I’m saying not enough have extra insight to make a difference. I’m saying they’re almost all making choices that are based on polls, hunches and “momentum”.
Nobody is saying that betting markets are a magic crystal ball whose estimates are perfect. (Nate Silver’s numbers may be about as good or even better — does he have a page showing odds on the nomination, and on November?) As Chronos points out, fees and regulations hamper the markets. If the U.S. legalized such bets, the market efficiency would increase greatly.
Markets for betting on football, basketball, etc. are very large. Some people can beat those markets overall, but it isn’t easy.
But it’s rather silly to say people are rejecting “sure-thing” bets because they have better investment opportunities. What are those better-than-certain opportunities? That the stock market will go up? That the price of gold will always go down??! I’m not betting on politics — *because I don’t think I’m smarter than the average punter. * Still it is … interesting … that several of you are certain you are smarter but can’t be bothered to pick up a few thousands in extra pocket money. 
There’s some chance that, due to illness or whatever, Trump or Biden will need to be replaced as candidate. If Biden should drop out before the Convention who would replace him. Neither logic not “The DNC” would make Sanders the choice, IMO. Klobuchar?? If not her, then who? I’m not sure that the idea of sending Hillary in as the late-inning reliever is so far-fetched.
Are you kidding? Of course it’s far fetched. And how about Cuomo? Is it “not far fetched” that he’ll drop being governor in the middle of the Corona outbreak and run for president?
And the money question: are those scenarios remotely more likely or even as likely as Sander winning big in the remaining contests?
You can buy ‘Yes’ for Biden to win the Dem nomination right now on PredictIt for $0.77 per share. Those shares will each be worth $1.00 in June if he wins the nomination. If you believe that Biden winning the nomination is a sure thing why aren’t you buying those? What other investment are you putting your money in that has an ROI of > 25% over the next three months?
Those gambling sites are illegal from my jurisdiction, so you’ll have to come up with some other stupid comeback.
It’s not a stupid comeback. I’m genuinely trying to understand your over the top hostility toward prediction markets. Did a prediction market kick your dog? Put its thumb in your beer?
There’s $14 million dollars in the PredictIt market for the Dem nomination. That’s a financial incentive for people to be making well researched, sound decisions. You seem to be emotionally invested in the idea that it’s just people guessing, but with that amount of money floating around there’s more than enough reason for people to take it seriously.
It is true that prediction markets are prone to favorite-longshot bias, so there will almost always be a little weirdness in the tails. However, that is something that can be corrected for and thus not enough reason to dismiss prediction markets entirely.
Joe Biden at $0.77 is capturing something real that is happening and might not be reflected in the polling or his delegate lead. There’s a non-zero chance that Joe Biden drops out of the race before the convention. Possibly due to COVID-19, maybe due to Tara Reade, or possibly some other reason that I haven’t considered.
I don’t personally think that’s the case, so I bought some Joe ‘Yes’ today, but it is ridiculous to completely dismiss his drop from $0.90 a week ago to $0.77 today as nothing. That represents a lot of real money moving around.
Finally, your objections to where PredictIt bettors live or the color of their skin are truly bizarre. I can’t imagine why you think that has a bearing on their ability to make informed decisions.
What it’s capturing is people dreaming of a longshot bet.
Markets don’t predict things. They match supply and demand. The reason I push back hard on the betting market is because I find the “magical powers of the market” shit extremely annoying.
You are simply wrong about this.
There’s a pretty solid body of academic research that shows this.
I recommend FORECASTING ELECTIONS: COMPARING PREDICTION MARKETS, POLLS, AND THEIR BIASES by David Rothschild.
"Conclusion
In 2008, Fivethirtyeight, a debiased poll-based forecast, offered to the general public a more accurate forecast than raw poll numbers or raw prediction market prices."
Hmmm. Also, Rothschild works at Predictit, right?
Wow. That is some really dishonest selective quoting there. Here’s the whole paragraph…
Intrade outperformed 538 when well known biases, like favorite-longshot bias, were corrected for.
But this thread is not about whether 538 is better than prediction markets. It’s about whether prediction markets have any value at all, and they clearly do even if 538 performed slightly better than them in some election or other.
Finally, David Rothschild does not, and has never, worked at PredictIt.
Uh huh. Taking out bias isn’t a switch you flip. It’s decisions based on past performance. Your article is him building a perfect “debiasing” based on making decision markets look good.
Oh sorry, he founded PredictWise. What a totally honest clearing up of my misunderstanding. Thanks.
And further, have you or anyone here that reports betting market numbers ever offered the “debiased” version? Is your Rothschild running a debiased betting tracker? Because I would honestly love to look at it.
https://www.biz.uiowa.edu/faculty/jberg/papers/ResultsIEM_2001May.pdf
“Results from a Dozen Years of Election Futures Markets Research” is cited in the above paper, and answers the question about raw market results.
“Polling data in the figure represents the absolute errors of the final pre-election polls averaged across all candidates and across all of the major polling organizations for which data was available. The market outperformed polls in 9 of 15 cases according to each measure (election eve closing prices and last week average prices). **Across all elections, the average poll error was 1.91% while the average market error was 1.49% and 1.58% by the two measures. **In a few cases (the 1988 and 1992 U.S. Presidential elections) the market dramatically outperformed polls. The worst outcome, the 1996 U.S. Presidential election, is a peculiar one that gets additional attention below. In the majority of other cases, the market does about as well as the average poll, sometimes worse but often better, even if by a small margin.”
I can confirm that he did make this statement on a recent 538 Politics Podcast (either “Is Even Coronavirus Partisan?” or “The Democratic Primary Is No Longer Partisan”, but I’m not going to bother going back and listening to them because Clare Malone isn’t on right now them to provide her dry sarcastic response when Nate Silver or Galen Druke say something absurd) because it made me laugh out loud, and it is definitely a shift from the previously more serious take on the “Scottish Teens” betting markets.
And I agree, “betting markets” are stupid. They are influenced by media focus and viral media trends without reflecting how the actual electorate is responding to those trends. Polls are often of questionable statistical accuracy and many are far from unbiased but at least they are a direct measure of public perception rather than some wonk’s belief about what the public should think.
Stranger
Seems like someone should be able to make lots of money with a few scripts that measure google news ngrams and twitter hashtags and measure the extent to which betting markets overweigh widely-discussed events.
Betting markets are a useful signal. They’re not perfect, they’re noisy. But they’re still useful. And the longer they are around, the better they are likely to get because there are financial incentives to fixing non-random measurable errors.
Yes and? Are you stating with confidence that nobody makes money betting against completely unrealistic bets? If you went and bet on Biden right now, it is free money.
Eta: well, pre-coronavirus. Now, there’s a probably a .5% chance he’ll be dead by then.
People have been betting on elections longer than there have been polling companies.
What is your problem. You asked, “Hmmm. Also, Rothschild works at Predictit, right?” I answered that question straightforwardly and directly. I don’t know how I could have more honest.
Predictwise, Rothschild’s site (which is not a prediction market), did have a debiased tracker. It hasn’t been updated since January. I’m not sure why. Link.
Ok, whatever. What debiased betting market numbers do you have on hand right now? Your twitter bot doesn’t offer them up, right?
My Twitter bots do, in fact, do some debiasing when it is warranted, but my algorithm is pretty crude all things considered.
But my Twitter bot is beside the point. Whether my debiasing is perfect or hopelessly flawed doesn’t change the fact that there is a large body of academic research that shows that prediction markets perform about as well as, and in some cases better, than aggregated polling.
There is value in looking at prediction markets. That’s a fact. You can turn a blind eye to it if you wish, but that’s to your detriment.