Kings are still a very good team and will win lots of games soon

[KingsCorona]

hey do you remember that team that won the cup in 2012? i remember them going through a strange mid season slump and having a real tough time scoring goals...their name is on the tip of my tongue but i cant quite remember it

I seem to remember that team going through a mid-season slump practically every single season regardless of who was on the team
It sure seems like it, anyway.

 

[Jason Lewis]

Just observation on my part. It took Williams a while too, granted I think he was "more hurt" than Gaborik.
Not worried.

Kings don't exactly have a super easy system to adapt to as well.

 

[Trolfoli]

Funny how you could make a 4 paragraph rant against advanced stats have 75% of the paragraphs be wrong....

Denying advanced stats is like arguing cigarettes don't cause cancer....

Regression to the mean only means over a large enough sample size you should end up at or close to the mean.

Sighting "regression to the Mean" as the reason for the better shooting percentage recently is Wrong. Past performance from a statistical stand point has no bearing on future performance. (in reality if you suck so bad you get traded/waved there can be a link between past and future performance.)

Much like the coin that was sighted in the article. Flip a coin you you get heads or tails. No matter what the out come is it has no bearing on future outcomes. The same is true with Shooting %.

Now when you flip a coin 100 time and get heads every time you are either lucky or there may be another issue with the coin you overlooked. When you increase the sample size to a 1,000 and keep getting heads you need to look at other reasons for this. Ask questions like is the coin heads on both sides, is one side of the coin heavier ... When the Kings had a terrible shooting % over half way through the season you need to ask why is there such a discrepancy between the league average and what the Kings are putting up? Questions that follow are Players/Coach/System.... It was not some statistical chance that the Kings had a horrible shooting %. They earned it.

The problem with numbers/advanced stats is that people use them for their own agendas. The Kings are just unlucky and it'll even out.... No there is nothing in probability/Stats that gives rise to that. More then likely there have been some changes that you may or may not have picked up on. The team is constantly trying to adjust and become better, or should be.

Laying it all on advanced stats saying it will even out is just a cop out for people that don't want or can't see the real problems.

But then here I am the fool.... some one says, "the regression to the mean is a beautiful thing." When it doesn't exist and shows the person has a limited understanding of statistics, and I'm trying to engage in a discussion when others don't speak the language. Please from a mathematical point of view tell me where I'm wrong?

Don't worry the Kings have stockpiled enough negative shooting energy in the begging part of the season to carry them deep into the off season. Put your faith in puck karma and don't ask questions about what the real problems are. Life is much easier that way.

 

[Raccoon Jesus]

Regression to the mean only means over a large enough sample size you should end up at or close to the mean.

Sighting "regression to the Mean" as the reason for the better shooting percentage recently is Wrong. Past performance from a statistical stand point has no bearing on future performance. (in reality if you suck so bad you get traded/waved there can be a link between past and future performance.)

Much like the coin that was sighted in the article. Flip a coin you you get heads or tails. No matter what the out come is it has no bearing on future outcomes. The same is true with Shooting %.

Now when you flip a coin 100 time and get heads every time you are either lucky or there may be another issue with the coin you overlooked. When you increase the sample size to a 1,000 and keep getting heads you need to look at other reasons for this. Ask questions like is the coin heads on both sides, is one side of the coin heavier ... When the Kings had a terrible shooting % over half way through the season you need to ask why is there such a discrepancy between the league average and what the Kings are putting up? Questions that follow are Players/Coach/System.... It was not some statistical chance that the Kings had a horrible shooting %. They earned it.

The problem with numbers/advanced stats is that people use them for their own agendas. The Kings are just unlucky and it'll even out.... No there is nothing in probability/Stats that gives rise to that. More then likely there have been some changes that you may or may not have picked up on. The team is constantly trying to adjust and become better, or should be.

Laying it all on advanced stats saying it will even out is just a cop out for people that don't want or can't see the real problems.

But then here I am the fool.... some one says, "the regression to the mean is a beautiful thing." When it doesn't exist and shows the person has a limited understanding of statistics, and I'm trying to engage in a discussion when others don't speak the language. Please from a mathematical point of view tell me where I'm wrong?

