[Game Takes] Flames 3 Jets 2

[Bend it like Bourgeois]

Originally Posted by Bingo

The only way they aren't factual is if they are recorded in error ... a shot outside the plate is considered in, the person missed a pass so it's classified wrong.

Otherwise it's pretty much the same as the guy counting traffic at an intersection. He probably misses the odd car, once and a while he hits the button twice by mistake but over the course of an hour the data is what it is ... largely factual.

However what do they mean is very much up for debate.

Ehh. Respectfully disagree.

Someone drew a box. Do the lines really define a black and white high danger chance or not? Even if they do generally it would be all over the map when applied to a given game or a given player.

Then someone else decides if a shot was from that box and meets other subjective criteria. In a fast moving game from a distance and sketchy angles. To get offside calls mostly right a trained linesman goes to a brightly painted straight line across the ice and watches intently. That’s not what we have. We have loosely defined criteria loosely judged.

Counting cars is pretty clear. Is it a car. Did it go past me.
Fancy stats are more like counting smart people at a food court. You can probably reinforce or challenge your assumptions and that might be interesting. I wouldn’t rely on it for facts though.

Edit: I should clarify I am generalizing about fancy stats but would agree there are some more accurate than others. Counting shots is probably pretty accurate as a number (if often poorly used). Counting shots plus some other criteria gets into pretty muddy territory. it’s downhill from there.

[Red]

Originally Posted by GioforPM

Nope. This was the toe save by Markstrom (right before their actual goal).

But again, these are individual circumstances whereas I think this stat is meant to be digested over a large portion of the season.

ETA: Even the most basic of stats, like assists, vary in individual plays as far as quality goes.

For sure. Looking at a statsheet from a game and seeing player A had a G+A may look like said player had a good game. But he may have been just lucky to capitalize on a couple plays while being terrible the other 18 shifts.

But for team stats, these variables can paint a very different picture than the eye test and often the end results prove. I feel like these stats remove the single most important factor, skill of individuals.
Facing a Laine or Ovechkin one timers, even from outside the home plate has to be seen as more dangerous than a weak backhand by Nordstrom or Bennett from up close.
At the same time, a Bennett weak backhand while the goalie is swimming may be more dangerous than a Kucherov snipe while the goalie is set to face it.

Just so many factors.

[Enoch Root]

Originally Posted by GioforPM

All stats can be misleading taken by themselves and this is moreso if you take a small sample size like a game (eg Markstrom's best save was not a high danger chance - it was his toe save on a shot from near the dot where he had to move all the way across).

But over an extended time, a disparity in high danger chances tells you something.

But no one uses them like that. They are almost invariably used as a way to illustrate how a single game went. Like last night.

My complaint has always been with fans' conclusions, not with the stats. They are what they are, a (very small) set of simple, counting data.

[Bingo]

Originally Posted by Red

Would a Kucherov, Laine or Ovechkin one timer from the top of the circle be a HD chance? Since it's outside the home plate I guess it would not.

Now let's say a guy like Nordstrom fights checks to get to the net and gets a weak backhand off in to the goalie chest while trying not to fall. That is a HD chance, correct?

How about a pinching D scoring from top of the circle on what is an almost empty net because the goalie is out of position? Think of a sneaky backhand pass from behind the net to a pinching D while the goalie is anticipating a wraparound.

Nordstrom's isn't a high danger chances unless he picked up a rebound, took a pass, or tipped it.

Both cases 1 and 3 are outside the plate, so wouldn't be a scoring chance or a high danger chance.

But really it has to be that way. I don't want people "judging" chances as that throws a pile of subjectivity into it.

[Bingo]

Originally Posted by old-fart

So, another question due to my lack of knowledge...

What about a 2 on 1 (for example) that results in an excellent pass across and a shot that is either blocked or goes wide? Is that NOT a high danger attempt or scoring chance because it didn't result in a shot on goal?

In the game last night in the second period (I think) Backlund made a hard quick pass to EatBread who tried to pull it across the front of the net and went backhand high... but missed that top corner by inches. He had the goalie beat (I think) but missed that top corner. High danger chance or even scoring chance based on the statistical criteria, or no? What about shots that hit the post?

Apologies for my lack of understanding on the methodology....

That would be a high danger chance because he took the pass just out of the crease and got a shot off.

Doesn't matter if the shot is blocked, wide, hits a post or saved it's a high danger shot attempt.

