That link describes fifteen statistical fallacies. Which of them do you feel are present in my post?
The most obvious one is the metric fallacy (that graphic references it as both “McNamara’s Fallacy” and “Danger of Summary Metrics”) but there is also “Data Dredging”, a variation of “Survivorship Bias”, and implicitly a sampling bias by assuming the parameters of successorship today are the same as they were 150 years ago.
In general, you can’t just look at the summary statistics of a type of event, especially one with a limited number of instances, and make the frequentist induction that the relative incidence of past events reflects the future probability, especially with respect to a single instance to any defined level of confidence. As an example, if you rolled a six-sided die ten times and came up with four rolls of ‘5’, you couldn’t conclude that rolling any six-sided die (or even this particular die) has a 40% chance of getting a ‘5’. It just happens that the small sample set that you used had an unlikely but not impossible distribution, and even if the die used to develop the samples is actually biased it doesn’t reflect a general bias in all six-sided dice in the world.
If someone wanted to build an actual model that might possibly reflect all of the major contributors to the likelihood of a Vice President being promoted or elected to the Presidency, you’d need to list out the pertinent parameters of different conditions in which that consequence would occur, develop prior estimates of likelihood, and then apply the previous presidential administrations (both where the consequence does and does not occur) to generate posterior estimates to approach a general expectation. Even then, the great variability in confounding factors and the relatively low instances of vice presidential succession would provide a model with substantial uncertainties about how it would apply to the case of J.D. Vance.
Again, I’m not trying to single you out for criticism but there is a general tendency to believe that some ‘analysis’ is better than ‘no analysis’ even if the fundamental basis for the analysis is flawed, and making the underlying assumption that the future will be statistically reflected by the past without consideration for confounding factors and biases is just not methodologically correct.
Stranger
I’m going with Zero. Harris is going to blow The Clown Car off the road in November. Vance will catch the heat for the loss and he is likely never going to rise farther. Might not be Done, but will never go higher.
I like your confidence but I think a lot of people don’t realize it’s a big country with many Americans who have a bizarre nostalgia for the first Trump presidency and/or are convinced that Joe Biden and Kamala Harris are the twin harbingers of doom. Vance doesn’t seem to be doing the GOP any favors with overall appeal but I think the media tends to understate the appeal of his views, as extreme as they are, among a demographic of low information and extremist people who often don’t show up in the polls or vote regularly. And as much as Trump no longer attracts the large crowds that he did in 2016, there are still a lot of people who would rather vote for him, even if reluctantly, than a ‘liberal’, Black-South Asian woman for president.
Thus far, I’ve seen people predicting a 400 Electoral College vote blowout for Harris, a 20-point popular vote beatdown, or Trump actually dropping out and going on the run to Russia or North Korea with zero credible argument behind it and in apparently obliviousness to how that same attitude worked out previously.
Stranger
I don’t feel you’re correct.
Quantitative fallacy (aka McNamara fallacy) is always a possibility. But it’s not something you can simply declare is happening. You need to demonstrate the predicted result isn’t occurring and then cite quantitative fallacy as a reason for why the results that were predicted aren’t occurring. That’s obviously not the situation here. Regardless if you feel that Vance’s chances of someday becoming president is zero percent or a hundred percent, you can’t cite that feeling as proof that a sixteen percent prediction is wrong.
Danger of Summary Metrics doesn’t apply here. It’s a potential problem that can arise when you are analyzing two sets of data and misinterpret the correlation, which is not what happened here.
Data dredging is the process of running a set of data repeatedly through a series of tests and then only using the results which confirm the desired goal. That obviously didn’t happen here; there was no repetition.
Survivorship bias is when there is a hidden gateway that isn’t acknowledged. An example would be analyzing the use of seatbelts by examining people who are treated in hospitals following car accidents. The hidden gateway is that some people who are in car accidents are killed and don’t go to any hospital. I don’t know what you imagine the equivalent is here.
You’re correct that you can’t use past results to make predictions of future results in all cases and correctly cite dice rolls as an example of this. But this is not about rolling dice. We’re talking about political careers not dice rolls. Past events in a person’s political career do affect future events in their career in a way that doesn’t occur in dice rolls.
So if you want to argue that presidential campaigning has changed from this point on and the last hundred and fifty years of presidential campaigning therefore have no bearing on future presidential campaigns, you’re going to need to back up that argument with evidence.
And if you want to argue that there are other factors which have a greater predictive value, I won’t dispute that they may exist. But the burden is on you to produce those other factors and offer the evidence that they have greater predictive value.
