Comments on: Does this prediction market chart look predictive to you? http://www.midasoracle.org/2007/08/13/does-this-prediction-market-chart-look-predictive-to-you/ Prediction Markets For All Tue, 02 Dec 2008 17:13:42 +0000 http://wordpress.org/?v=2.6.5 By: Chris. F. Masse http://www.midasoracle.org/2007/08/13/does-this-prediction-market-chart-look-predictive-to-you/#comment-16063 Chris. F. Masse Mon, 13 Aug 2007 17:02:03 +0000 http://www.midasoracle.org/2007/08/13/does-this-prediction-market-chart-look-predictive-to-you/#comment-16063 Wait. In the Harry Potter case, the literature experts were able to analyze the past (the past writings from J.K. Rowlings) and the NewsFutures prediction market did output an objective probability (around 75% climbing later to 85%). Is that OK if I call "absolute accuracy" the determination about whether a prediction market got it right or wrong, and "relative accuracy" whether a prediction market got it better or worse than a competitive institution? Wait. In the Harry Potter case, the literature experts were able to analyze the past (the past writings from J.K. Rowlings) and the NewsFutures prediction market did output an objective probability (around 75% climbing later to 85%).

Is that OK if I call “absolute accuracy” the determination about whether a prediction market got it right or wrong, and “relative accuracy” whether a prediction market got it better or worse than a competitive institution?

]]> By: Michael Giberson http://www.midasoracle.org/2007/08/13/does-this-prediction-market-chart-look-predictive-to-you/#comment-16062 Michael Giberson Mon, 13 Aug 2007 16:54:23 +0000 http://www.midasoracle.org/2007/08/13/does-this-prediction-market-chart-look-predictive-to-you/#comment-16062 Secrecy? False Signals? Lack of inside information? Sounds like the Harry Potter markets to me. I guess that was why experienced traders stayed away. ;) I think ESS is absolutely right as you point out in claim #2, above. But the claim in #1 is fundamentally flawed. What do you mean by the term "absolute accuracy"? Is a normal six sided dice absolutely accurate only if you roll exactly the expected value of 3.5 each time? Or is it only good if you roll a number "close enough" to 3.5 (say 3 or 4), and therefore you take a roll of 1 or 6 as symptomatic of a problem to be explained. Secrecy? False Signals? Lack of inside information?

Sounds like the Harry Potter markets to me. I guess that was why experienced traders stayed away. ;)

I think ESS is absolutely right as you point out in claim #2, above.

But the claim in #1 is fundamentally flawed.

What do you mean by the term “absolute accuracy”? Is a normal six sided dice absolutely accurate only if you roll exactly the expected value of 3.5 each time? Or is it only good if you roll a number “close enough” to 3.5 (say 3 or 4), and therefore you take a roll of 1 or 6 as symptomatic of a problem to be explained.

]]>