Why you should *never* trust David Pennock when he brags about the accuracy of his predictions

Posted on March 14, 2010 by Chris F. Masse

He cherry-picks positive outcomes in hindsight, and he measures probabilistic predictions categorically and in isolation.

Don’t be fooled by any “research scientist”. :-D

About Chris F. Masse

Founder and President of Midas Oracle
This entry was posted in Analysis (Accuracy & Precision), Economics, Humor, Prediction Post-Mortem, Science and tagged , , , , . Bookmark the permalink.

5 Responses to Why you should *never* trust David Pennock when he brags about the accuracy of his predictions

  1. Paul Hewitt says:
    March 14, 2010 at 10:15 AM

    David Pennock was really bragging about simple statistical prediction models and web search data. Statistical (base line) models outperform web search models, and combining the two leads to better predictions. Neither is a prediction market, though the web search model would be classified as “collective forecasting”. I see you added the “Humor” tag to this post. You must agree that David’s post was at least partly in jest.

    Of all the researchers, David Pennock is not one I would doubt too much when it comes to discussing accuracy. I’m about to publish a post about “Prediction Without Markets” where he and the other authors conclude that prediction markets offer insignificant improvements in forecasting accuracy. He is one of the few that seem to understand the concept of materiality.

  2. Chris F. Masse says:
    March 14, 2010 at 1:22 PM

    Looking forward to your post.

  4. David Pennock says:
    March 14, 2010 at 2:57 PM

    Paul: thanks for the kind words and strong defense. Thanks for you interest in our work and for the fantastic post about it. I will comment further on your post on your site. I agree that Chris’s post matched jest with jest.

  5. Paul Hewitt says:
    March 14, 2010 at 3:01 PM

    You’re more than welcome, David. I think you “get it”. I look forward to reading your comments!

    Paul

Leave a Reply