SF/SA Problem -- What fixes it?

[mr grieves]

Here's something I found concerning last season. It's basically the "regression to the mean" problem, but put a slightly different way. I hope I'm not just howling UNSUSTAAAAAAAAAIIIIIIINABLE!! and predicting doom here.

First, the problem:

In 2013, the Leafs gave up 1,441 shots (27th in the league), allowing 128 goals (17th), giving them a .917 save percentage (7th). I buy the more shots, lower quality theory as being a sustainable form of defense. But I think it means you're not playing a very good offensive game though.

In 2013, the Leafs scored 145 goals, and that was good for 6th in the league. Remarkably, they did it on only 1264 shots (28th), which gave them that 11.5 SH% (1st overall, obviously).

Now, if that last stat falls from the to the very-good-but-not-six-or-so-times-in-a-decade-amazing 9.5 SH% (or top third of league), you'd be going from a 248 GF and +29 season to a 205 GF and -13 in an 82-game season. That doesn't make the playoffs. To get back to, say, +10 you'd need to give up 23 fewer goals, or go from giving up 223 goals in 82 games to 200 GA. And if they're giving up shots at the same rate, that'd call for a .926 S%. 

While 11.5 SH% was highly unlikely (something like 6 teams have done in the last decade excluding the post-lockout PP-mania), even fewer teams -- most of them the Boston Bruins -- have posted a S% like that since the last lockout.

So, the problem, in shot:

I can't see the team's defense getting much better than it was (7th in the league is good), and I can't see them repeating the SH% they had. They need to get more shots on goal. Randy Carlyle, in fact, has said the team needs to have the puck more often and get it on the opponent's goal more often for the team to be successful next season.

The solution, then, is pretty simple:

Get more shots on goal so the SH% can be something more repeatable.

The questions that follow:

1. What moves has the team made this off-season that make you think they'll be better at getting the puck out of their zone and into the opponent's in a way that threatens a GA them?

2. What moves would you like them to make that they haven't yet that would make this even more likely than you think it now is?

 

[Kin]

I'm confused. When you say that you "can't see the team's defense getting much better than it was(7th)" what's the stat there? Save percentage? That's the defensive metric?

 

[Bill_Berg]

Gardiner will play more than 12 games this year. That's a good start.

 

[mr grieves]

I'm confused. When you say that you "can't see the team's defense getting much better than it was(7th)" what's the stat there? Save percentage? That's the defensive metric?

Yes.

 

[Kin]

Yes.

That's not, say, a goaltending metric?

 

[mr grieves]

Yes.

That's not, say, a goaltending metric?

Team save percentage. It covers both starter and back-up.

But, okay. You're saying Bernier/Reimer will combine for Thomas/ Rask numbers?

 

[mr grieves]

Gardiner will play more than 12 games this year. That's a good start.

One would hope.

One good thing to say about the off-season so far is even if Nonis spent a bunch of money making sideways (at best) moves up front, the playoffs defense is still in tact. If that group plays -- and not Fraser, Holzer, etc. -- the team will have a much better transition out of their end.

 

[Kin]

Team save percentage. It covers both starter and back-up.

But, okay. You're saying Bernier/Reimer will combine for Thomas/ Rask numbers?

Um...no? I'm saying that defense is something that's very hard to quantify. That, for a very young team like the Maple Leafs, we should expect a lot of defensive improvement to come from the young players on the roster improving.

As to the general point I'd want to see more than a big shot differential in a shortened season before I believed that there was an inherent problem that needed solving. The shot differential for the Leafs, for instance, was more than twice as bad last year as the year before. I'm not sure we, or the Leafs, should approach it as something concrete that needs immediate attention.

 

[mr grieves]

Team save percentage. It covers both starter and back-up.

But, okay. You're saying Bernier/Reimer will combine for Thomas/ Rask numbers?

Um...no? I'm saying that defense is something that's very hard to quantify. That, for a very young team like the Maple Leafs, we should expect a lot of defensive improvement to come from the young players on the roster improving.

