Hello Professor Tyler Cowen and all the commenters,
#1. Professor Lance Fortnow made a specific point: taken one day before Election Day, the TradeSports’-s prediction markets of the individual races for the US Senate were accurate (provided that Virginia and Montana go democratic).
#2. Professor Lance Fortnow DID NOT SAY that the TradeSports’-s prediction market for the control of the US Senate was accurate. Please, don’-t put words in his mouth.
#3. Analysis reports from economists and statisticians are coming, but, please give them time to digest the data…- once the dust has settled.
Thanks for your attention,
I haven’t read MR’s commentary on that, but…I love markets too, but the markets were all over the place on this race, on the Montana race, and pretty much every other competitive one. Believe me–I was alternately hemorrhaging and raking it in depending on what hour it was last night. Saying “the markets got everything right” is no different than taking a sample of all DC pundits, or getting 100 different estimates from one pundit, and saying that “the DC pundits” or “pundit X predicted every single race correctly.”
Even one day before Election Day, I am pretty sure (not positive) that VA fluctuated on both sides of the fence. He might have selected a fortuitous point in time, but I’m (99%) sure that Missouri and Virginia were both mispriced at different times yesterday. And if he’s using the closing numbers, well, I don’t know how Tradesports calculates closing prices, but I’ve noticed multiple instances in which the ‘closing price’/last trade for a contract did not exist in the previous 24 hours. As in, there was no trade at that price within 24 hours… So take those numbers with a grain of salt.
Additionally, TS’ numbers, when aggregated, implied six Democrat pickups, but also stated that the Republicans had a much better than even (70%) chance of holding the Senate. As you pointed out, Prof. Fortnow said nothing about that, but it’s equally damaging to his overall implication that the prediction markets are our most efficient ‘polling’ tool. If you say, “All these races are [ex.] 60-40 Dems, but we think that the Dems will fail to pick up one random seat out of the six,” that has a very different meaning from “each individual race was predicted correctly.”
PS–in my experience, poli futures have three possible stages: that a race is “close,” which I’d define as 58-42 _going either way_; “leans decidedly (wherever)” maybe 59 to 73ish; and “likely (party),” as longer odds than that. Now, if you got the mean weighted price within each of those bands, then you’d be talking…but it certainly seems to me that, at any given point in time, the “current price” within those bands is a random function of the biases/intuitions of the major traders of that future. AKA, TS showing MO as favoring McCaskill over Talent 53-47 vs. Talent over McCaskill 53-47 mean the same thing.