Don't worry the Kings have stockpiled enough negative shooting energy in the begging part of the season to carry them deep into the off season. Put your faith in puck karma and don't ask questions about what the real problems are. Life is much easier that way.

You're not wrong, but your'e not right, either. You complaining about people putting all their eggs in one basket (advanced stats) yet you're putting all your eggs in exactly the other. Obviously reality lies somewhere in the middle, there's no need to be an absolute extremest railing against it. When the Kings were shooting 4% there were certainly other systematic factors at play, but **** luck was one of those factors. When I was talking about regression to the mean, I was referring to the luck coming around as well as the play getting back on track. It's been proven time and time again over a full-season sample size that most teams will have certain statistics even out. So we may not be talking mathematically/statistically enough for you, but the colloquial "regression to the mean" here simply means move back towards league averages for things like shooting %, PDO, and the like. Drop the condescension.

 

[lexlavender]

Regression to the mean only means over a large enough sample size you should end up at or close to the mean.

Sighting "regression to the Mean" as the reason for the better shooting percentage recently is Wrong. Past performance from a statistical stand point has no bearing on future performance. (in reality if you suck so bad you get traded/waved there can be a link between past and future performance.)

Much like the coin that was sighted in the article. Flip a coin you you get heads or tails. No matter what the out come is it has no bearing on future outcomes. The same is true with Shooting %.

Now when you flip a coin 100 time and get heads every time you are either lucky or there may be another issue with the coin you overlooked. When you increase the sample size to a 1,000 and keep getting heads you need to look at other reasons for this. Ask questions like is the coin heads on both sides, is one side of the coin heavier ... When the Kings had a terrible shooting % over half way through the season you need to ask why is there such a discrepancy between the league average and what the Kings are putting up? Questions that follow are Players/Coach/System.... It was not some statistical chance that the Kings had a horrible shooting %. They earned it.

The problem with numbers/advanced stats is that people use them for their own agendas. The Kings are just unlucky and it'll even out.... No there is nothing in probability/Stats that gives rise to that. More then likely there have been some changes that you may or may not have picked up on. The team is constantly trying to adjust and become better, or should be.

Laying it all on advanced stats saying it will even out is just a cop out for people that don't want or can't see the real problems.

But then here I am the fool.... some one says, "the regression to the mean is a beautiful thing." When it doesn't exist and shows the person has a limited understanding of statistics, and I'm trying to engage in a discussion when others don't speak the language. Please from a mathematical point of view tell me where I'm wrong?

Don't worry the Kings have stockpiled enough negative shooting energy in the begging part of the season to carry them deep into the off season. Put your faith in puck karma and don't ask questions about what the real problems are. Life is much easier that way.

I think Doty explained some of this as well, but your arguement hinges on the fact that you say there is no real "mean" to regress to. However, PDO regression exists, and there is years upon years of data from all teams that PDO, over a full season, usually averages out to one. We are regressing to that mean, our insanely low shooting percentage is climbing to match our normal save percentage values. We are also regressing to the "mean" set by our Corsi/Fenwick/possession values. While there is no set mean here (Its not like a certain possession percentage guarantees you a certain amount of goals) we are regressing in a sense to the fact that higher amounts of shots will lead to a higher amount of goals. And while you are correct that technically past performance does not effect future performance, continued performance at a certain expectation should yield a certain amount of success. You may flip a perfect coin 100 times and get all heads, but if you keep flipping after that, eventually, when we extend out to an infinite number of flips, the outcomes WILL be 50/50. In the same way, when we extend our sustained possession over a long enough period of time, it will eventually lead to goals.

But, I'm more playing the devil's advocate here, there is alot in play besides just math. I personally believe shot quality does account for alot, as well as player confidence and systems.

I believe the only time you can discount the math is if the Kings purposely played a system to cheat the math; they put on a high amount of low quality shots in order to inflate possession values. But, if you look at each game as a random sample, where the kings simply take the shots and opportunities they are given, the math is valid, and given a large enough sample size, the math will play out. Much in the same way that my coin analogy requires a perfect coin, in order for possession to be a solid metric, we require a perfect, random game . Even though at times it seems our system is the former scenario, especially during our slump (the eye test would say that) I FIRMLY believe the Kings play to the latter scenario (ie our numbers are a true measure of our possession).