As to the general point I'd want to see more than a big shot differential in a shortened season before I believed that there was an inherent problem that needed solving. The shot differential for the Leafs, for instance, was more than twice as bad last year as the year before. I'm not sure we, or the Leafs, should approach it as something concrete that needs immediate attention.

Okay. Then set aside the shot differential for a moment. Is it not concerning that there were so few SF and the team only scored enough to make the playoffs because of a remarkably high SH%?

 

[Kin]

Okay. Then set aside the shot differential for a moment. Is it not concerning that there were so few SF and the team only scored enough to make the playoffs because of a remarkably high SH%?

Yes and no. If we take shooting percentage as a heavily influenced by randomness number then it's clearly problematic. Heck, even if it's only a slightly random it's not good. That said:

1) I'd need to see things stay that way for longer than 48 game before I thought it was an inherent flaw in the composition of the club that needed to be addressed via personnel changes.

2) To the extent that personnel changes can help in that regard, I think the ones they've made have been largely positive.

3) I'd need to see the completed roster before I considered them deficient anywhere.

 

[mr grieves]

Okay. Then set aside the shot differential for a moment. Is it not concerning that there were so few SF and the team only scored enough to make the playoffs because of a remarkably high SH%?

Yes and no. If we take shooting percentage as a heavily influenced by randomness number then it's clearly problematic. Heck, even if it's only a slightly random it's not good.

But that's not really incompatible with my point. The Leafs team managed the SH% equivalent of rolling double sixes. It very well might not happen again. Those numbers move around a lot. They don't often move much higher than 11.5%.

Which is why I'd think the sensible thing to do would be to control what you can: shots on goal.

That said:

1) I'd need to see things stay that way for longer than 48 game before I thought it was an inherent flaw in the composition of the club that needed to be addressed via personnel changes.
2) To the extent that personnel changes can help in that regard, I think the ones they've made have been largely positive.
3) I'd need to see the completed roster before I considered them deficient anywhere.

So not as significantly tied to personnel as I think... Okay. Then what's the systems change that gets more shots on goal? Or do think they're awfully random too?

 

[Kin]

But that's not really incompatible with my point. The Leafs team managed the SH% equivalent of rolling double sixes. It very well might not happen again. Those numbers move around a lot. They don't often move much higher than 11.5%.

Which is why I'd think the sensible thing to do would be to control what you can: shots on goal.

But that's what I'm saying. I'm not certain that the evidence is great that shooting percentage is a random number that should be likened to a roll of the dice as opposed to being, in large part, a result of the abilities of the players on a team. It may dip and weave but that's not the same thing as saying it's wildly random.

So not as significantly tied to personnel as I think... Okay. Then what's the systems change that gets more shots on goal? Or do think they're awfully random too?

Well, like I said, a lot of the improvement is going to have to come internally through players already in the organization. That's just a fact that comes from a young team growing and developing. You can say "but what if it doesn't" but if the young players on the team aren't going to take significant steps forward then the team is in a bad spot no matter who they signed today.

Seriously, you need to step back from your weird soapbox because it's messing with your basic ability to hold a conversation. You're the one saying shooting percentages are random and the idea that I'm anti-statistical analysis in sports is, well, people who've been here longer than you can tell you how wrong that is. I just don't like bad stats.

 

[Bill_Berg]

Clarkson was 8th in the league with 3.75 shots per game last year, if that makes you feel better.

 

[mr grieves]

But that's not really incompatible with my point. The Leafs team managed the SH% equivalent of rolling double sixes. It very well might not happen again. Those numbers move around a lot. They don't often move much higher than 11.5%.

Which is why I'd think the sensible thing to do would be to control what you can: shots on goal.

But that's what I'm saying. I'm not certain that the evidence is great that shooting percentage is a random number that should be likened to a roll of the dice as opposed to being, in large part, a result of the abilities of the players on a team. It may dip and weave but that's not the same thing as saying it's wildly random.

The opposite of a dip is a peak. And, given what you can see among 30 teams -- several more offensively talented than the Leafs -- over 9 years, 11.5% is a peak. 30 teams x 9 years of SH% stats show that.

