Sheng Peng: Striking Stats: Kings 1/2 Season Review

Nice work Peng. I really did notice Brown pick up his physical game this year, mostly to start the year. Feel like hes kind of gone unnoticed. But perhaps that's a good thing.

First 10 mins and Muzzin is playing like hes heard his name in the trade rumors.

the transition game was as good as i've seen this season. crisp passes, tape to tape, hitting guys in stride. a bit of a lull in the third period, but a large lead will do that.

in regards to the blog TL;DR - fancy stats bore the crap out of me but I do see in a way, how they factor in, but damn do I hate them. I am sure it was a brilliant article though.


Who was that team that showed up last night and rang up 5 goals?

One thing that was a slight factor but not the main reason I saw such a good game last night: No Matt Greene or Tom Gilbert. as much as I think McNaab sucks the D core looked real good last night.

in regards to the blog TL;DR - fancy stats bore the crap out of me but I do see in a way, how they factor in, but damn do I hate them. I am sure it was a brilliant article though.


Who was that team that showed up last night and rang up 5 goals?

One thing that was a slight factor but not the main reason I saw such a good game last night: No Matt Greene or Tom Gilbert. as much as I think McNaab sucks the D core looked real good last night.
- MikeOxbyg

He hasn't been playing as well since coming back as he was before his injury, that's for sure. I'd like to give him a little more time to see if he can get back to form. Early on this season he was outplaying all the other D for more than a couple games.

He hasn't been playing as well since coming back as he was before his injury, that's for sure. I'd like to give him a little more time to see if he can get back to form. Early on this season he was outplaying all the other D for more than a couple games.
- PancakesPenner

I'd have to agree. Before the injury he was checking and hitting like Blake use to hit.
Give him some time to get the confidence back. Thanks for the write up Sheng!

King's 5v5 possession numbers have lagged without Kopitar (as we'll see shortly). However, his shorthanded work has remained robust. His 47.85 4v5 Fenwick Against/60 is second-lowest among team forwards and the NHL (out of 113 forwards, 50+ 4v5 minutes). Essentially, teams aren't getting a lot of chances on King while he's killing penalties.
- Sheng Peng

Shore is first on the team and in the league with a 47.75 4v5 Fenwick Against/60 (out of 113 forwards, 50+ 4v5 minutes).
- Sheng Peng

That said, Muzzin is still doing a lot things right defensively. His +2.79 5v5 Relative Corsi For % and 53.27 4v5 Fenwick Against lead the defense.
- Sheng Peng

To be fair, I'm sure that that's due to them being on the 2nd PK unit. All of the 2nd-unit penalty killers have better 4v5 Fenwick Against/60 than the 1st unit killers. Kopitar and Doughty are the worst on the team in that stat not because they're the worst on the team at penalty killing, but because they always get the other team's 1st PP unit when it's fresh. The 2nd-unit PKers get the other team's 2nd-unit PPers or the 1st-unit PPers when they're dead tired.

Also, with respect, a good Corsi % doesn't mean that you're doing well defensively. Forbort's and Muzzin's Corsi Against/60 are practically equal, suggesting that they're equally effective at limiting overall shot attempts against, but Muzzin has far more shot attempts for, enough that he leads the team in Corsi % because of it while Forbort is one of the worst on the team in that stat. If you looked at that and concluded that Muzzin is playing well defensively and Forbort isn't, you'd be wrong. The difference in that stat for them is how much more involved in offense Muzzin is. If you're an offensive player like that who believes in doing what it takes to get shots on net, even if they're not of the highest quality--i.e. if you have that "shot mentality" that you discussed--then you're liable to have good Corsi %. That's why Dustin "No shot is a bad shot" Brown has typically been one of the team's Corsi leaders.

If you want to evaluate a player on just defense, you shouldn't use Corsi or Fenwick %, IMO. You should look at purely defensive stats, like Scoring Chances Against/60 and SV% (both of which Muzzin is last among defenseman on the team in, suggesting that he's not doing well defensively). Those are much better indicators of how well a player is defending, since they're 100% defensive and absolute stats, not just 50% and relative.

Nice work
- poisondhearts37

Thanks!

Thanks for the write up Sheng!
- arh777

You're welcome!

