March 10, 2013
Shots On Goal
The Rundown, Week 8

by Ryan Schwepfinger

One of the greatest challenges of fantasy hockey is finding undervalued gems—players that exceed expectation, providing a winning return on investment. In essence, the game is all about return on investment. If your roster is littered with players providing you a high return relative to the price you paid for them, you are on the path to success.

One way to identify undervalued players is to look at trends. In fantasy, we do not have to make a long-term investment to a player like a real-life NHL general manager would. Our margin for error is naturally much larger. As soon as we decide we like a player one week, we can cut him the next with no remorse.

One of my favorite tips for success is based on this: utilize a few roster spots dynamically. Constantly chase hot streaks and role increases; do not be afraid to pull the trigger on free agent moves. While many of these speculative acquisitions will be "misses," the goal is to land a few big "hits" along the way—that is, to pick up a player that continues his break out, providing strong production for your team all season long.

How do you identify who to pick up, or even who to trade for? Focus on the sustainability numbers. One of my favorites is PDO, a favorite at this site. The best feature about PDO is its strong predictability; we know that over time, it will even out towards an average of 1000.

I charted PDO on February 5, and did so again today (March 10). In roughly a month, 28 of the 30 teams saw their PDO either progress or regress towards 1000 as we would expect. That is an astoundingly high predictive index for future production. Making that conclusion even stronger is that the two exception teams not complying with the "law of 1000" were ranked very close to that number to begin with. They were not obvious candidates for movement in either direction.

To summarize: on February 5, 14 teams had a PDO above 1000. All but one regressed, seeing their number come down. 16 teams had a PDO below 1000. All but one progressed, improving their number since. The top 12 teams all regressed, and the bottom 11 teams all progressed.

The value of this information is obvious from a fantasy hockey perspective. Many fantasy regular seasons have about a month remaining. Identifying the teams that will break out over the next month could go a long way towards making that charge towards the playoffs.

For reference: the top 10 teams in PDO: Anaheim, Montreal, Chicago, Toronto, Carolina, Tampa Bay, Detroit, Vancouver, Pittsburgh, Boston.

The bottom 10: Calgary, Florida, St. Louis, Philadelphia, Edmonton, New York Islanders, New Jersey, Winnipeg, Columbus, Minnesota.

While we can surely expect the percentages for the top 10 to decrease over the next month, and vice versa for the bottom 10, the strategy goes beyond simply acquiring players on teams with a low PDO, and selling players on a team with a high PDO. It is important to note exactly which "branch" of PDO (save percentage or shooting percentage) leads to a team's ranking, and to factor in possession.

Some quick notes:

Philadelphia and St. Louis have fallen to 27th and 28th in the league in PDO, respectively. Their falls are almost entirely based on save percentage—Philadelphia's 5v5 shooting percentage actually increased over the past month. This may create an opportunity to buy Ilya Bryzgalov and Jaroslav Halak very cheaply, based on recent perception. Remember, these are two strong offensive teams—at the very least, the goaltending should improve enough to get you wins.

While all teams in the top eight have seen regressions, only two have not been "severe" regressions of over 15 points. Those two are Anaheim and Chicago, who regressed from 1047 to 1041 and 1028 to 1025, respectively. Anaheim's 1041 mark today still leads the league, and does so by 15 points. The difference between these two teams is obvious: Chicago ranks fourth in the NHL in close-game Fenwick, while Anaheim ranks 25th. One of my strongest feelings about PDO is relating it to possession. Chicago is more likely to sustain a higher PDO based on their affinity for the puck. I continue to be baffled how Anaheim is winning games based on their percentages. Tampa Bay was an extremely similar case to Anaheim on February 5, and their PDO has since plummeted 59 points from 1077 to 1018.

Anaheim might have been bucking the trends all season, but a similar low-possession team to jump recently is Toronto, whose PDO climbed to fourth in the league, up 29 points from 995 to 1024. Their 5v5 shooting percentage has jumped from 8.4% to 10.1% over the span. It is difficult to see the Leafs maintaining this level of play for much longer. Carolina saw a similar jump from 987 to 1023, but that was based on a shooting percentage leap from an abnormally low 6.6% to 9.9%, not to mention Carolina is a stronger possession team. The differences are obvious: Toronto is a better candidate for regression.

Edmonton continues to rank last in 5v5 shooting percentage (5.7%), followed closely by Ottawa (5.8%) and San Jose (6.3%). Edmonton's number actually has trended up from 5.3% since February 5, but Ottawa's is down from 7.1%, and San Jose's is down by 8.8%. Ottawa has plummeted in possession without Erik Karlsson, but San Jose has held steady, and thus their percentage drop is more a factor of total luck than the other two teams. Interestingly enough, Ottawa and San Jose are the two teams with the highest 5v5 save percentage at the moment (.947 and .938, respectively).

Keep in mind PDO and sustainability when considering our weekly matchup coding. Possession and goals data tells us where teams have been, while PDO can tell us where they are going. Putting the knowledge together can give you the upper hand in a playoff chase.

This week's Hot Matchups: Buffalo, Tampa Bay, Washington (Edmonton, Colorado, New York Islanders, Florida, Phoenix).

Cold Matchups: Boston, Chicago, New York Rangers, Montreal, Detroit (Los Angeles, Vancouver, San Jose).

Hot Offenses: St. Louis, Chicago, Pittsburgh, Montreal (Los Angeles, Boston, Carolina, Philadelphia).

Cold Offenses: Columbus, Buffalo, Edmonton, Nashville, Minnesota (Colorado, Florida, Ottawa).

A great example of combining PDO data with our matchup data—if Carolina is to show any offensive regression, it likely will not occur this week. The Hurricanes have phenomenal matchups this week in the notoriously offensive Southeast. However, the team's eighth-ranked 5v5 save percentage of .924 is a strong candidate to regress in my eyes, especially with Dan Ellis and Justin Peters between the pipes as opposed to Cam Ward.

Much like Cam Atkinson returning with little overall fanfare a few weeks ago, Kris Versteeg is back in the lineup for Florida. It is not often that a top-line, big-minutes winger is available for pickup. It is difficult to recommend him based on anything beyond his role and ice time, but consider him if you have the flexibility and can afford the plus/minus hit.

Eric Fehr had displayed strong peripherals all season, but was not getting the ice time to make him worthy of fantasy consideration. That has changed over the past month or so, as he has been lining up with Nicklas Backstrom and Troy Brouwer. His strong possession game has translated into scoring, with four points in the last week alone. The downside for Fehr is that his on-ice shooting percentage is far outpacing his linemates' marks, suggesting regression, but in deeper leagues, he is a player to take a shot on.

A defenseman comparable to Fehr is Jake Muzzin of the Kings, who also had strong possession numbers but a small role earlier in the season. That seems to be changing as well. The solution to the failed Keaton Ellerby experiment has been slotting Muzzin alongside Drew Doughty at both even strength and on the power play. Consequently, his scoring numbers are starting to balloon, along with his ice time. Muzzin is a name to watch moving forward.  

For the first time all season, Columbus is not coded as a "hot matchup," due to their team defense improving strongly enough to push them to the top half of the league in fewest goals allowed/game. This is the case despite a 5v5 save percentage that has changed minimally over the past month (.902 to .909). Jack Johnson is a name to watch here—his possession issues are well-known, but his relative Corsi is nearly doubled from his time in Columbus last season, and his big-minute, power play quarterback role is incredibly appealing for fantasy. The concern with players like this is always plus/minus, so be wary, despite Johnson's plus-one mark in the three games since he returned from injury.