Second Thoughts: Assessing the Bubble of the NCAA Bubble
by Ryan Lambert/Columnist
At this point in the season, if you're not in the top 20 of the Pairwise, it usually takes a minor miracle to get you anywhere near the tournament picture.
Invariably, though, someone gets hot, another team gets inexplicably cold, and there's a shakeup or two. That is to say, the current top 20 isn't hard and fast by any means.
See: Northeastern dropping from seventh to 16th and out of the tournament in about two months in 2013-14.
See also: Northeastern going 19-1-2 post-Christmas to make the tournament in 2015-16 after starting 3-12-3. They jumped from 49th to 13th in two and a half months. (What goes on with those guys?)
However, it's fair to say that the current NCAA tournament bubble — let's say teams Nos. 14-20 — are fairly hard and fast, and that any teams in the 20-25 range would have a hard time cracking it barring a run that at least somewhat approximates those 2015-16 Huskies.
With this in mind, we must consider the fact that the two most predictive stats in college hockey as it relates to long-term success are the percentage of shots taken in all situations (SF%) and the sum of shooting and save percentages in all situations (PDO).
Shots-for percentage is a pretty simple metric to understand: The more you outshoot your opponents, the better chance you have of scoring more goals than them. PDO is a little trickier, because in college hockey in particular, teams with a lot of skills or a high-end goalie will be more likely to maintain what would otherwise be considered unsustainably high percentages.
So the only real question one needs to ask themselves at this time is, “How good do I think this team actually is?” If you're just basing your projections on wins and losses, you often find the analysis begging; certainly good teams are going to win a lot more than bad ones, but sometimes bad teams can win a lot of games because they shoot at a higher percentage than is reasonable, or a goalie gets surprisingly hot.
A classic example of this comes in the form of UNH earlier this season; remember when everyone thought the Wildcats were destined to make a big run for Dick Umile's final season? They started the season 6-1-1, the third-best record in the country at that time, and had people writing stories about how good everyone on the team was. This despite the fact that they were controlling fewer than 50 percent of the shots in their games, but shooting almost 13 percent as a team and getting .936 goaltending from a guy who entered the season with a .911 career save percentage.
Since then? UNH is 3-6, clinging to .500 hockey for dear life (in what is probably a doomed effort), still controlling fewer than 50 percent of the shots in their games, but have been shooting a much more predictable 8.6 percent, while getting a much more predictable team save percentage of .904.
That's not to pick on the Wildcats, but it's quite illustrative of the forces we're talking about. Subpar teams can only outrun the effects of mediocre systems and low-level talent for so long before everything falls apart. Likewise, that 2015-16 Northeastern team was a hell of a lot better than their 3-12-3 start suggested, and at the end of the season they ended up with 22 wins, which felt about right (not that any team should win 19 of 22 in most circumstances).
So then, which teams in the 14-25 range (and beyond) are most likely to see their NCAA tournament chances change over the second half of the season? Well, using SF% and PDO will probably give you a good picture:
Generally speaking, the teams in the top left are those that have played well but perhaps not gotten the best luck, while those in the bottom right are mostly playing poorly overall and getting some bounces to go their way. A few interesting teams stand out on both sides of the ledger, but four or five the strike me as most likely to push their way into the NCAA tournament:
The Crimson, which technically isn't close to the bubble right now (36th in the Pairwise at 4-5-1). The Crimson are playing well overall controlling more than 51 percent of shots — a respectable but not overwhelming number — and not getting results despite a PDO of nearly 102, which is high but sustainable for a team with that much skill and a good goaltender.
One has to imagine that the small number of games they've played coupled with the fact that they're about to play primarily a weak ECAC schedule going forward is going to get them a lot of wins in short order here. They have the personnel and the coaching, but their games have mostly been blowouts one way or the other, which explains the PDO to some extent; their wins have come by a combined score of 19-4, their losses 8-16, with a 4-4 tie mixed in.
