Archive for Analysis
It took seven grueling games, but the when the Capitals defeated up the upstart Islanders on Monday night, D.C.’s finest secured a date with the Rangers for the fifth time in seven years. This brings us to our second round goaltending preview of Braden Holtby.
When I sit down to write these posts, I always take a look back to see if I’ve done previews before and see how my current analysis stacks up against my observations from prior seasons. I realized this time around that I’ve already done two (!) previews of the twenty five year-old Saskatchewan native. Those prior looks can be found here and here. Seems like only yesterday he was making his playoff debut.
After flashing serious potential in his first few seasons on a defensively porous Capitals team, Holtby put everything together this season under Barry Trotz. His numbers (2.44 GAA and .923 save percentage) were both single season bests. Additionally, he started 72 games this year, by far a career high (we’ll get to that later). He really cemented himself as an upper-echelon goaltender this season. Read More→
As the season has progressed, two things became very apparent for this year’s New York Rangers: They are not a great possession club, but they are still a very good hockey team. A lot of stats folks are already comparing them to last year’s Colorado Avalanche, who rode an absurdly high SV% from Semyon Varlamov to the playoffs, only to be ousted in the first round. There have been other comparisons to teams that missed the playoffs all together (last year’s Toronto Maple Leafs, the Minnesota Wild of w few years ago, etc).
The facts are pretty simple: The Rangers, when looking at their full season numbers, are not a great possession team. It’s something that’s been beaten to death, and a lot of folks have the Rangers bowing out in the first round. The combination of minimal possession and a high “luck” factor (SPSV%/PDO) have made the Rangers a target for early elimination, the same way the opposite scenario made them a dark horse to make a Cup run last year.
But that’s on the surface. The common “anti-stats” argument is that stats can be bent to prove any argument. While that is true, you don’t see those that apply stats properly doing this. The lazy narrative for the Rangers is taking the full season numbers and marking them as “ripe for the picking.” More analysis needs to be performed.
The New York Rangers are in the playoffs. But Ryan Lambert of Puck Daddy raised a question that struck a nerve with a lot of fans: Should the Rangers scare anyone? The article uses raw SAT%/CF% and SPSV%/PDO to justify that the Rangers are a poor possession team that are riding a great season of luck. These were some concerns I raised two months ago. I also revisited it one month ago, and found some interesting trends.
But here’s the issue with using just those numbers: It doesn’t take into account certain variables (like systems and rush shots, more to come), it doesn’t break down trends, and it looks at a full season instead of how a team is playing now. Not all numbers can be taken in a vacuum, some need context. This is one of those times where context is needed.
First things first: Lambert is correct when he says the Rangers aren’t the best puck possession team and that they have an unnaturally high SPSV%/PDO. But that’s about the only thing he’s 100% right with. I’ve tackled New York’s possession in the past, but we can revisit here.
Earlier this season, I wrote a post about how the bubble may burst for the New York Rangers. While an unpopular opinion, it wasn’t without statistical proof. Their possession numbers –a huge correlation exists between possession, playoffs, and Stanley Cup winners– were way down, they had an unsustainable SH%. That combination can spell death for a team.
However the Rangers are winning, somehow, and it forced me to revisit those numbers a few months later. The SH% was still abnormally high, but the possession numbers were starting to improve. At that point, I was cautiously optimistic, but not as optimistic as I was last season, when I correctly predicted a run to the Finals.
Now, after the trade deadline, I find myself wondering: Why not this Rangers team?
By any measure, Cam Talbot’s temporary reign as New York’s No. 1 goaltender has been a resounding success.
The Prince has been an adventure at times, mishandling the puck and giving up the occasional softie. But overall, it’s impossible to complain about Talbot’s sparking 11-2-3 record, 2.35 GAA and .920 SV% since taking over for Henrik Lundqvist (not including yesterday’s game). In fact, as MSG noted on Sunday night, Talbot’s numbers over the last two seasons stack up against any goalie in the league.
The Blueshirts are still counting down the days until Lundqvist’s return, but in the meantime Talbot has quashed any concerns about needing to bring in outside veteran help.
Talbot skeptics will wonder if the Rangers are playing so much better in front of Talbot as a team that they’re masking any of his deficiencies, and it’s a question I’ve asked myself. Indeed over the last few years, it has seemed like the Blueshirts tighten up defensively in front of their backup netminders and feel a little freer to take chances knowing the ultimate safety valve was sitting on his throne in the crease.
Earlier this month, I wrote a post about the New York Rangers and how their underlying numbers weren’t those of a team that can win a Stanley Cup. While that post may have been a bit doomsday/worst case scenario, the numbers didn’t lie. It’s a position that I still take: I don’t believe this team has been playing a dominant puck possession game, like they did last season. Of course, they haven’t been dressing their best lineup, but that’s a horse I’ve beaten to death.
Let’s revisit those possession numbers, using the last two months of games as an indicator of how they’ve been playing. Now this is where things get interesting, as the Rangers have actually played to a 52.9% SAT%/CF% over the past 23 games, since January 1, 2015. All this while playing a possession vacuum on the fourth line regularly (ok, sorry).
