Capital Trends Game 20

20 games in the books and almost a quarter of the season’s worth of data to analyze. Let’s have a looksie under the hood.

Nothing is quite as sobering to a hockey fan as a regulation loss to the Buffalo Sabres. A team that is pushing the envelope in suckiness. The whipped cream on the shit sundae that is that loss, is that it capped the Sabres’ current three game winning streak. Which is coincidentally the longest winning streak the Caps have held this season (all of once). So, what I’m trying to say is that the optimism is gone. I’m not in full fledged titanic sinking mode, but there’s evidence out there to suggest that teams not in a playoff position come Thanksgiving, often don’t change their fortune.

Welp, onto the plots. Remember any observations or comments I make are strictly on the data presented in the graph discussed. Not making sweeping generalizations here.

The recent games have posed a troubling pattern of not dominating possession like I had foolishly gotten used to during the beginning of the season.

YTD_possession_game_20 rolling_possession_game_20
As to be expected, when you don’t win the possession battle, your cumulative season possession drops. The 10-game rolling plot shows a much more troubling fall from greatness, down to just ok. Still better than last season though. And at least the Buffalo and Colorado games provided a nice spring board to get it going in the right direction.

YTD_PDO_game_20 rolling_PDO_game_20
The Caps are still looking a little unlucky. Keeping things under 100, and even under 99 in the more immediate data. Let’s check back with the breakdown of this.

YTD_percentages_game_20 rolling_percentages_game_20
Well, I bet no one predicted that Holtby would begin to bounce back to respectable levels like he’s shown capable of before. Weird, that. But seriously it is nice to see that save percentage coming back, and in a hurry too. The shooting percentage though should be no surprise, the forwards are still struggling to get the puck in the net.

YTD_shots_game_20 rolling_shots_game_20
The shots for the Caps are looking steady, but the shots against are not conducive to winning, unless Holtby goes beastmode. Maybe those shots for will start to translate to goals…?(!)

YTD_ES_shots_game_20 rolling_ES_shots_game_20
These plots would lend evidence to the Caps doing something different during 5’s. I was hoping that the spike in shots against was due to the number of PKs they’ve been subjected to lately. But it would appear to be something with their even strength play. There are even stretches where they’ve been getting outshot by their opponents.

YTD_ST_game_20 rolling_ST_game_20
Special teams are still nothing to complain about. Especially the current state in the 10-game scheme of things. Seriously, a team with the possession metrics the Caps are sporting and an STI over 107% should be running away with a playoff spot. Alas, I said should, not are.

YTD_PP_Performance_game_20 YTD_PP_Shots_game_20 rolling_PP_Performance_game_20 rolling_PP_Shots_game_20
Power play still looks pretty solid on all fronts. Except for maybe opportunities. And I guess the shooting percentage is down a bit, but they’re still converting on nearly 1/4 of their opportunities. Which is certainly not what’s been ailing them.

YTD_SH_Performance_game_20 YTD_SH_Shots_game_20 rolling_SH_Performance_game_20 rolling_SH_Shots_game_20
Maybe I spoke too soon about how they’ve been subjected to quite a few PKs. They’ve actually been getting penalized less lately. They’re still penalized more than their opponents, but that can’t account for the unimpressive standings position. They’re still suppressing shots like crazy, and the goalies have been doing their part. 88% should be a solid shorthanded save percentage to maintain. That’s usually close to the league average for a season.

YTD_fenwick_score_game_20 rolling_fenwick_score_game_20
Man, they’ve really been taking their foot off the pedal when they have the lead lately. There hasn’t been a whole lot of qualifying TOI to swing that 2+ data (i.e. they haven’t been leading by 2 or more all that often), but the up 1 data is in a nosedive. Tied was looking pretty dire too, until they got to play the teams at the bottom of the possession boards. I guess to get a better idea of what’s really happening it could be useful to look at events for and against.

YTD_fenwickfor_score_game_20 rolling_fenwickfor_score_game_20
Well they certainly have been forsaking offense lately when leading on the scoreboard. In ┬áthe past 10 games, they’ve dropped their Fenwick-For events per 60 by about 24%. That’s…not good. Let’s see if they’re at least decreasing the attempts against.

YTD_fenwickagainst_score_game_20 rolling_fenwickagainst_score_game_20
Uh…that’s a big nope. In fact, over the past few games they’ve been allowing shots more so than in the past. So, either they’re really bad at the defensive shell, or the other teams are just really motivated, or just spot the Caps a lead because they know they’ll be able to walk over them after they’re leading. That negative change in the possession also seems to stem mostly from the extra Fenwick events allowed against

I’m just going to throw all the same kind of plots for Corsi below without really commenting on them. They’re just there for your perusal.

YTD_corsi_score_game_20 rolling_corsi_score_game_20
YTD_corsifor_score_game_20 rolling_corsifor_score_game_20
YTD_corsiagainst_score_game_20 rolling_corsiagainst_score_game_20

And then just to look at 4 major aspects of the game in one plot:



So there you have it. Just about every bit of data you could want to look at. You got any ideas why the Caps are so mediocre? I’m having trouble discerning anything, other than their penchant for soul-crushing boners.



Capital Trends Game 12

It’s time for the inaugural Capital Trends post for the 2014-2015 season for the Washington Capitals. We’re twelve games into the season and the results are conflicted with my general opinion of the on-ice product. The Capitals haven’t set the league on fire in the standings. But they have been in the fancy stats world. This is why you may see such a disparity in the power rankings done across the internet. From 23 (out of 30) to 7, the Washington Capitals have good underlying numbers, but not good standings points. If they were playing in the CORSI Hockey League they’d probably be top 5 in all power rankings. But much to Mr. Simmons’ delight, that’s not a real thing.

