Nate Silver’s model vs. InTrade’s prediction markets

Posted on February 23, 2009 by Chris F. Masse

The information is below.

However, before that, just a short note.

-

  1. The real comparison to do would be to compare the prediction markets with a panel of very diverse Oscar predictors.
  2. The real question to ask is, “Was forecasting each Oscar category an easy task or a difficult task?”. If it is considered easy, then the prediction markets have no merit.
  3. The ultimate question is, “Do the prediction markets help us approaching omniscience?”. And the honest answer would be: epsilon.

-

Nate Silver:

intrade-versus-nate-silver

-

Wikipedia:

(In black, what either Nate Silver and/or InTrade missed.)

Best Picture = Slumdog Millionaire

Best Director = Danny Boyle

Best Actor in a Leading Role = Sean Penn

Best Actress in a Leading Role = Kate Winslet

Best Actor in a Supporting Role = Heath Ledger

Best Actress in a Supporting Role = Penélope Cruz

(Nate Silver missed 2, whereas InTrade missed only 1.)

-

UPDATE: Nate Silver’s confused post-mortem

About Chris F. Masse

Founder and President of Midas Oracle
This entry was posted in Analysis (Accuracy & Precision), Analysis (Meta), Exchanges & Markets, Forecasting (Science & Practice), Market Expiry, Market Prices & Probabilities and tagged , , , , , , . Bookmark the permalink.

3 Responses to Nate Silver’s model vs. InTrade’s prediction markets

  2. Chris F. Masse says:
    February 23, 2009 at 4:43 PM

    Guessing all the frontrunners correctly is something to brag ONLY if the reported confidences are high enough. If they are not and you get them all correctly, then the markets have biases and are NOT accurate.

    http://behind-the-enemy-lines.blogspot.com/2009/02/why-failing-to-fail-is-failure.html

  3. Jason Ruspini says:
    February 23, 2009 at 7:03 PM

    As we saw in both the election predictions by State and the Oscars, Nate Silver tends to have less error in favorite/longshot situations while PMs did better where the predictions were roughly between 30-70%.

    Here, Nate’s error was so great in the middle band that the favorite/longshot bias of the markets did not matter.

    The biggest danger for Nate’s modeling is one of those “0%” longshots coming in though. He could have done much worse last night. Because of the longshot issue, when comparing the error of Nate and PMs, PMs will tend to bleed while Nate will be more vulnerable to blow-ups.

Leave a Reply