Comments on: Are they afraid? http://www.midasoracle.org/2009/01/08/are-they-afraid/ Prediction Markets + Market Predictions = Collective Forecasting That Pays Off Sat, 20 Mar 2010 08:05:32 -0400 hourly 1 By: Medemi http://www.midasoracle.org/2009/01/08/are-they-afraid/#comment-23254 Medemi Sat, 10 Jan 2009 12:28:47 +0000 http://www.midasoracle.org/?p=12579#comment-23254 hear-hear. - You're getting smarter every day Chris. :) There's too much focus on politics/polls anyway. Why ? Don't answer that, I already know the answer. hear-hear.
-
You’re getting smarter every day Chris. :)
There’s too much focus on politics/polls anyway. Why ?
Don’t answer that, I already know the answer.

]]> By: Chris F. Masse http://www.midasoracle.org/2009/01/08/are-they-afraid/#comment-23253 Chris F. Masse Sat, 10 Jan 2009 00:04:53 +0000 http://www.midasoracle.org/?p=12579#comment-23253 The Lee--Moretti paper leads me to think that the social utility of the prediction markets will not emerge if we focus on comparing them with polls. - (I agree with many of your points.) The Lee–Moretti paper leads me to think that the social utility of the prediction markets will not emerge if we focus on comparing them with polls.
-
(I agree with many of your points.)

]]> By: Jason Ruspini http://www.midasoracle.org/2009/01/08/are-they-afraid/#comment-23251 Jason Ruspini Fri, 09 Jan 2009 21:11:14 +0000 http://www.midasoracle.org/?p=12579#comment-23251 Let's not conflate the different questions here. Can poll analysis do better than markets? Yes, sometimes. Can markets work well where (timely) polls and other rich data are lacking? Yes, sometimes. Do markets have utility beyond forecasting? Yes. Should markets incorporate poll data and poll analysis? Of course, now let's consider Lee/Moretti paper. If you think this paper says anything deep, you are just being fooled by academic signals and terms like "autoregressive", "exogenous" and "Bayesian updating". Markets should incorporate poll data and all this paper says is that, yes, markets smooth noisy poll data. (It doesn't mention that this also reduces error). The paper doesn't actually go anywhere interesting with the math. The idea that earlier polls should have more impact on one's outlook as this model predicts is absurd, and doesn't hold as the authors conclude. The paper is underdeveloped and based on a single event. Let’s not conflate the different questions here.

Can poll analysis do better than markets? Yes, sometimes.

Can markets work well where (timely) polls and other rich data are lacking? Yes, sometimes.

Do markets have utility beyond forecasting? Yes.

Should markets incorporate poll data and poll analysis? Of course, now let’s consider Lee/Moretti paper. If you think this paper says anything deep, you are just being fooled by academic signals and terms like “autoregressive”, “exogenous” and “Bayesian updating”. Markets should incorporate poll data and all this paper says is that, yes, markets smooth noisy poll data. (It doesn’t mention that this also reduces error). The paper doesn’t actually go anywhere interesting with the math. The idea that earlier polls should have more impact on one’s outlook as this model predicts is absurd, and doesn’t hold as the authors conclude. The paper is underdeveloped and based on a single event.

]]>