More on Prediction Markets

if you bring too many EYES into the panopticon, it will drive you mad.

Recently, I published my off-the-cuff take on Scott Alexander’s prediction contest for 2023.

The 2023 Prediction Dispatch

Now, you can see a market-based prediction for the same questions at Manifold Markets.

Let’s take a closer look at a few markets:

• 21. Will Donald Trump make at least one tweet in 2023? Manifold says 69%, I said 90% on the theory “it’s likely, and I’m guessing 90% for everything that I say is likely”. I think it is more likely than 70%, but I’m not particularly confident in that prediction.

• 9. Will a nuclear weapon be used in war (ie not a test or accident) and kill at least 10 people in 2023? Manifold says 5%, I said 1%. I really don’t want to gamble on this, and it is a known weakness of prediction markets that they over-emphasize low-probability events.

• 2. Will Ukraine control the city of Sevastopol at the end of 2023? and 3. Will Ukraine control the city of Luhansk at the end of 2023? I said 10% for both, on the theory that it is most likely that nothing will change, and in an off-the-cuff response I’m probably better off not analyzing the situation further. Manifold gives 14% for Sevastopol and 26% for Luhansk. I do think Manifold’s estimates are more accurate¹.

• 15. At the end of 2023, will prediction markets say Joe Biden is the most likely Democratic nominee for President in 2024? I said 20%, based on the “he won’t run again” theory. The prediction market says 79%. I think I’m right, but I am also quite biased: I desperately do not want an octogenarian Biden to run again, and see all sorts of reasons why he shouldn’t.

• 18. At the end of 2023, will prediction markets say Ron DeSantis is the most likely Republican nominee for President in 2024? I said 80%, the market says 54%. If anything, I think my 80% is too low and the real odds are closer to 90%. Donald Trump’s campaign is flailing, and there is no figure in the GOP who can run a credible campaign against Trump and DeSantis. There is also no real doubt that DeSantis will announce a run after the Florida Legislature ends its session for 2023. As the failings of the 77-year-old Trump’s campaign become more apparent, and as DeSantis continues to do well in polls, it is almost certain he will be in the lead on prediction markets.

• 35. Will Bitcoin end 2023 above $30,000? I said 10%, the market says 31%. I have retired from speculating on prices of cryptocurrency, after determining a year ago that it was impossible to short it without participating in exchanges with existential risks.

• 49. Will AI win a programming competition in 2023? I said 80%, the market says 32%. I see this as largely a matter of resolution semantics and the willingness of the media to promote stunts, rather than a matter of the future state-of-the-art of AI.

There are (at least) two dimensions² to a prediction. Likelihood, and confidence. For “Biden runs again”, I say 20% but have low confidence in that prediction. For “DeSantis leads Trump in December”, I say 80% and have high confidence in that prediction.

The magic of prediction markets is that you, me, or anyone can go and change the numbers.

Carbon Removal

I am betting “NO” on this. Carbon removal is a terrible idea promoted by incompetent bureaucrats.

• Doesn’t every bit help fight climate change? No. This is the type of bad logical argument that contributes to incompetent bureaucracy.

• Won’t it become cheaper? Probably not. There are two main types of carbon capture.
  In one, the CO2 is changed back into hydrocarbons such as glucose or ethanol. (Plants tend to do a lot of that). There is no possibility whatsoever this will be more effective than “just use an electric car/lawnmower” — the laws of thermodynamics forbid it.
  In the other, the CO2 is filtered, bottled, and stored underground³ forever (or, possibly, mixed with rock). This is also kind of silly. I don’t see how this cycle can possibly be better than “just use something else for energy”. Especially considering that the crude oil could be used for any number of products rather than burning it.

• What about dealing with the existing CO2 in the atmosphere? I have become extremely suspicious over the years that the powers that be actually want some amount of global warming. Even if there were a magical⁴ lever to change the CO2 concentration, I doubt they would push it back below 400PPM⁵.

Biden’s Credibility

I started this market, so the summary there says it all:

Joe Biden's administration recently said it will end COVID-19 emergency declarations on May 11.

Biden also said to a reporter that the emergency will end "when the Supreme Court ends it", suggesting that the plan is to find a reason to extend the emergency yet again by May.

Will the COVID emergency end on (or before) May 11?

My interest in the market is more hedging and knowledge discovery than in making a profit. Do other people trust Biden on this, or not?

---

as a reminder: you can comment at Manifold. there is no “anti-canvassing” rule, so feel free to show up there.

1. If I cared about winning the contest, that would be important.

2. A third dimension could be “likelihood of some unforeseen event preempting the question under debate”. Some unlikely examples include “Turkiye invades Crimea” or “a Tunguska-sized meteor hits Crimea”.

“Potential for insider trading” is also a dimension, but one that everyone appears to have agreed to ignore for now.

3. Do not get me started on the “CO2 pipeline” discussion.

4. Perhaps “magical” causes people to tune out. You can also imagine aliens showing up with a CO2 pump. Yes, like in Spaceballs. No, I’m not trying to be serious in this footnote.

5. If you read this Newslettr, you have probably heard my rant that focusing on “degrees of global warming” is fundamentally unserious compared to focusing on “increase in CO2 PPM”.