If I'd thought shooting percentage were "wildly random" or as variable as rolling dice, I would've been worried they'd have shooting 4% next season. Rolling dice was a metaphor for hitting an unlikely peak, not SH% generally. I said from the outset 'let's say 9.5%' because that seems to be what the top 3rd of league generally shoots at. That seemed reasonable because reasonable people would put the Leafs there based on the talent on their roster.

So, yes. I know SH% isn't random, it is related to the talent of players taking shots. But it also varies somewhat. The only premise I was asking you grant me is that it's more likely than not that the Leafs will convert shots at the rate that a top-10 offensive team tends to next season.

Seriously, you need to step back from your weird soapbox because it's messing with your basic ability to hold a conversation. You're the one saying shooting percentages are random and the idea that I'm anti-statistical analysis in sports is, well, people who've been here longer than you can tell you how wrong that is. I just don't like bad stats.

Yeah, there's no soapbox here. And your taking a simple and not controversial argument -- "a 9.5 SH% is more likely than 11.5 SH%" -- to mean "heavily influenced by randomness" isn't helping further any sort of conversation. Unless you think the Leafs have stealthily put together among the best roster of shooters in the league since the lockout, you can't seriously think 11.5% isn't going to drop somewhat -- especially in a year where Nonis is stealthily dismantling the roster.

And I don't think you're "anti-statistical analysis of sports" -- I just think you're happy to not bother looking at any numbers, and to dismiss those that exist, when you've chosen form an opinion using other forms of evidence. 

 

[mr grieves]

Clarkson was 8th in the league with 3.75 shots per game last year, if that makes you feel better.

It helps.

So does the playoffs (hopefully) convincing Carlyle that Franson and Gardiner deserve to play a lot more than they did last season.

But it's also somewhat concerning that, at the same time, management dismantled a line that was excellent at controlling play from 2010/11 to 11/12.

Not sure whether the team comes out ahead.

 

[Kin]

The opposite of a dip is a peak. And, given what you can see among 30 teams -- several more offensively talented than the Leafs -- over 9 years, 11.5% is a peak. 30 teams x 9 years of SH% stats show that.

Except I don't think that shooting percentage is necessarily a straight link to offensive talent either. 11.5 might be a high number but I think a team having a higher shooting percentage than another given a roughly comparable level of offensive talent can be a result of design/execution.

So, yes. I know SH% isn't random, it is related to the talent of players taking shots. But it also varies somewhat. The only premise I was asking you grant me is that it's more likely than not that the Leafs will convert shots at the rate that a top-10 offensive team tends to next season.

Sure, and I addressed that. I don't think that this year's relatively low shot totals is necessarily something inherently problematic going forward and to the extent that it is, I think the personnel changes they've made help in that regard.

Unless you think the Leafs have stealthily put together among the best roster of shooters in the league since the lockout, you can't seriously think 11.5% isn't going to drop somewhat -- especially in a year where Nonis is stealthily dismantling the roster.

Again, you're not actually responding to what I'm saying. I'm fine with the idea that shooting percentage isn't a constant. I even allowed for the different opinions on the randomness of the number rather than taking a position on it.

My position, and this is now the third time I've said it, is that I don't think that the shot total or shot differential this year is necessarily set in stone going forward.

And I don't think you're "anti-statistical analysis of sports" -- I just think you're happy to not bother looking at any numbers, and to dismiss those that exist, when you've chosen form an opinion using other forms of evidence.

But that's the fundamental problem with the position you keep taking. I can look up Corsi. I can look at ES P/60. In order to refute them as good concepts I have to be familiar with them which is what I've been doing.  When I say things like ES P/60 is bogus because the results clearly contradict what we've seen.

I appreciate that people want to turn things into baseball. The statistical analysis of baseball has tremendously broadened understanding of the game. The problem is that the people big into the advanced numbers in hockey have learned all of the bad lessons from that movement in baseball and none of the good ones.

Being able to reject lousy analytics is fundamental to the process. Hockey folk haven't picked that up yet.

 

[mr grieves]

The opposite of a dip is a peak. And, given what you can see among 30 teams -- several more offensively talented than the Leafs -- over 9 years, 11.5% is a peak. 30 teams x 9 years of SH% stats show that.