[Bingo]

Originally Posted by Bend it like Bourgeois

Ehh. Respectfully disagree.

Someone drew a box. Do the lines really define a black and white high danger chance or not? Even if they do generally it would be all over the map when applied to a given game or a given player.

Then someone else decides if a shot was from that box and meets other subjective criteria. In a fast moving game from a distance and sketchy angles. To get offside calls mostly right a trained linesman goes to a brightly painted straight line across the ice and watches intently. That’s not what we have. We have loosely defined criteria loosely judged.

Counting cars is pretty clear. Is it a car. Did it go past me.
Fancy stats are more like counting smart people at a food court. You can probably reinforce or challenge your assumptions and that might be interesting. I wouldn’t rely on it for facts though.

Edit: I should clarify I am generalizing about fancy stats but would agree there are some more accurate than others. Counting shots is probably pretty accurate as a number (if often poorly used). Counting shots plus some other criteria gets into pretty muddy territory. it’s downhill from there.

Agree with that.

But the categorization as it's built today is a simple counting exercise without opinion or subjectivity.

That isn't to say the definition is correct or optimal though.

[old-fart]

Originally Posted by Textcritic

I believe the model factors in all shot attempts. So, in the instance above this indeed would count as a HDSC.

Ah.... "shot attempts"... not just shots. Makes much more sense in my head now.

But on my two on one example then... if the one forward attempted the pass to the other forward that was blocked and/or bobbled into the corner with no attempted shot, it doesn't count as a HDSC as no shot was attempted?

(last question, I swear)

[Bingo]

Originally Posted by old-fart

Ah.... "shot attempts"... not just shots. Makes much more sense in my head now.

But on my two on one example then... if the one forward attempted the pass to the other forward that was blocked and/or bobbled into the corner with no attempted shot, it doesn't count as a HDSC as no shot was attempted?

(last question, I swear)

Exactly.

Have to "attempt" a shot.

[Red]

Originally Posted by Bingo

Nordstrom's isn't a high danger chances unless he picked up a rebound, took a pass, or tipped it.

Both cases 1 and 3 are outside the plate, so wouldn't be a scoring chance or a high danger chance.

But really it has to be that way. I don't want people "judging" chances as that throws a pile of subjectivity into it.

Ya, until there is a more accurate way of bridging the gap of eye test and probability to meeting a random formula the stats guys and deniers will continue to disagree.

Meeting a formula for a HD chance seems more important than watching the actual play because we need to remove potential bias. And I get why.
But, and this is a big but. We often hear fan excuses like this; we lost but were the better team because we had more shots on net. We all know that this is not always true, but again, the stats can support the "better team" argument.

Couple of unskilled players passing the puck in "an area of choice" resulting in a weak shot that had little chance (ran out of room, out of balance, backhand while falling etc) of beating the goalie should not equal a HD chance. Make that same play with players that are square to the play and have time to take a good shot and it's a different story. But the formula is the same.

While a shot from outside by a guy that scores from there several times a year and has done it all career should be a HD change. It must, really. I mean coaches and players work a lot on nullifying that very play by certain players.

Would be nice to take a game and break it down play by play with both, stats and non stats people. Re-play a sequence and discuss them before marking them as HDs or not.
I think the stats and non stats people would agree a lot. And I think they would not align closely with the advanced stat websites.

[Bingo]

^I think at some point shots will be deemed dangerous based on a sensor suggesting the goalie had to move.

And that will be a big leap in my mind.

If you're shooting from a bad angle ... goalie doesn't move at all. But even a shot from the middle can sometimes not result in the goalie having to move much.

[Calgary4LIfe]

Originally Posted by Bingo

^I think at some point shots will be deemed dangerous based on a sensor suggesting the goalie had to move.

And that will be a big leap in my mind.

If you're shooting from a bad angle ... goalie doesn't move at all. But even a shot from the middle can sometimes not result in the goalie having to move much.

I think in the next couple of years, stats are going to take a massive leap forward. Even something as simple as 'possession' has to be indirectly measured, thereby adding a degree of inaccuracy.


I actually look forward to a more reliable and accurate set of statistics, some of which I am going to guess most people here probably haven't even thought of, but that we will all come to use regularly.



This is not to say that I don't enjoy what is available now (and I appreciate you always including them in your Game Takes). I also appreciate you breaking down some of those definitions (and patiently doing so).

[Enoch Root]

^ seconded

[SnipeShow]

When Tanev shoots from our blue line is it considered a high danger shot against Winnipeg?

