By Budd Bailey
It’s never easy to guess how an NHL team will do from year-to-year. Like everyone else in sports journalism, my record on predicting the future is shaky at best.
However, baseball writer Bill James once came up with methods of determining the probability of whether a team will move up or down in the standings. Let’s see how they work in hockey. It only takes three questions, and two of them are easy and require little mathematics (sorry about the other one):
How did the team do last season?
Teams that have winning records and losing records both tend to move toward the .500 mark. Let’s group the teams that had more wins than losses in 2018-19, and identify them as candidates to slip in 2019-2020.
Winners: Tampa Bay, Boston, Toronto, Montreal, Washington, N.Y. Islanders, Pittsburgh, Carolina, Columbus, Nashville, Winnipeg, St. Louis, Dallas, Calgary, San Jose, Vegas,
Losers: Florida, Buffalo, Detroit, Ottawa, Philadelphia, N.Y. Rangers, New Jersey, Colorado, Chicago, Minnesota, Arizona, Vancouver, Anaheim, Edmonton, Los Angeles.
How did the team do in relation to the previous season?
It’s another simple one. If a team improves in points in one season, it is likely to take a step back the following season. Let’s look at which squads had more points in 2018-19 compared to 2017-18:
More points: Tampa Bay, Montreal, Buffalo, Detroit, N.Y. Islanders, Carolina, Columbus , N.Y. Rangers, St. Louis, Dallas, Chicago, Calgary, San Jose, Arizona, Vancouver, Edmonton.
Same points: Pittsburgh.
Fewer points: Boston, Toronto, Florida, Ottawa, Washington, Philadelphia, New Jersey, Nashville, Winnipeg, Colorado, Minnesota, Vegas, Anaheim, Los Angeles.
Was a given team lucky?
Let’s say your favorite team scores 225 goals and gives up 225 goals in a season. It should have a record of .500, more or less, if you count overtime/shootout losses as just plain losses. It doesn’t always work out that way, and luck is a logical reason why. Sometimes a team will win some one-sided games but lose more than its share of one-goal decisions. That usually means the puck just wasn’t bouncing its way.
Some arithmetic is needed to calculate it. The Sabres had a winning percentage of .402 (33-49). James takes the goals-for figure squared, and then divides it by goals-for squared plus goals-against squared. That means the number is 226-squared divided by 226-squared plus 271–squared. That translates to 51076/51076 + 73441. The percentage comes out at .410, which means the Sabres were a little unlucky.
Tampa Bay had a winning percentage of .756 and an expected winning percentage of .682, so the Lightning comes out on the fortunate side. You might expect that from such a great regular season. Let’s look around the league:
Lucky – Tampa Bay, Boston, Montreal, Washington, N.Y. Islanders, Carolina, Columbus, Philadelphia, Nashville, Winnipeg, Dallas, Minnesota, San Jose, Anaheim, Edmonton, Los Angeles.
Unlucky – Toronto, Buffalo, Florida, Detroit, Ottawa, Pittsburgh, N.Y. Rangers, New Jersey, St. Louis, Colorado, Chicago, Calgary, Vegas, Arizona, Vancouver.
We have 16 lucky teams and 15 unlucky teams, with Arizona only .001 on the unlucky side. The luck apparently was well distributed. You’d expect the good teams to have a little luck on their side, and they do – so it’s surprising to see Toronto on the list. The biggest deviation was Tampa Bay, but Colorado was second (.527 expected, .463 actual winning percentage).
Who could be better in 2019-20?
Let’s break down the teams on how the indicators work. Three ups predict a good chance at more points, and three downs are a less than good sign.
Three ups – Florida, Ottawa, New Jersey, Colorado.
Two ups, one down – Toronto, Buffalo, Detroit, Philadelphia, N.Y. Rangers, Chicago, Minnesota, Arizona, Vancouver, Anaheim, Los Angeles.
One up, one even, one down – Pittsburgh.
One up, two down – Boston, Washington, Nashville, Winnipeg, St. Louis, Calgary, Edmonton.
Zero up, three downs – Tampa Bay, Montreal, N.Y. Islanders, Carolina, Columbus, Dallas, San Jose, Vegas.
Admittedly, there are plenty of factors that come into play in predicting a season, such as free agents (in both directions), trades, rookies, and so on. But, if you had to pick four teams that would have more points in 2019-20 than they did last season, you could do worse than the Panthers, Senators, Devils and Avalanche. We’ll see how we did in April.
(Follow Budd on Twitter @WDX2BB)
Leave a Reply