To be fair, I'm sure that that's due to them being on the 2nd PK unit.

Also, with respect, a good Corsi % doesn't mean that you're doing well defensively.

You should look at purely defensive stats, like Scoring Chances Against/60 and SV% (both of which Muzzin is last among defenseman on the team in, suggesting that he's not doing well defensively).
- Osprey

Re: Shore & King, absolutely, I wasn't trying to imply that they were the best penalty killers on the team. Their results are definitely helped by their weaker competition -- that said, being #1-2 in the league means they're also being ranked with other second-unit penalty killers in the league and besting them. It's an impressive stat -- to a point.

Many times, best defense is a good offense. I certainly believe in "traditional" defense -- down-low battles, sticks in lanes, blocked shots -- which Muzzin does well -- but I also believe "traditional" defenders also need to be competent puckmovers. Not great, just competent.

From 2013-16, Muzzin is second among LA defensemen in SCA60 (behind Greene), by the way. As for this season, your SCA60 leaders on defense are Gravel & Gilbert, which suggests deployment -- Gravel & Gilbert don't get as many defensive zone starts -- means a lot. Whereas your three worst defenders, according to just this stat, are Martinez, Doughty, and Muzzin. Are you sure that you want to be stressing SCA60?

Personally, I don't bother much with SV%. Too many factors out of defender's control to use that.

From 2013-16, Muzzin is second among LA defensemen in SCA60 (behind Greene), by the way. As for this season, your SCA60 leaders on defense are Gravel & Gilbert, which suggests deployment -- Gravel & Gilbert don't get as many defensive zone starts -- means a lot. Whereas your three worst defenders, according to just this stat, are Martinez, Doughty, and Muzzin. Are you sure that you want to be stressing SCA60?
- Sheng Peng

Deployment does mean a lot; I agree. If you look at the seasons individually, instead of summing them up, you'll see that Muzzin's best SCA60 was in his rookie season (kind of like how Gravel leads now) and he's gotten worse in that stat each season since (5.21->6.31->6.44->7.11->8.03). His two worst marks in SCA60 have come in these last two seasons, when he's been deployed as a #2-3 (and separated from Doughty). This all suggests that he was fairly good when he was being treated and sheltered like a #4-6 defenseman, but has fallen off as he's been used in a role (as a #2-3) that's at least a little more than he's cut out for.

The fact that a player does well in a limited, lower role doesn't always mean that he's capable of handling a big promotion to an important role. It would really behoove LA to find one or two solid defensemen (like they had back when they were winning Cups) to force the rest (not just Muzzin, but Forbort and even McNabb) down a notch or two into roles that are better suited for their abilities.

Personally, I don't bother much with SV%. Too many factors out of defender's control to use that.
- Sheng Peng

I disagree. The only factor that SV% adds over something like SCA60 is the goaltender, but whether the goaltender makes the save or not is very much in the control of the defender because the defender has a big effect on the quality of shots and the quality of shots tends to have a big effect on SV%. It's especially so because a goaltender's performance is consistent across the team. All defenders on a team play in front of the same goaltender. If there's a noticeable difference in SV% between the defenders, then the only explanation is that some defenders are allowing better quality shots than others.

Muzzin's SV% has been consistently near the bottom of the defenders for many seasons now, which goes against the idea that it's out of his control. In fact, I pointed out above how Muzzin's SCA60 has been going up since his rookie season. Well, his SV% has been going down since his rookie season, as well. That's too perfect of a complement to be coincidence. If SV% were really out of a defender's control, it would tend to be all over the map, not consistently bad or consistently proportional to a stat that is within his control. Yes, deployment affects both stats, but that's the opposite of having no control, since it means that your presence and usage makes a difference.

Deployment does mean a lot; I agree. If you look at the seasons individually, instead of summing them up, you'll see that Muzzin's best SCA60 was in his rookie season (kind of like how Gravel leads now) and he's gotten worse in that stat each season since (5.21->6.31->6.44->7.11->8.03). His two worst marks in SCA60 have come in these last two seasons, when he's been deployed as a #2-3 (and separated from Doughty). This all suggests that he was fairly good when he was being treated and sheltered like a #4-6 defenseman, but has fallen off as he's been used in a role (as a #2-3) that's at least a little more than he's cut out for.
- Osprey

My point is that SCA60 doesn't tell you what you think it does. Starting from Muzzin's deployment as #2 dman in 2014-15, Muzzin's SCA60 was 6.44 -- Doughty's was 7.70. In 15-16, Muzzin's was 7.11, while Doughty's was 7.54. By your rationale, is Doughty not cut out to be a #1 dman?

