Getting better…

Well, here we are again. Last night was a big night for some, and not so big for others. RL and SL scored the most points, and the forecast only really blew it on three participants (over 100% error for JG, AT, and JW, all who under-performed). Here’s a summary:

20161026

On the whole, the average error was 2.7 (which is worse than last week’s), and the maximum error is 7.7 (also worse than last week). That said, the three anomalies are removed (slackers!), the average difference is 2.3, and that max error is 5.5 – so, not so bad after all! Now, let’s take a look at why these errors happened on a player by player basis.

Last night’s over-performers were:

  • Kucherov (4 points total, over f/c by 3.4 points)
  • Stamkos (4 points total, over f/c by 3 points)
  • Hall (3 points total, over by f/c 2.4 points)

Typically, those that outperformed their forecast had one (or more) of these players on their roster.

Likewise, the top under-performers were:

  • Karlsson (no points, 1.8 under f/c)
  • Marchand (no points, 1.8 under f/c)
  • Tuuka Rask (no points, 1.6 under f/c)

Given that a majority have Karlsson on their roster, this helps in explaining the general under-performance.

I’m hoping to start building longer-term forecasts that I can spread the error over, so stay tuned!

 

Leave a comment