Might of rambled a bit, its late.

 

[Trolfoli]

Sorry, flipped out about all the puck luck, hot goaltender, excuses. I guess the problem I have about regressing to the mean is that it makes it seem like the Kings offensive problems were more of random chance. Just a slump.

There have been some changes since the break. Like adding Gaborik, more depth, pushing less skilled players down the line up. I'd argue that the increase in the offense/shot % is not random, but caused by other changes.

With any mathematical model you have to look at and understand it's limitations. So the coin being flipped and coming up heads a 100 times makes me ask is this a random event or is there something interfering with the experiment? Like the Kings 3% shooting or whatever it was. This number is so far outside the norm, standard deviations would be handy here, that you have to stop and ask why is this the case and are the Kings doing something to contribute to this. Low shot quality, less skilled players playing more minutes... Not buying the Hot goalie thing, also keep in mind it was to the point that a lot of teams were playing their backup goalies against the Kings.

Another way the numbers can deviate quickly is they play style. Say P.Kane tries 4 cross ice passes and on the 4th one connects for a high quality chance. First 3 passes don't register as shots, 4th goes in... that's 100% shooting.... Now the Kings get the puck on net 4 times for low quality chances, then maybe another shot or 2 from the rebounds and they are credited with 4+ low quality chances instead. That will change your shooting % fast by cutting down on the number of shots.

With the numbers capable of swinging so much from different play styles I don't really see the league average being comparable with the Kings play style. Especially when Quick went down and the Kings went uber defensive with Jones/Scrivens in net. With the way the Kings were playing (overly defensive, with grinders in the top 6) I'd say the Kings were outside the population of the average hockey team and there's no reason to think that their shooting % should be near the league average. I guess when I get bored enough I'll see if I can find the standard deviation for the league shooting %. It'd be interesting to see just how far outside the Kings were.

 

[Ron*]

Sorry guys, you are all getting caught up in math and regression to the mean and sample size and bell curves and the value of pi and the value of e but the truth of the matter is it all comes down to effort.

We are talking about human beings and a long ****ing season and these guys know in their heads they are good enough to coast when they need to and turn on the switch when they need to.

It's not about math.

It's about effort.

 

[King'sPawn]

Sorry, flipped out about all the puck luck, hot goaltender, excuses. I guess the problem I have about regressing to the mean is that it makes it seem like the Kings offensive problems were more of random chance. Just a slump.

There have been some changes since the break. Like adding Gaborik, more depth, pushing less skilled players down the line up. I'd argue that the increase in the offense/shot % is not random, but caused by other changes.

With any mathematical model you have to look at and understand it's limitations. So the coin being flipped and coming up heads a 100 times makes me ask is this a random event or is there something interfering with the experiment? Like the Kings 3% shooting or whatever it was. This number is so far outside the norm, standard deviations would be handy here, that you have to stop and ask why is this the case and are the Kings doing something to contribute to this. Low shot quality, less skilled players playing more minutes... Not buying the Hot goalie thing, also keep in mind it was to the point that a lot of teams were playing their backup goalies against the Kings.

Another way the numbers can deviate quickly is they play style. Say P.Kane tries 4 cross ice passes and on the 4th one connects for a high quality chance. First 3 passes don't register as shots, 4th goes in... that's 100% shooting.... Now the Kings get the puck on net 4 times for low quality chances, then maybe another shot or 2 from the rebounds and they are credited with 4+ low quality chances instead. That will change your shooting % fast by cutting down on the number of shots.

With the numbers capable of swinging so much from different play styles I don't really see the league average being comparable with the Kings play style. Especially when Quick went down and the Kings went uber defensive with Jones/Scrivens in net. With the way the Kings were playing (overly defensive, with grinders in the top 6) I'd say the Kings were outside the population of the average hockey team and there's no reason to think that their shooting % should be near the league average. I guess when I get bored enough I'll see if I can find the standard deviation for the league shooting %. It'd be interesting to see just how far outside the Kings were.

While all results are independent of eachother in the coin model and corsi, the point is you are more likely to get more "heads" results if you flip a coin 50 times versus 30 times, which is the point of corsi.

