewise for a break-out pass.

It's often necessary to conduct a lengthy review involving a video team in Toronto to verify whether a goal is a goal, and even then sometimes the agreement is not total.

 

[Doctor No]

Well, in my proposal at least, it would be pretty straight-forward.

Green - successful breakout (retained possession)

Yellow - puck goes out but team doesnt retain possession

Red - turnover inside the defensive zone

Things are oftentimes straightforward on paper.

But hockey games aren't played on paper - they're played inside television sets.

A "shot on goal" is a pretty straightforward definition (on paper).

 

[allan5oh]

breakout numbers are tracked and definitely talked about, but some people think successful breakouts per 60 is a good number to go by when it's not.

 

[Ola]

I am not sure exactly how telling it would be, but it would be interesting to look at PIMs by team when a specific defender is on the ice. Possibly made as a factor showing the impact in relation to teammates avg when the defender. Possonly adjusted for score.

Alt 1
Nik Lidstrom (ie a player that can get the puck up ice) -- Penalty by team when player is on the ice/60 = 2.33
Dan Girardi (ie a player that screws up a lot, puts his team in tough situations) -- Penalty by team when player is on the ice/60 = 12.76

Alt2
You just run the above data for all players, and then gather the avg impact on the these numbers when they are on the ice with the defender in Question.

Is this available anywhere? Is raw data available anywhere to create it?

 

[Number 57]

breakout numbers are tracked and definitely talked about, but some people think successful breakouts per 60 is a good number to go by when it's not.

Link?