So I wish I had a plot that fully captures just how bad of a coach Adam Oates was. But you’d have like 30 different lines on the plot all telling you the same thing, this year’s version of the Washington Capitals is far superior to last year’s. Many of the early returns of the coaching change and upgrade on defense have been very promising. At least that’s what I’m telling myself even if the standings don’t necessarily reflect it just yet. Things are promising enough that I find myself rather optimistic for this season, even if I rage quit on a game here and there. And even though the team is only on pace for 75 points and another season of missing the playoffs, I’m not angry. Frustrated, maybe. Disappointed, surely. But let’s get to the data.

This first take won’t include any playoff pacing, or rolling game data since the season is still young. Which should also serve as encouragement to step off that ledge fellow Caps fans.

Without ExtraSkater around, I kind of missed the game-by-game bargraph of Fenwick Close differential. So I made it, and added a different dimension.

The score-adjusted differential is just the score-adjusted percentage multiplied by the number of events that qualify. The score-adjusted percentage formula can be found here, I used Tulsky’s values for average percentages and time on ice. It’s definitely not correct in the sense that it’s actually adjusting for league averages, but I’m not collecting league-wide data, and it’s better than nothing. But you’ll notice, the Caps have had the puck a lot more than last season.

It’s pretty crazy to look back at how woeful the Caps were as a possession team last year and then to look at this plot and to see them above 55% (a really good number) in almost every way I can think of to measure possession using shot attempt proxies. The score-adjusted Fenwick gets dragged down by the Caps’ tendencies when holding a lead as you’ll see later.

Ah, so this is where you can see a precipitous fall that coincides well with the team’s current losing streak. But let’s look at what is driving the bus here:

Well the shooting percentage is looking good. Maybe a bit of a dip in even strength shooting %. But good god, that save percentage. I’m gonna go on record and say that is unsustainably bad. Holtby is far better than he’s playing. And the team is still adjusting to new systems and teammates. I’m holding out hope that Holtby will rebound nicely. But I can’t tell you when that will happen unfortunately.

Here’s another graph that shows Adam Oates had a pretty awful system in place. The shots against are way down from last year. And the Caps are outshooting their opponents by a large margin, currently by about 5.5.

Here’s a look at how they’re doing at even strength. Still quite good.

The special teams aren’t anything really worth commenting. They’re doing better than the rest of the league overall with an STI > 100. The special teams PDO is pretty high though at 1027, but with the best shot in the league manning the left faceoff circle, that doesn’t seem unreasonable.

YTD_PP_Performance_game_12 YTD_PP_Shots_game_12
And a look at the power play is about what you’d expect. The Caps are still about where they were last year in this regard, except they may be getting a few less power plays.

YTD_SH_Performance_game_12 YTD_SH_Shots_game_12
The shots allowed while shorthanded were starting to spike a bit, especially at the Coyotes game, but they brought it back down with a better effort against Calgary. Either way, these numbers are better than the historically bad numbers from last season. Woof.

And then just a look at most aspects of the team’s performance in one graph. That point pace drop isn’t pretty.


Now I wanted to take a deeper look into why they may be losing. Surely it can’t be just that the goaltending is atrocious (it has been though). I wanted to look at how the team is performing in different score scenarios to see if there’s a discernible pattern there.

YTD_fenwick_score_game_12 YTD_corsi_score_game_12
I was really hoping to glean something out of these. Especially since the team has held a lead in each of the games they’ve lost. But there’s really nothing to note here. They shoot the puck more than half the time in all represented scenarios, except when they’re up by 2 or more. There hasn’t even really been any noticeably bad trends the past five games, other than their tied-score performance coming down to a dreadful 55% (!). I guess that means I’m all aboard the goaltending-will-right-itself train, until it doesn’t.




About this Site

So I wanted to move away from my old blog (Laich the Lamp) for a couple reasons. One was because I wanted to not have links to the garbage I wrote last year, in an attempt to make the content available more appealing and semiprofessional. The other reason being that the name really started to bother me. So I started this blog to focus more on just the plots I’ve been generating. And to hopefully have a name that has some staying power; I don’t want to turn into Andy Dwyer from Parks and Recreation (Mouse Rat forever!).

So the posts won’t be at published at any particular pattern, if you like what you see, check back occasionally and maybe there will be new content. I have trouble finding the time to provide commentary and actually saying worthwhile things in a post. If people don’t care what I have to say, I can just make posts where I publish only the plots and shut my trap. We’ll see where this goes, but for now posts may be kind of sporadic. They may require a 5 game losing streak to prompt me.

I started making these plots because, unfortunately I don’t get to watch every game. I wanted to be able to see what kind of trends the Caps were creating throughout the season. I also don’t like Excel plots. I don’t know why, could be because of the limitations, or because I’m a MATLAB elitist. Additionally, I’ve never played hockey and only started watching diligently within the past few years. So I don’t usually have anything meaningful to add on what a player is supposed to do or how difficult a certain task is. Nor do I have a vast knowledge of the players of the past.

About the plots – they are auto generated using some custom MATLAB code I wrote. I started last year by manually inputting data from a number of sites – mainly ExtraSkater. But when that got shut down I had to write something that could import and reduce the data automatically from what NHL releases. So this is young code, and it has not been peer reviewed. If there’s anyone out there proficient in MATLAB and who knows the formulas used to define some of the advanced statistics, I’d gladly welcome your time to look over the stuff so that I’m not publishing incorrect garbage.

Well that about sums it up, I hope you enjoy the content. Feel free to comment or provide feedback.