Except I don't think that shooting percentage is necessarily a straight link to offensive talent either. 11.5 might be a high number but I think a team having a higher shooting percentage than another given a roughly comparable level of offensive talent can be a result of design/execution.

Yes. I've only really watched the Ducks in the finals, so don't really know if RC's instituting the same system there that he's used elsewhere. If RC helps make up the design that's executed the players, it might be worth knowing his team's shooting percentages:

2005/06 -- 9.7% (21st), with fifth in SF
2006/07 -- 9.8% (12th), 7th in SF
2007/08 -- 8.8% (24th), 25th in SF
2008/09 -- 9.6% (13th), 12th in SF
2009/10 -- 9.4% (9th), 15th in SF
2010/11 -- 10.1% (1st), 27th in SF
2011/12 -- fired, so dropping it.
2012/13 -- 11.5% (1st), 28th in SF

With available talent plus how RC's game changes, along with several other variables, these numbers mightn't mean much to you. But make of them what what you will.

Unless you think the Leafs have stealthily put together among the best roster of shooters in the league since the lockout, you can't seriously think 11.5% isn't going to drop somewhat -- especially in a year where Nonis is stealthily dismantling the roster.

Again, you're not actually responding to what I'm saying. I'm fine with the idea that shooting percentage isn't a constant. I even allowed for the different opinions on the randomness of the number rather than taking a position on it.
[/quote]

My position, and this is now the third time I've said it, is that I don't think that the shot total or shot differential this year is necessarily set in stone going forward.
[/quote]

And, as should've been evident from the existence of this thread, neither do. My question was what roster moves, if any, help move it in a better direction? What untapped potential on the roster or likely-to-rebound off-seasons will do so? And, after you suggested they weren't decisive one way or the other, what changes in the way the Leafs play do you think will move it in a better direction?

 

[Kin]

And, as should've been evident from the existence of this thread, neither do. My question was what roster moves, if any, help move it in a better direction?

The additions of Bolland, Clarkson and Bernier.

What untapped potential on the roster or likely-to-rebound off-seasons will do so?

A year of growth from the prominent young players on the roster. Chiefly Kadri, Gardiner and JVR.

And, after you suggested they weren't decisive one way or the other, what changes in the way the Leafs play do you think will move it in a better direction?

This seems like the kind of question that can really only be answered once we have a much better picture of what the team's actual roster is.

 

[mr grieves]

And, after you suggested they weren't decisive one way or the other, what changes in the way the Leafs play do you think will move it in a better direction?

This seems like the kind of question that can really only be answered once we have a much better picture of what the team's actual roster is.
[/quote]

"The good news for me is, it's not my problem any more, that's up to Randy," Nonis quipped when questioned on the breakdown of his forward complement with the Clarkson addition. "I've done what I could do."

I don't know how much more the roster is going to change.

But it's a simple question that it can, in part, be answered without waiting to see if D'Amigo's on the third or fourth line: how does Randy Carlyle hockey get enough shots on goal that you don't need a 11.5% shooting percentage to make the playoffs? Maybe it calls for someone who watched the Ducks more closely than anyone here has.

All's I know is I saw one season of it with a roster that seems to have had more offensive talent than this one. I'm curious is all.

 

[Kin]

I don't know how much more the roster is going to change.

There's still room for it. I don't think we know what the back-end of the defense is. I don't think we know what the bottom six is going to be. Guys can surprise at camp and force their way onto a club prompting trades then. I wouldn't take what Nonis said as gospel because I can guarantee that if you asked him if he'd listen to trade offers right now he'd say yes.

But it's a simple question that it can, in part, be answered without waiting to see if D'Amigo's on the third or fourth line: how does Randy Carlyle hockey get enough shots on goal that you don't need a 11.5% shooting percentage to make the playoffs?

But I've answered that. Improvement throughout the line-up both via development and new additions. Another year of getting acclimated to the system he wants to play. Carlyle doesn't need to come in with a completely different playbook to make that happen even if we accept the premise, and just to be clear I don't accept this premise, that the low shot total from last year wasn't just as much of a statistical aberration as the shooting percentage was.