Sent from my GM1917 using Tapatalk

[FireGilbert]

The rebounds being a high danger chance are interesting. If a goalie with poor rebound control turns a low danger shot into a high danger rebound shouldn’t that be similar to just letting in the low danger shot?

[Bingo]

Originally Posted by FireGilbert

The rebounds being a high danger chance are interesting. If a goalie with poor rebound control turns a low danger shot into a high danger rebound shouldn’t that be similar to just letting in the low danger shot?

I wouldn't say so.

Because his mistake creates a high danger chance against ... rebound in the home plate area.

The designation makes sense when you think about it. Goalies have to adjust to a tip, a rebound, and a pass that makes them move from a post. All more dangerous than a guy just shooting from the middle of the ice with no deception at all.

Someone said before they should make an exception for shots out side the plate that come from circumstances too though ... and I see that. But that goes back to the objective measure issue.

An example ... guy passes post to post for what looks like a tap in, but technically he's beyond the post and therefore outside of the plate.

[Hot_Flatus]

Originally Posted by Textcritic

Sure, but it's not like the Flames have not already dominated the Canucks twice already this season, and pounded them into the ground. In fact, I would say that Game #2 v. Vancouver on 16 January was easily the Flames most complete game of year before last night.

The calling card of this team is traditionally 2 steps forward, 1 step back. The more realistic winning this series in convincing fashion against the Nucks is made to seem, the more likely it seems to me that this team will revert to it's grassroots and play some awful hockey and make this a struggle.

I truly hope this is not the case, but until they prove me wrong, I'm no longer giving them the benefit of the doubt. Smoke them 3-1 or 4-0 with similar outings to the last game and I'm onboard.

[united]

The unfortunate thing about high danger chances is that generally while "high danger" sounds good, and intuitively the concept makes sense, it under performs almost all other metrics as it relates to correlation to future goal differential and repeatability.

I've looked at Natural Stattrick's high danger metric, which DOES NOT use passing data, and in every single season it lagged every other "advanced" statistic fairly significantly.

Early in the season, high danger chances is significantly less predictive and significantly less repeatable than both corsi and expected goals. The gap continues toward the 50th game of the season, at which point the spread between the three slowly shrinks until the end of the season. Of course, historically around game 60 actual goal differential first equals then surpasses corsi and expected goals (and obviously high danger chances).

So in the end you have a metric - high danger chances - that is significantly less meaningful than corsi and expected goals from the start of the season to about two-thirds into the season, at which point it is less meaningful than actual goal differential. In other words, at no point in a season does high danger chances outperform other metrics so I'm not sure why it's referenced so commonly in its current form.

Of course a big part of that is the sample size high danger chance needs to stabilize which is the largest problem I think: people who don't understand the importance and impact of sample size, and that it isn't equal across all metrics, shotgun the metrics out, for instance quoting a player's or team's corsi, expected goals, and high danger chances across the last 10 games. A 10-game sample of corsi has a much different meaning than a 10-game sample of high danger chances and should be treated accordingly, but it isn't unfortunately. It isn't done intentionally just out of innocent ignorance.

Now, all the above was using a high danger metric without passing data. Bingo has mentioned above that his model does use passing data which would undoubtedly make it significantly more powerful than anything on Natural Stattrick's site and likely up there with private data.

[Jay Random]

Originally Posted by united

It isn't done intentionally just out of innocent ignorance.

Oh, the ignorance isn't entirely innocent. This kind of thing will go on as long as the media recruit most of their staff from the segment of the population that flunked high-school math.

I used to know a senior editor for a major newspaper, which I shall not name to protect the flagrantly guilty. This editor divided the world into ‘word people’ and ‘number people’, with no overlap. If you could handle numbers at all, that disqualified you from being a ‘word person’. And ‘word people’ were categorically superior and only they, I inferred, were fit to be journalists.

It seems bizarre that people could actually be proud of being innumerate, but there you are.

[DeluxeMoustache]

Originally Posted by Bingo

^I think at some point shots will be deemed dangerous based on a sensor suggesting the goalie had to move.

And that will be a big leap in my mind.

If you're shooting from a bad angle ... goalie doesn't move at all. But even a shot from the middle can sometimes not result in the goalie having to move much.

Totally agree. Huge step forward.

So much of how likely a shot is to go in depends on where it is shot towards more so than where it is shot from, and of course what the goalie has to do to save it