If you’ve read my work on other websites, you’ll know that I love statistics. All kinds, really…simple stats,
advanced stats, the whole ball of wax. With a lockout getting closer and closer and not much off-season
news to cover, I thought I would do a series of posts on different kinds of statistics. Today’s piece
focuses on how to measure luck.
We now have the ability to track exactly who was on the ice for each shot taken during a game, and
therefore we can measure what percent of shots turned into goals when a player is on the ice, whether
it was his own shot, a teammate’s shot, or an opponent’s shot. This technology has allowed us to
quantify a whole bunch of other things too, but that’s another article for another day.
So we know what percent of shots became goals when a player is on the ice (even strength only), we call that on-ice
shooting percentage. We know what percent of shots against turned into goals too, we call that on-ice
save percentage. We also know that if you add those two numbers, the league average comes out to
about 1,000. For example, Cal Clutterbuck’s On-ice Sh% last year was 6.9%, and his on-ice Sv% was .930,
so his PDO was (.069 + .930 = .999) or right about average. Ditto Nick Johnson, 7.4 on-ice Sh% and .926
On-ice Sv%, (.074 + .926 = 1.000)
PDO is poorly named—it’s not an acronym for anything, it gets its name from the internet handle of
the guy that first came up with the idea. However, it has been proven to be a very useful tool because
players reliably regress toward the mean: someone with a high rating comes back to earth the next year
and someone with a low rating tends to see a bump the following year.
Now let’s take a look at some Minnesota Wild players and how their Puck Luck shaped up last year. For reference,
the team’s collective even-strength Sv% in 11-12 was .927. An on-ice Sh% of 10 is pretty standard for
most top-level talent, with 12 or more being superstar and 5-7 expected for grinders.
Fun with Small Sample Sizes
The following players showed some outlier numbers, but each played five games or less, and anything can happen in a small sample. There isn’t really anything to be learned from the following cases, but to me it’s always fun to look at the extreme cases.
Jeff Taffe (5 GP) On-ice Sh% 19.05, On-ice Sv% .976, PDO 1166. This guy should go out and get himself some lottery tickets. He was on the ice for 4 GF and 2 GA, and you know there must not have been a lot of rubber flying at either net, but a couple Powerball tickets might not hurt. Just saying.
Chay Genoway (1 GP) On-ice Sh% 14.29, On-ice Sv% .800, PDO 943. It was just one game, with 1 GF and 1 GA, but still, that’s one really big number and one really, really low number.
Kris Fredheim (3 GP) On-ice Sh% 0.0, On-ice Sv% .882, PDO 882. While Taffe got all the breaks, this poor guy was not on the ice for a single goal for, and two goals against. It wasn’t just a one-game stint either, it was a week’s worth of ice time. Better luck next year Kris.
Mikko Koivu (55 GP) On-ice Sh% 8.53, On-ice Sv% .935, PDO 1020. Based on what we know about these types of stats, it’s a pretty safe bet that next season, his on-ice shooting percent will be a little higher, and the goalie will probably save a little less than .935, but his final number will settle a little closer to 1000.
Jared Spurgeon (70 GP) On-ice Sh% 7.08, On-ice Sv% .942, PDO 1013. For a guy that played seventy games, it’s rather remarkable that Spurgeon’s on-ice sv% was .942. Hopefully with the better talent the Wild will be fielding and some more development, his on-ice sh% will creep up a bit.
Dany Heatley (82 GP) On-ice Sh% 8.31, On-ice Sv% .928, PDO 1011. Heatley’s days of scoring 40 goals (heck, even 30) are probably behind him so his on-ice sh% is probably about what we can expect going forward, although skating with Zach Parise and a full year of Koivu may help.
Matt Cullen (73 GP) On-ice Sh% 7.60, On-ice Sv% .906, PDO 982.
Devin Setoguchi (69 GP) On-ice Sh% 7.30, On-ice Sv% .904, PDO 977.
Both these players posted an on-ice Sh% in the seven range, so the hope is that will come up a bit next year, but it’s interesting to note that for two guys who played in seventy games, the goalie behind them was only saving 90% of the shots faced. As mentioned, PDO tends to regress to the mean, so with on-ice Sv% closer to team-average next year, these two players are in line for some better performance next year.