When I cited Muzzin's 2012-17 SCA60 total with Greene leading in this category, my point was to show that a low SCA60 doesn't necessarily mean best defensive defenseman. Just as we agree that Shore and King's very low FA60 doesn't mean they're the best penalty killers in the league.

If there's a noticeable difference in SV% between the defenders, then the only explanation is that some defenders are allowing better quality shots than others.

That's too perfect of a complement to be coincidence.
- Osprey

We will simply disagree on this. There are so few goals relative to shots and even scoring chances, which is why most advanced stats proponents emphasize the larger sample sizes. There are some who care about Save %. I'm not one of them.

I kind of hate to reply again, and almost didn't, since we disagree and we don't want to beat a dead horse, but there isn't a new blog or topic to comment on, so I apologize if this is arguing for the sake of arguing. There's nothing else to do today.

My point is that SCA60 doesn't tell you what you think it does. Starting from Muzzin's deployment as #2 dman in 2014-15, Muzzin's SCA60 was 6.44 -- Doughty's was 7.70. In 15-16, Muzzin's was 7.11, while Doughty's was 7.54. By your rationale, is Doughty not cut out to be a #1 dman?
- Sheng Peng

From 2009-10 (his sophomore season and first season as #1) through 2012-13, Doughty didn't once have a SCA60 over 6.55. Since then, he hasn't once had a SCA60 under 7.52. A huge jump of almost a full point occurred once he was paired with Muzzin and he's maintained 7.52-7.70 through Muzzin, McNabb and Forbort. It seems pretty clear to me that those three players have been dragging him down defensively, which is not surprising, considering that they were all in their first or second full seasons when they were thrown into the fire on the top pairing. In judging Doughty's defensive stats, you really have to take whom he's been paired with into consideration.

Also, I wasn't suggesting that one use SCA60 by itself. I mentioned SV% in the same breath because they complement each other and because those are two examples of purely defensive stats. The point was that you can't read much into Corsi as a defensive stat because it isn't purely one. Only half of what goes into Corsi is defense; the other is offense. You can't just quote a good Corsi figure and suggest that it means that someone is good at defense. It can just as easily mean that the player is average or even mediocre at defense, but good or even great at offense. The way that you can tell which it is is to take a look at other stats that are more purely offensive or defensive, such as SCA60 and SV%, among many others.

We will simply disagree on this. There are so few goals relative to shots and even scoring chances, which is why most advanced stats proponents emphasize the larger sample sizes. There are some who care about Save %. I'm not one of them.
- Sheng Peng

You did, in this very blog, liberally cite Shooting %, Expected Goals For %, Points/60, 5v5 Goals For / 60, 4v5 Goals Allowed / 60, Penalty Kill % and Powerplay %, indicating that they are all meaningful (for assessing both team and individual performance), despite goals being few and far between. You even cited Budaj's SV% as the main bright spot for him, even though it's the exact same measurement as a defenseman's SV%. If you care about SV% for goaltenders, but not for defensemen, that could be justifiable, but if the reason for dismissing defensemen SV% is that it's based on goals, then that makes no sense. Similarly, to emphasize all of those other goal-based stats and then suggest that a different goal-based stat isn't meaningful because it's based on goals, which are far fewer in number than shots and scoring chances, doesn't many sense, either.

Large sample sizes are great, but they're only meaningful when you're gauging the frequency of something much smaller within that sample size. For example, how meaningful is it to know that a player took 1000 shots on goal in a full season unless you know how many actually went in the net? If only 10 went in, that's a miserable shooting % of 1% and it suggests how worthless those player's shots were. If you were to argue that 10 goals is too small of a "sample size" to draw a decent conclusion, you would be completely missing the point, since the true "sample size" here is the full season and the sub-"sample size" is the 1000 shots, plenty of data to draw a likely very accurate picture of the player's shooting ability, even if it looks like you're basing it on "only" 10 goals. Similarly, that's why it doesn't make sense to disregard other goal-based stats. They're always measured as goals out of something much larger (ex. a high quantity of shots, scoring chances, TOI or the season), and it's that larger quantity that is the sample size, not the number of goals.