They've dug themselves a bit of a hole, but not an impossible one.
BU and Wisconsin
The Terriers and Badgers control the majority of the shots on goal in their games, and BU is having a little bit of trouble putting the puck in the net (9.2 percent shooting is fine but not for a group with that talent level).
For these teams, the real issue is that Kyle Hayton and Jake Oettinger have been awful for much of the year, especially relative to his quality.
BU's league schedule is pretty easy in the second half of the season, and one imagines they'll figure things out despite being two games below .500 at the break. Wisconsin gets the (potential) benefit of the Big Ten seeming like a huge crapshoot in terms of who beats who from one game to the next.
This is a pretty simple team to break down: They're not nearly this bad, just logically.
You look at the River Hawks as having struggled a lot in terms of shooting the puck and also getting a bit lucky in terms of their PDO, but that's because they're maintaining a high shooting percentage (11.2 now 12.6 last season) but getting much worse goaltending (.898 versus .911).
But what's interesting about Lowell here is how bad that team save percentage is, and more accurately, why it's that bad. After all, Christoffer Hernberg is having himself a very good season at .931 for a 9-2-0, while Tyler Wall has been so far beyond bad as to be inexplicable at .823 and 0-6-0. Let's put it this way: Wall has played about a quarter of Lowell's minutes and given up more than half of its 49 goals; his highest save percentage in any single game this season is .875.
Let's assume, then, that Hernberg's save percentage is unsustainable at its current level and drops to even the national average. Meanwhile, by dint of Wall's putrid performance, he doesn't start another game (giving up 6 on 33 against BU should do it for his season).
Wall has cost Lowell about 12 goals versus the national average in just six starts, three of which he's been pulled from. That's equivalent to roughly two wins. If Lowell entered the break 10-7-0, no one is having a conversation about why they're bad.
This one's a bit of a long shot but a reason the Bobcats are so awful in terms of record (currently four games below .500 and 47th in the Pairwise) but so good in terms of how well they play. They take more than 58 percent of all the shots in their games.
But because they suffer from a low shooting percentage (7.2 and likely to come up) and poor goaltending (.885 and who knows?). Here, too, there's a team suffering from a serious goaltending discrepancy; Andrew Shortridge has been worse-than-average at .903, while Keith Petruzzelli has been horrific at .860. If Shortridge is The Guy going forward — and why wouldn't he be? — this team probably starts winning a ton of games.
Enough to overcome a 32-spot PWR deficit? Well, they might be this year's Northeastern?
So if we're saying those four teams theoretically could do it, which teams could drop out to give them a chance? The three most obvious candidates to drop out are:
Minnesota-Duluth and Nebraska-Omaha
Both of these teams are currently below .500, which makes them ineligible for the tournament in the first place, and getting awful goaltending. They're propped up entirely by the strength of their NCHC schedule, which isn't going to change as more conference matchups approach, but they're also pretty likely to rack up more losses since most of the NCHC teams are, y'know, better than them.
Either one of these teams could get their goaltending sorted out, sure, but being sub-.500 and in a top-16 spot is tough to square in the long term.
As I said in this column, the Eagles are looking at their second straight season of not really being able to beat good teams but regularly and absolutely humiliating those worse than them. In furtherance of the point, since that column published a month ago, the Eagles went 1-2-0 against Northeastern (one meeting) and BU (two). They got outshot by 11 (for a SF% of 47.1) and were outscored by three (for a GF% of 43.4).
In the second half, BC plays Providence, UNH, and UConn once each, UMass twice, and both Lowell and Maine thrice in Hockey East games. And in the Beanpot, it's Northeastern then one of BU or Harvard.
Tough to see a lot of guaranteed wins in there, right? But they also have Michigan Tech and one of Arizona State or Northern Michigan, so hey, that's probably two more wins, right?
That might not be enough to hold off other contenders, but it's something.