Tuesday, April 17, 2012

Odds And Ends On First Round Series

I had this long article planned about how the head to head matchup went in the regular season isn't correlated with the first round playoff record. That's true - it isn't. However, it sure does look like it matters, to some degree, since no one cares whether their team went 4 or 7 games if they ended up losing - a series loss is a series loss.

I sorted the 96 post-lockout playoff teams into first-round series winners and losers. I don't think it's particularly significant, but the first-round series winners had a .583 regular season winning percentage (OT/Shootout results removed and called 'ties') against first-round losers. They had a .569 winning percentage against everyone else. The team with the better head to head record in the regular season was 23-10 in first round playoff series.

Goal differential appears to be huge - the team with the better OT/shootout removed goal differential is 35-13 in first-round playoff series post-lockout.

Meanwhile, better Fenwick Tied doesn't appear to be an enormous advantage, as the better Fenwick Tied team is only 18-13.

Let's put all of this into a helpful table:

Criteria1st R W1st R L
Better H2H Record2310
Better Goal Differential3513
Better Fenwick Tied 1814

None of this is predictive - we don't know that GD is more predictive than Fenwick Tied from these results - but it will be interesting to see how things shake out this year.


  1. I looked at the same thing, and almost posted on it, but decided the utility was so low because there just wasn't enough data to come to any conclusions. I found it will take about 8 more years of post-lockout playoff hockey for any of these stats to be significant enough for predictions.

    BTW- why doesnt every row in your table add to the same number? Maybe not every team had a h2h, but shouldn't GD and Fen Tied have the same number?

  2. I'm sure it took you way less time to do this, too, since you know how to do these things. There's hours of data collection here.

    Why doesn't every row add to the same number -

    H2H: There's 48 H2H results since the lockout, but 15 of them are .500.

    GD: This has 48 results.

    FenTied: There's only 4 years worth of data, so only 32 results, plus there were two teams who were 'tied', rounded to three decimal places. Of course they couldn't possibly be actually tied, but I decided to treat them that way.