I kind of hate to reply again, and almost didn't, since we disagree and we don't want to beat a dead horse, but there isn't a new blog or topic to comment on, so I apologize if this is arguing for the sake of arguing. There's nothing else to do today.

From 2009-10 (his sophomore season and first season as #1) through 2012-13, Doughty didn't once have a SCA60 over 6.55. Since then, he hasn't once had a SCA60 under 7.52. A huge jump of almost a full point occurred once he was paired with Muzzin and he's maintained 7.52-7.70 through Muzzin, McNabb and Forbort. It seems pretty clear to me that those three players have been dragging him down defensively, which is not surprising, considering that they were all in their first or second full seasons when they were thrown into the fire on the top pairing. In judging Doughty's defensive stats, you really have to take whom he's been paired with into consideration.

Also, I wasn't suggesting that one use SCA60 by itself. I mentioned SV% in the same breath because they complement each other and because those are two examples of purely defensive stats. The point was that you can't read much into Corsi as a defensive stat because it isn't purely one. Only half of what goes into Corsi is defense; the other is offense. You can't just quote a good Corsi figure and suggest that it means that someone is good at defense. It can just as easily mean that the player is average or even mediocre at defense, but good or even great at offense. The way that you can tell which it is is to take a look at other stats that are more purely offensive or defensive, such as SCA60 and SV%, among many others.

You did, in this very blog, liberally cite Shooting %, Expected Goals For %, Points/60, 5v5 Goals For / 60, 4v5 Goals Allowed / 60, Penalty Kill % and Powerplay %, indicating that they are all meaningful (for assessing both team and individual performance), despite goals being few and far between. You even cited Budaj's SV% as the main bright spot for him, even though it's the exact same measurement as a defenseman's SV%. If you care about SV% for goaltenders, but not for defensemen, that could be justifiable, but if the reason for dismissing defensemen SV% is that it's based on goals, then that makes no sense. Similarly, to emphasize all of those other goal-based stats and then suggest that a different goal-based stat isn't meaningful because it's based on goals, which are far fewer in number than shots and scoring chances, doesn't many sense, either.

Large sample sizes are great, but they're only meaningful when you're gauging the frequency of something much smaller within that sample size. For example, how meaningful is it to know that a player took 1000 shots on goal in a full season unless you know how many actually went in the net? If only 10 went in, that's a miserable shooting % of 1% and it suggests how worthless those player's shots were. If you were to argue that 10 goals is too small of a "sample size" to draw a decent conclusion, you would be completely missing the point, since the true "sample size" here is the full season and the sub-"sample size" is the 1000 shots, plenty of data to draw a likely very accurate picture of the player's shooting ability, even if it looks like you're basing it on "only" 10 goals. Similarly, that's why it doesn't make sense to disregard other goal-based stats. They're always measured as goals out of something much larger (ex. a high quantity of shots, scoring chances, TOI or the season), and it's that larger quantity that is the sample size, not the number of goals.
- Osprey

Yea, I think we just disagree here. Not too much more to say.

I will note that xGF, despite its name, is not a goal-based stat, it's a shot-based one.

I will stress again, just so you know where I'm generally coming from, that I believe the best defense is a good offense (which doesn't necessarily mean scoring) -- which is why I stress CF% over just CF or CA. Also, goals being the result of a process, I generally stress "process" stats -- while I do cite goal-based stats from time to time, I usually believe the process tells us more.

Just call me Sam Hinkie lol.

For my "Striking Stats" pieces, I just pick up on a variety of interesting stats, some with more meaning to me than others. Just because I mention it doesn't mean I value it much -- such as Brown's Hits/Game.

Anyway, I do really appreciate your thoughtful, measured responses, but I figure we're just at an impasse. Hopefully, I'm not being too preachy, but there's nothing wrong with that! We're both trying to be smarter about the game we love, and honestly, nobody has all the answers, even the guys who have won Cups.