Lightning Numbers: Adjusting for score effects, the impact of injured blueliners, and new goaltending stats

This will be the first #fancystats report of the season, looking at some of the underlying metrics for the Lightning's first seven games.

This is as good a time as any to take a baseline picture of the Lightning's #fancystats. It's very early to be drawing a lot of conclusions from these numbers, but things have changed drastically for the Bolts over the past week, so a snapshot of where they are right now can be informative after we see where they end up in a month or two.

1. Fenwick (Puck Possession)

Fenwick rates are pretty good measures of the level of team play, although it's quite early in the season for strong conclusions to be drawn. At the most right now we get indications, and what we leave in and out can change the picture to some extent.

At 5v5, the Lightning are in 8th in the league in Fenwick with 54.7%. The problem with using all 5v5 data, though, is that it doesn't account for score effects, which is a fancy way of saying that when teams fall behind they shoot more, and when teams get ahead they shoot a lot less. If we take just situations where the score is close (within one goal in the first or second period), we can eliminate the times when a chasing team starts to throw everything at the net, but we do that by making an already small sample size even smaller.

One way to deal with the score effects issue without decreasing the sample size is to look at Score-Adjusted Fenwick. This essentially compares what a team does in various score situations to what everyone else in the league does in those situations. It has the benefit of leaving all 5v5 play in the sample but it also accounts for score effects.

In Score-Adjusted Fenwick, the Lightning are in 11th with 52.28%. The concern here is that when the score has been tied, which is where the Lightning have spent the most time, they're only possessing the puck 47.25% of the time.

They're good when they're good (up one or two goals) or when they're down by one, but when they're just starting a game or when the other guys catch up, they're not so good. And that's a little worrying. That's when the Lightning ought to be setting the pace. Instead, they're chasing.

2. Injuries on defense

Currently, injured blueliners Victor Hedman and Radko Gudas both have Fenwick percentages above 50%, but so do most Lightning players. Hedman has a relative Fenwick of -2.24% and Gudas has a relative Fenwick of +3.75%, but by looking at With or Without You breakdowns we can dig a little deeper into that.

I pulled the WOWY data from Hockey Analysis for Hedman on Tuesday and for Gudas on Wednesday. These give us insight into how one player is affecting his teammates (and vice versa). It's tough to truly untangle all of the ins and out, and this is a very small sample size, but it does shed a small bit of light on how the lineups are working.

Better with Hedman

Worse with Hedman

Better with Gudas

Worse with Gudas

Steven Stamkos Valtteri Filppula Ryan Callahan no one
Ondrej Palat
Alex Killorn
Tyler Johnson

The truly informative part of this exercise, however, was discovering that Hedman played much better with Anton Stralman than without him, but Stralman experienced no corresponding bump from being paired with Hedman. The same was true of Jason Garrison and Radko Gudas. Gudas played better with Garrison, but Garrison didn't see a great benefit from Gudas.

It's impossible to tell whether that will continue to be the case, but those two offseason pickups appear to be having some effect at this moment.

3. Goaltending:

As of Wednesday morning, the Lightning have a team save percentage of .925. In five games, Ben Bishop has put up 3 quality starts. Both of Evgeni Nabokov's starts have been quality starts. Both have positive Goals Saved Above Average currently, but neither are in the top echelons in that number.

In a testament to the weirdness of small samples, Bishop's even-strength save percentage is much lower than his performance when down a man. He has a .908 at even-strength and a .921 PK save percentage. (For whatever reason, denotes a goalie's save percentage while on the penalty kill as "Power Play save percentage" while "Shorthanded" means the goalie's team is on the power play. It's very confusing, since almost no one else does that.)

Generally speaking, ESSV is higher and the higher rate of scoring on the power play is what brings goalies' overall save percentages down. In this early going, however, the Lightning have been turning up heads on the penalty kill, and that's affecting Bishop's numbers. I'd expect his ESSV to go up even as his PKSV goes down, though it's not really knowable how fast that will happen.

There are some interesting goalie stats available this season that we didn't have last season, as new sites pop up to try to take the place of Extra Skater. One of the fun ones is War-On-Ice's Adjusted Save Percentage. ASP weights save percentages on shots from three zones to give a more nuanced picture of goaltender performance than straight save percentage can give. (See their glossary for an explanation of the three zones.)

As of Wednesday morning, Bishop's ASP (5v5) is .9066, due to low rates in the medium scoring zone (basically the "home plate' area, slightly expanded, minus the lower slot areas). Out of 60 goaltenders, he's 50th in the league in saving shots in the medium scoring zone (25% of the 5v5 shots he's seen). He's 34th of 60 on shots from the high-scoring zone (15% of the 5v5 shots he's faced.) That's probably deeply influenced by randomness, but the deeper style he plays can sometimes be a factor.

As things stand, we don't really have a good grasp on exactly what these numbers are telling us. It's tough right now to know even how common various shot locations are across the league or what kind of difference shot location really ought to make in evaluating goaltender performance.

It's obvious that it ought to be taken into account, but that's definitely not the whole story. For example, it's more likely that a goaltender will have a chance to set up on a perimeter shot, but that's not always the case. In addition, this data tells us nothing about the height/trajectory of the shot or what part of the net it was aimed at, and those things are crucial to understanding what kind of space the goalie covered or didn't cover.

In any event it will be very interesting to see whether we can glean more information about these new save percentage breakdowns and figure out how best to incorporate them into goalie evaluation.

4. Forecast

The Lightning's 5v5 PDO is currently 100.2. Overall, slightly elevated goaltending (.9333) is offsetting slightly low scoring (6.88). I would expect, given this roster, that the team shooting percentage ought to rise to around 8% by the end of the season, although it is possible that won't happen. I judge that the upward pressure on that stat is moderate. Nonetheless, right now, good luck, if you want to call it that, is balancing bad luck.

But if "luck" (really randomness) isn't having a huge effect on current results, special teams might be. Special Teams Efficiency gives an indication of the effect of man-up and man-down situations on a team's record. The Lightning have an STE of 113.9 (26.9 PP% + 87.0 PK%), which is very high. The power play rate in particular is very likely to come down.

The timing of when this decline happens in relation to if and when 5v5 scoring rates increase could lead to some sticky situations. If the PP scoring decreases but 5v5 scoring doesn't, the Lightning will be relying on goaltending to bail them out, and that's always a bit of a gamble. It looks to be an especially big one in the immediate future.

Over the long-term, however, the key will be whether the Lightning can once again figure out how to set the pace in 5v5 tied situations. The past few games have shown a team that has been more than a little sloppy. They may have gotten away with some points, but that's not a recipe for long-term success, especially if Ben Bishop's goaltending isn't otherworldly like it was last season. Fix the fundamentals of 5v5 play and the rest won't matter so much.