clock menu more-arrow no yes

Filed under:

Reading tea leaves: creating stories out of random noise

New, comments

Let's try something new.

Instead of using statistics to dispute narratives after they have been written, let's see if we can use statistics to predict which narratives will soon be written. After the jump, you'll find a list of articles that may appear sometime this year as people try to assign significance to simple regression.

  • Andrej Meszaros was great for the Flyers last year, but has not found the same success this year. Statistic subject to variance: +/-. Last year's +30 was driven by the team having a 10.2% shooting percentage and 93.3% save percentage when he was on the ice at 5 on 5. Expect numbers closer to league-average 8.1% for 5v5 shooting and Bryzgalov's career 92.6% for 5v5 save percentage this year.
  • Sean Couturier started the season well, but is hitting the rookie wall. Statistic subject to variance: shooting percentage. Couturier is at 20.8% now, after averaging 14.2% in Juniors and 11.7% last year.
  • The 9-8 lost-in-the-woods game was a wake-up call for Bryzgalov that turned his season around. Statistic subject to variance: save percentage. Bryzgalov was at 87.0% after that game, yet has a career 91.5% average.
  • Brayden Schenn is finally healthy and/or fitting in. Statistics subject to variance: +/-, shooting percentage. The team is shooting 0.0% and stopping 82.1% of shots at 5 on 5 with him on the ice, which resulted in him being a -5 through four games but won't last. His individual 0.0% shooting percentage won't hold up either.
  • Ever since (injury, change of lines, roster move, other random event), the Flyers offense isn't clicking like it was early in the season. Statistic subject to variance: shooting percentage. The Flyers are currently shooting 11.3% at 5 on 5, versus a league-average of 8.1%, and 17.1% at 5 on 4, versus a league average of 13.4%.

Which of these articles do you expect to see by the end of December? What other storylines do you see emerging?