I can click on Google and discover that where I live there is a 20% chance of rain tomorrow. (This number is derived via satellite photos, difficult calculations, etc.) I can also use historic data to guess the probability of rain on the same date two years hence; suppose that number is also 20%.
It is likely that BOTH these numbers are "correct" in the sense that if we amass a large collection of such 20% predictions we find that about 20% of them come true. Yet clearly the long-distance probability estimate is fuzzier and certainly seems "unreliable." But I'm not sure how best to quantify statistically the difference between the short-term and the long-term probability estimate. One way is to observe that the long-term estimate, if repeated periodically, will fluctuate over a broad range, while the estimate closer in time to its target won't have time to fluctuate. (Anyone have a better approach?)
This examples were for a binary choice. Multi-choice is more difficult; prediction markets themselves indicate that the nominees or next Potus may well be someone not yet "on radar screens."
If we trust prediction market, there is about a 55% chance that Trump will face Biden in November. Trump will be about a 3-2 favorite.
There is a 23% chance that Trump will face some other Democrat. Who is favored then? It depends on whether the Democrat is some last-minute substitute for a faltering Biden, or is himself a viable star. Prediction market thinks Newsom would be a favorite against Trump.
If Trump is not the nominee, the nod likelihood will split (in roughly equal portions) to DeSantis, Haley or Other. DeSantis would be favored to lose, Haley favored to win, and who knows about Other?
Overall, the GOP is very slightly favored to win the electoral vote, with a 3% chance for an Independent (RFK Jr??). The Democrats of course are heavily favored to win the popular vote.
I think that we still have a year to go before the election, and nobody has actually won the nomination for either major political party. A lot of things will change over the next 12 months, so we don't even know what will be on people's minds then....
With exceptions, I think what you write is correct. But that doesn't make the current predictions irrelevant. Some people might want to make travel plans, or even purchase weapons based on such year-away predictions. Foreigners seeking a better life may be in the process of changing their aspired target away from U.S.A.
If Newsom really is more likely to beat Trump than Biden is, kingmakers should be exploring options NOW.
(Though beyond the scope of this sub-debate, please note that we're not home free in the 22%-likely event that Biden beats Trump in the electoral college. The election then is likely to be as close as 2020's was; the insurrection that follows may dwarf the January 6 event in its casualty count and other undesired results.)
And anyway, is the 2024 election campaign really so difficult to predict? Almost half of America will prefer Putin's side in any European war, because Trump tells them to. Welfare and infrastructure will be big issues, but many Americans will prefer to focus on firing teachers who "preach wokeness." The economy will be important but "Biden's economy" is better than any before by some metrics, yet Trump fans insist the economy was better under Trump.
Judged from right now, the most likely single outcome in the 2024 election is that Trump will eke out a narrow electoral college win over Biden. Let's not "hide our heads in the sand" about this.
