With the Stanley Cup playoffs right around the corner, odds-makers and analysts are rushing to pump out their prediction models. For analysts, the hope is to simply be correct, while for odds-makers, the idea is to profit off of fans’ decision making.
In either case, we get a sense of which teams are favored to have post-season success. Keep in mind, predictive models are probabilistic in nature, so they only tell us which teams are the most likely to win based on certain inputs, not which ones will win.
Here’s what some of the more popular models floating out there have to say in terms of the Anaheim Ducks:
DTM About Heart
“Don’t Tell Me About Heart”, a prominent analytics blogger, has released a model using expected goals to predict playoff outcomes.
For those unfamiliar with the expected goals metric, it essentially tells us how many goals a team or player is expected to score based on a shot’s location and type. However, in DTM’s model, he also factors in individual shooting talent. Here’s what his model says:
xG Playoff Probabilities pic.twitter.com/WjeoQonfXN— DTM About Heart (@DTMAboutHeart) April 10, 2017
Anaheim comes out looking pretty good here, with a 66.2 percent chance to move on from round one. From there on out, there chances are still solid in relation to their Western Conference peers.
This all makes intuitive sense —- Anaheim hasn’t fared well in raw shot attempt differential, but they’ve done a tremendous job of controlling scoring chances all season long. Scoring chances are naturally higher quality shots, so it would make sense that the Ducks look better in a model that favors shot quality.
Moneypuck.com’s prediction model uses a slew of variables to come up with their probabilities. The variables used are: expected goals from non-rebound shots, expected goals in special teams situations, unblocked shot attempt share, save percentage, even strength shooting percentage, and share of power play time.
There’s a bit more nuts and bolts to it, explained here. Anyway, here’s how Anaheim looks:
The Ducks are still favorites in the first round, but by a much slimmer margin than the one we saw in DTM’s model. Again, this makes intuitive sense — Anaheim can’t possibly look as good statistically if shot attempts are factored in.
Even so, having a 6.4 percent chance to win the Cup in a model where the favorite has a 14 percent chance doesn’t sound all that bad.
Micah Blake McCurdy, also known as “Ineffective Math”, who runs the fantastic hockeyviz.com, also released his own predictive model. Unlike the previous two models, this one doesn’t factor in expected goals, but instead, “offence and defence, shooting and goaltending results, special teams, penalties drawn and taken, rest, and home ice advantage”. Here’s the results:
"Official" thing will go on the site for posterity/shame tomorrow, but today I did the important work: I fixed the colours. pic.twitter.com/r4MpDhXbUY— Micah Blake McCurdy (@IneffectiveMath) April 11, 2017
Once again, we find Anaheim as a favorite to move on in round one. It’s interesting that without expected goals factored in, the Ducks come out looking pretty good with the other inputs that McCurdy includes.