You may have a lower % of results, but the overall quantity is what matters. In the end, a high shooting % is meaningless if you are taking a low number of shots.

Corsi isn't the answer to all, but teams that shoot more and have a lower than average % will still eventually yield results. As the Kings are showing now.

 

[kinghock]

Which NHL teams are the best and worst closers?

http://sports.yahoo.com/blogs/nhl-p...are-the-best-and-worst-closers-165814868.html

 

[Raccoon Jesus]

Sorry, flipped out about all the puck luck, hot goaltender, excuses. I guess the problem I have about regressing to the mean is that it makes it seem like the Kings offensive problems were more of random chance. Just a slump.

There have been some changes since the break. Like adding Gaborik, more depth, pushing less skilled players down the line up. I'd argue that the increase in the offense/shot % is not random, but caused by other changes.

With any mathematical model you have to look at and understand it's limitations. So the coin being flipped and coming up heads a 100 times makes me ask is this a random event or is there something interfering with the experiment? Like the Kings 3% shooting or whatever it was. This number is so far outside the norm, standard deviations would be handy here, that you have to stop and ask why is this the case and are the Kings doing something to contribute to this. Low shot quality, less skilled players playing more minutes... Not buying the Hot goalie thing, also keep in mind it was to the point that a lot of teams were playing their backup goalies against the Kings.

Another way the numbers can deviate quickly is they play style. Say P.Kane tries 4 cross ice passes and on the 4th one connects for a high quality chance. First 3 passes don't register as shots, 4th goes in... that's 100% shooting.... Now the Kings get the puck on net 4 times for low quality chances, then maybe another shot or 2 from the rebounds and they are credited with 4+ low quality chances instead. That will change your shooting % fast by cutting down on the number of shots.

With the numbers capable of swinging so much from different play styles I don't really see the league average being comparable with the Kings play style. Especially when Quick went down and the Kings went uber defensive with Jones/Scrivens in net. With the way the Kings were playing (overly defensive, with grinders in the top 6) I'd say the Kings were outside the population of the average hockey team and there's no reason to think that their shooting % should be near the league average. I guess when I get bored enough I'll see if I can find the standard deviation for the league shooting %. It'd be interesting to see just how far outside the Kings were.

Good post! And no worries buddy, I agree that there are plenty of other factors at play that you can't just take the big picture in a vacuum. I agree that during that slump it almost even looked like the Kings were trying to game the system but just taking 40 low-quality shots a game with no energy. There's no doubt your shooting percentage and other metrics will take a hit with that. Just like conversely the Kings look more dangerous lately and thus things are coming up. Others have covered it much more eloquently than I for sure.

Some of the issues you bring up are limitations with the advanced stats that the By The Numbers board and hockey statisticians/bloggers/nerds alike are trying to juice further. I.e. the shot quality project and other things. For example, the Leafs are the major exception to all the existing fancystats--they have a ton of guys shooting high % so they defy the Corsi/Fenwick stats to an extent. Much of it has to do with the quality of player shooting, no doubt; one shot from Kessel is probably roughly equal to infinity from Trevor Lewis. Anyway, long story short, they're not perfect and other things have to pass the eye/smell test to get a true interpretation of the stats (to paraphrase what Ron said), but over our slump, we were exposed to many things, including absolute crap luck, which inevitably had to come back around (not literally had since we're talking math, but likely had).

 

[Duc620]

Sorry guys, you are all getting caught up in math and regression to the mean and sample size and bell curves and the value of pi and the value of e but the truth of the matter is it all comes down to effort.

We are talking about human beings and a long ****ing season and these guys know in their heads they are good enough to coast when they need to and turn on the switch when they need to.

It's not about math.

It's about effort.

Still, math has its way with the universe.... even hockey.

True, it's always when and how the effort is made. I never expect a team to be 100% every night of a 80-100 game season. Especially a team with the Kings system. It's not physically, emotionally, or mentally possible. A team can only keep that up over a 20 -30 game stretch. And by that I mean keeping the whole team, every player, coach, etc, in sync and on point all the time. That's playoff.

Don't get me wrong, professionals give 100% attention and effort to whatever they're doing - that's what makes champions. But they also know when and how to get to that playoff level.

Some of us seem to expect a team to play like that all the time.