Let me be blunt: The New York Rangers did not deserve to win last night’s game over Florida. They started slow, special teams bailed them out, and then they finally played well in the third. This is the exact same way the Rangers played against Carolina, minus the special teams. Hank bailed them out in the first period of that game. Those are two wins with poor performances that they only won because they played against bad teams.
The Rangers lost the two previous games to the Habs and Isles, also games that they did not deserve to win. They beat the Senators before that, and the Rangers certainly played better defensively in that game. However, there was something missing from that game too. That’s five straight games where the Rangers were not the better team. They came out of this 3-2, beating the bad teams and losing to the good teams.
Anyone that has seen Rick Nash at any great length this season knows he’s playing at a Hart Trophy level. Despite his play however, he’s ‘only’ 16th in total points. Everyone knows the major award decision makers usually –heavily– favour the point producers; it’s how Erik Karlsson has a Norris trophy in his cabinet after all. With that in mind, what does Rick Nash have to do to be in the running for the game’s most prestigious individual prize? In short, he needs to get even better.
If Nash merely matches his production from the first half of the season he’ll likely set career highs in goals and points (currently has career highs of 41 goals and 79 points). Despite playing a strong two way game all year and having been remarkably consistent, Nash will need to smash his previous offensive totals. Take a look at the past four Hart winners.
As you know, we focus a lot around using both qualitative (eye test) and quantitative (#fancystats) analysis around here. Both have their pros and cons, and both should be used when evaluating talent and performance. These are not mutually exclusive measures.
One of the major drawbacks of #fancystats –specifically Corsi/Fenwick– is that it doesn’t measure the quality of the shot. While shot quality has been proven to be an unrepeatable skill, meaning we can’t quantify it at a statistical level, it doesn’t take away its importance. Luckily, the folks over at the indispensable war-on-ice have come to our aid with their scoring chance stat.
As we continue to digest last night’s win over the Kings (is winning getting boring, yet?), I thought we could go off-topic on this snowy Friday afternoon. For our regular readers, you know that I have been clamoring for a more comprehensive goaltending statistic than our standard rate stats; GAA and save percentage. As the world of Corsi and Fenwick and the like continue to evolve, goaltending statistics remain woefully underdeveloped. Enter former Rangers goaltender Steve Valiquette, who in addition to Sportsnet’s Chris Boyle (Shot Quality Project), have begun comprehensive research into how shot quality effects goaltender performance. (s/t to Kevin Power)
Every time a new goalie metric is conceptually introduced, I am forced to feel like an ungrateful cynic. Smart, hardworking, dedicated people are attempting to give me what I’ve asked for. Maybe it’s the delivery; “Steve Valiquette is going to change the way we think about goaltending!”, “This new statistic is going to revolutionize goaltending!”. No, it’s not. But it’s a great start.
I don’t want to get into the mechanics too much, but here is pretty much how Valiquette’s theory works:
- The zone is divided in half, vertically from the center bar of the net to the top of the circles, and is bisected by a horizontal line going laterally across the top of the circles.
- The centerline is Valiquette’s “Royal Road”. This line represents the lateral marker a puck travels across within that zone, which leads to higher percentage shots.
- The types of shots are broken down into “Green” and “Red” shots.
- Green shots include
- Possession across “Royal Road”
- Passes across “Royal Road”
- One-timers on same side of “Royal Road”
- Broken plays
- Green Rebounds
- Red shots include
- Shots from outside the designated area
- Red rebounds
- Green shots have accounted for roughly 76% of NHL goals this season, and are obviously converted at a much higher rate than Red shots (24%).
Ok, with that out of the way, here is my take. I think it’s fantastic the work Valiquette is putting in. I hope that is leads to a wealth of new information about how goaltending is evaluated. Once you get past all the terminology, his theory is pretty simple: shots that come what we consider “dangerous” areas of the ice are converted at a higher rate, more so if the goaltender is forced to move laterally. There is also a much higher chance of rebounds, deflections and scrambles resulting in goals than shots from the perimeter.
This all seems pretty obvious, no? I don’t mean to sound overly negative, but it seems to dress up a lot of concepts we simply take for granted, even if they aren’t currently quantified. Don’t get me wrong, I think this tracking concept has a ton of value as a foundation to a more comprehensive, value-based statistic (any statisticians out there, hit me up. I’ve got ideas).
I think that the greatest value this methodology has is situational analysis of current form. Craig Custance over at ESPN spoke to Valiquette about it (Insider post), and applied the method to Dallas goaltender Kari Lehtonen to evaluate whether his poor (relatively) rate stats this year are a true dip in form or whether he is getting hung out to dry. This is a really effective use of this statistic, but there is no basis for comparison that you can really hang your hat on to assess value.
Either way, this is a big step in the right direction for the continued evolution of advanced statistics. I believe a foundational concept now exists to build on, from a quantitative standpoint, and that is incredibly impressive in itself. But, as I said with GSAA, it’s a great start, and an evolution, but certainly not a revolution.