Over the last few days, I've looked at Jimmy Howard to see how last season brought him down in terms of Quality Starts and at his consistency relative to other goaltenders. Today we're going to dig a bit deeper into Howard's starts to see what we can learn about the different levels of performances and how often one might lead to another.
For a refresher, we're dealing with only three types of goalie performances here:
- Quality Start (QS): A performance where the goaltender put up a .917 or better save percentage or, if the opponent failed to get more than 20 total shots, a game where he stopped 88.5% or better.
- Really Bad Start (RBS): A performance where the goaltender failed to attain even an .850 save percentage.
- Non-Extreme Start (NES): Any performance which qualifies as neither a Quality Start nor a Really Bad Start.
The definitions for the first two of those come from Hockey Abstract 2014. The third is a term I made up so I could remember the Nintendo Entertainment System.
To do this analysis, I looked at every regular season game Jimmy Howard has played since the beginning of the 2009-10 season. That was the first season of his as a starter, so I'm throwing away 9 games spread over three seasons prior to that. This data set includes 276 total games. I have identified three of those games where he wasn't the starter, but for the purpose of overall numbers, I have kept them in as though they are starts, only skipping over them when it comes time to look at games grouped before and after certain starts. I didn't feel it was necessary to cut 120 minutes worth of gameplay out of this analysis to obey the pure definition of a start, especially since Howard was actually able to win two of those.
Looking at 5 years instead of 3 also sheds a bit more light on Howard's career QS total. Thanks to a bad 2010-11 regular season, his QS% here is 54%. However, that year's Wings' squad was also an offensive powerhouse, which gave Howard a lot of leads to play with. Including all the data lowers his RBS% down to 13.8%.
Why are Quality Starts Important?
A goalie has to give his team a chance to win. While a lot of that rides on the defense's ability to help the goalie out by keeping shots to the outside, limiting overall number of shots, and helping to clear rebounds, when push comes to shove, the glory and the blame falls on the shoulders (and other body parts) of our James Tiberius Howard. When Howard is able to do his job and put in the work to earn a Quality Start, the Wings have a much better shot of winning.
How does 75% strike you?
That's the percentage of Jimmy Howard's quality starts over the last five seasons which have turned into victories for the Wings. If you add in the unfortunate overtime/shootout losses over that stretch, Howard's quality starts have earned 81.5% of the points available in the 149 of those games he's given them.
So I'm Guessing that Avoiding Really Bad Starts is Also Important?
Ah, Specific Story-Guiding Question-Asking Disembodied Strawman, how I've missed your leading questions. That was an excellent one. Rob Vollman stated that a RBS gave a team only a 10% chance at winning the game. Howard falls very well into this. Thanks to the luck of both shootout coin-flips going against them, Detroit won only 3 out of 38 total games, or 7.9% (winning one shootout jumps that just over the 10% threshold). Points-wise, the Wings were able to earn only 19.7% of those possible.
How About Non-Extreme Starts?
When you factor up the 89 games where Jimmy Howard put up a save percentage between .850 and .917 (leaving out the three Quality Starts where he was below .917, but also saw 20 or fewer shots), the Wings earned wins 41.6% of the time and just 47.2% of possible points.
Jimmy Howard 2009-10 through 2013-14 Starts
Amazingly, it's hard for a goalie to get a shutout in a game that's not a quality start.
Surrounding Starts: Follow-Ups
So what does it look like after Howard has a Quality Start or a Really Bad Start? Is he more likely to fall off right after being good? How well does he rebound from cocking it up out there? For this, I looked at his next performance after each of those types of starts and also his next three.
|1 After RBS||3 After RBS||1 After QS||3 After QS|
Well that's interesting, if not surprising. Howard is better after sucking and he sucks more after being better. As a reminder, the QS and RBS total percentages over this time frame are 54% and 13.8%. Howard is more likely to rebound positively from a bad start and much less likely to repeat one. This effect tends to carry forward for several games as well (although the odds of avoiding a follow-up RBS go up quite a bit, they still hang out below his career average).
On the other side, it's troubling how much more likely that Howard is to follow a confidence-boosting win with a craptastic outing. It's slightly more than that even, as you can see that his save percentage over the three games following a good one tends to hover below his career average (although almost 90% of his career shutouts come from that set of games, making up 44% of those total Quality Starts).
Surrounding Starts: Prequels
Can we see a bad start coming based on the previous start?
|1 Before RBS||1 Before QS|
Holy smokes; no wonder it's so hard to build confidence. A full one-third of Howard's career shutouts come directly before a Really Bad Start. This matches up with the chart above about how how many more RBSs there are after QSs as well. But hey, at least he bounces back well, right?
How Hard Are These Starts?
It would be nice if we could casually say that Howard gets cocky after a dominant start and lets off the gas or that he gets fired up by a loss and bears down harder on the next go. It's entirely possible that this is what's happening, but there's also five other assholes on the ice in front of him who have a say in how easy it is for him to look like a hero or a chump.
Unfortunately, I don't have location data for all 7,786 shots over this time span, so I can't tell for certain how much harder it is, but we do have a decent proxy in number of shots. Here's how the shots against Howard per 60 minutes works out for each game type.
|Total||1 After RBS||1 Before RBS||3 After RBS||During RBS||1 After QS||1 Before QS||3 After QS||During QS||Non-Extreme|
There's not a ton of difference between the amount of shots the Wings allow in front of Howard under these circumstances, so coming to conclusions is much harder. The biggest differences from average come 1 start before and 1 start after an RBS. These are also the smallest sample sizes by a fair margin. What does seem to fit is that if Howard is falling victim to the "gets cocky/bears down" narrative, it seems the team is generally doing the same. The smallest rate of shots in a game for Howard comes immediately after he stinks. The biggest shot rate pops up in the warning games right before a fail.
This piece warrants a bit more study because not all shots are equal and that's not just about distance. I'm going to look more into this, but factoring only 5-on-5 close numbers from stats.hockeyanalysis.com, the defense has averaged out to 29.09 SA/60 overall during this time, which isn't too far off from the overall figure.
How to get the best shot at a great start
If you're curious about the best way so far to get Jimmy Howard a quality start while cutting the chances of a RBS in half while bumping up his career save percentage by 8%, then the simple answer is to always make sure Jimmy Howard plays both games in back-to-back nights.
There you go. Tougher scheduling, please.