I’m not really interested in “produc[ing] those other factors and offer[ing] the evidence that they have greater predictive value” because I don’t think it is actually possible to make a statistical prediction about Vance’s likelihood to promote up to the Vice Presidency, and I don’t think anyone else can, either based upon past history. I did lay out the approach that could be used to construct and validate (to the extent possible) a Bayesian predictive model but I don’t think it could actually be done to a useful credible interval, and epistemologically I think it is kind of pointless because whether the estimate is 99% or 1%, Vance remains a disturbing person with a dangerous ideology who should be vigorous opposed in all public forums regardless of whether you think he is likely to become POTUS or not.
Stranger
Can’t we limit a few key factors reasonably well and use those to reach a conclusion? For example, we know Trump is not 75% likely to become POTUS in 2024, and he’s not 25% likely. So we limit him to a range of about 40-60% likely, and pick a number somewhere in there. Likewise, with his life expectancy. Based on that sort of guesswork, I don’t see why we wouldn’t come up with some reasonable estimates, and I think that’s what we’ve done here collectively. Based on the responses, I’d say in this thread we’ve guessed that Vance has a chance of serving as POTUS somewhere between 10% and 30%, probably on the lower end of that range. I haven’t seen an argument so far that places his chances much higher, and 0% or anything close to that is ridiculous. Just being a US Senator under 50 years old probably gives him a 1% shot, or a non-zero shot for sure.
Gerolamo Cardano, Pierre de Fermat and Blaise Pascal felt you could, so go argue with them.
To answer the title question, I hope it’s 0%.
I’m generally a pessimist, but the last time this Asshole ran, he was somewhat of an “unknown”. That’s not the case anymore. The people that rejected him 4 years ago will do the same, with a lot more people on board. How many Republicans are openly denouncing it now? I don’t think he has a Pack-Rat’s chance in a room full of Dachshunds. (This is something I actually know about!)
I believe the Americans with a bizarre nostalgia are a tiny but vocal group, in a huge state of denial. And are about to be given a Boot to the Throat.
Shit. I’ve been a Republican my whole life. I won’t vote for anything or anyone with his taint on it.
The last time “this Asshole ran,” it was 2020, and he had been President for four years, so definitely not an unknown. I think you’re thinking of 2016.
Indeed. Thank you. I’m going to drink a beer in your honor!
Cheers!
Ah! I just pulled out a Chocolate Raspberry Stout! Didn’t know I had any of those left!
Listing the names of three pioneers of probability does not bolster your argument that we can make a credible estimate of J.D. Vance’s likelihood of becoming president through crude summary statistics.
Not nearly enough, especially compared to the number of ‘normal’ pre-MAGA Republicans that have been driven from public office and the ranks of GOP leadership, nor those voters who will hold their nose and vote for Trump anyway because Fox News is constantly beaming at them what tragic devastation Biden and now Harris have done to America, or what a draft-dodging “Stolen Valor” criminal Tim Walz is.
Trump doesn’t need a majority of voters to win; he just needs a (slim) majority of voters in swing states, and maybe not even that if the GOP-cum-MAGA party clutches enough statehouses to contest vote certification or throw out enough unfavorable votes on thin but legally sustainable premises as adjudicated by conservative-leaning federal district courts or (Og help us) the United States Supreme Court which has already aptly demonstrated bias and corruption.
People like to highlight how Biden had a record turnout in 2020 and (narrowly) achieved a majority of the popular vote but he also cleared three critical states (Wisconsin, Arizona, Georgia) by less than a percentage (in the case of Georgia, less than a quarter of a percentage), while Trump also had a record turnout (by over eight million votes) despite having one of the worst net negative popularity rankings of any candidate since modern political polling began. The notion that Trump doesn’t have, in your words “…a Pack-Rat’s chance in a room full of Dachshunds” is just not born out by evidence even before consideration for the voter registry manipulation, “Stop the Steal” fuckery, and just how much control the GOP has (and how much more it may well gain) at the state level to throw legitimate vote counts in question regardless of objective evidence.
The idea that a Harris win is a fur-sure thing that you can take ot the bank is one of the most dangerous notions imaginable, a virtual repeat of what got us into this fucking mess eight years ago in the first place. Fortunately, the Harris campaign is not (at least for now) taking a win for granted, but if you think there is a default majority block of voters who will unquestionably elect Harris you haven’t really sampled all the beer on tap.
Stranger
Evidence? I don’t have no Evidence! I don’t need no steenking Evidence! 
Mostly wishful thinking on my part. But what I’m seeing is Donny is displaying (to a greater degree) his Bat-Shit Crazy and people are finally starting to acknowledge it and talk about it. I think he’s done, and Vance will pay, and never make it to the Oval Office.