I honestly don’t think he can do it twice. If Nate is saying Trump has a 10% chance of winning and Trump still takes the presidency, then what future relevance does Nate still have? 10%? Might as well be 0.1% or 99%. Nate makes himself relevant or irrelevant tomorrow.
2 Likes
That’s my point. Nate isn’t in the guessing game. They never have and never will make a prediction about who will win, they only state the probability. Then everyone blames 538 for making the wrong guess… that they never made.
9 Likes
You said guessing, not me. I was talking about relevance. If Nate says Trump has 10% chance again, and Trump still wins, then Nate’s models, as correct and well formulated as they might be …are proven irrelevant.
let’s say you have this fancy model for when the ice cream truck is going to show up. By your calculations, you give the ice cream truck a 90% chance of showing up in November every 4 years. But for two cycles running, the ice cream truck does not show up. How relevant is your fancy model now?
Nobody is going to pay attention to you because to most people it is like the little boy who cried wolf. It has nothing to do with math or quality of the construction of the models. If it doesn’t explain reality accurately, then what good is it?
I would like to make a friendly wager.
If Trump wins there will be widespread violent protests in the streets.
If Biden wins, there will not. And Trump pisses and moans a little and leaves office the same as every other president.
Then there is this other problem with polls:
There are two types of people who are going to vote for Trump. Proud to vote for him (trucks with big flags), and those who don’t really like him but can’t stomach a Biden Harris future.
The first group will answer polls truthfully. The second group are so gunshy to admit they will vote for trump for fear of being called a racist, misogynistic, bigot that they will not even admit it to a stranger on the phone.
People care so much about the perception of others that it borders on pathological.
2 Likes
In ll fairness, IIRC the prediction 4 years ago was 30% that Trump would win, or about the same likelihood of a batter getting a base hit.
1 Like
So the final projection on the Fivethirtyeight website is now “locked down” and will no longer be updated. I guess I can stop frantically hitting “refresh” on my browser and find a new outlet for my nervous energy.
The official odds:
Trump 10%
Biden 89%
(I’m assuming that extra 1% includes the “giant meteor” scenario.)
5 Likes
By that reasoning though, Nate is in an impossible situation from the get-go.
A 50/50 coin model says you can flip heads 100 times in a row and the model is still right. That’s perfectly reasonable and is how probability works. Most people don’t understand that, however. Peoples’ intuition about probability is mostly wrong and we don’t generally have good heads for it (pardon the pun).
It sounds like you’re saying that to be “relevant” Nate must make predictions and be right by luck sometimes. That’s what everyone else does, but it’s not statistically rigorous nor what Nate is trying to do. People see “outcome X is 90% likely” and think “outcome X will happen”. If outcome Y happens, the prediction was still right. Nate is one of the few people willing to stick with the probabilities. Weather reports are rarely “correct” in the way people want them to be, but they are still relevant. We still check them every day.
8 Likes
IDK about anyone else, but stories like that help give me a sense of scale that brings emotional grounding to other events and decisions.
Truer words were never spoken. And is another way to state my point: if the models don’t match reality in some form at the end of the day, people will stop paying attention. The meteorology example you gave is a good one: the weather people are often wrong. But they are right far more than they’re wrong. If Nate Silver is not eventually right more than being wrong, in the colloquial sense, people will stop paying attention. This has nothing to do with understanding probability. I’m talking about people.
2 Likes
I suspect that Biden will win, and there will be throughout the coming week isolated spasms of violence, caused by rural Trump supporters at various state capitals, urban areas, et cetera. Not at home, no, these thugs will try to “invade” those neighbourhoods they consider hostile.
If we are lucky, these episodes will remain impotent attempts to intimidate. But I fear a dozen Kyle Rittenhouse style incidents will happen, as armed, hate-drunk males decide to strut around.
5 Likes
6 Likes
It could be the reverse, especially in places where half the houses have confederate flags, so the same thing could happen for democrats.
I don’t know. That’s where we depart. 538 doesn’t “cry” anything; it’s not their role. Anyone giving them shit fundamentally misunderstands what they’re doing. That’s not their fault and any outlet looking for the most comprehensive and accurate modeling will still look to them because they’re the platinum standard. Unless polls become absolutely irrelevant, which ain’t gonna happen, we’ll still be having this conversation in 4 years (or two when Dems get a supermajority in the Senate 
).
This is like all of the jocks thinking they’re cheating when the nerd tells them to read CliffsNotes instead of The Tempest and then they beat him up for failing the test. 
Here’s my take, as I’m sure everyone was anxiously awaiting.
Look at the map on electoral-vote.com, and focus only on very red states. Hover on each state and it shows their vote percentages in previous elections. Make note of which year is most similar to the most recent poll numbers for that state. Repeat as needed until you agree with me that this thing seems to be shaped a lot like 2008.
I’m normally a bourbon kind of guy, but this week we bought Bailey’s, Butterschotch Schnapps, and Ketel One. Last night I stocked up on dairy and juice.
So while a Bulleit on the rocks would definitely be inappropriate at this hour, I can put together one hell of a coffee drink and not feel too bad about it. And if you’ve ever tried Knudsen’s pineapple coconut juice, you know that that and some regular pineapple orange juice will cover up a lot of “sins”, and do so with a luscious mouthfeel that would make any prude very uncomfortable.
1 Like
You people are LITERALLY trying to kill me!
That sounds droolicious. Oh man. Forget all this talk about statistics. Let’s just have some fun. The stress has, in all seriousness, been close to unbearable. I was up at 4:45AM, anxious and had to go for a run.
1 Like
Meteorologists don’t make predictions either. They state the probability of weather events happening based on data. In regions with unpredictable weather, it’s a running joke among laypeople that the weather guy is always wrong” but people still watch them, because the information is still useful. People hear “40% chance of rain” and they make their own choice about whether to bring an umbrella. That’s not a prediction but it’s still helpful.
You keep conflating statement of probabilities with making predictions, except when Nate does it.
4 Likes
Either way, we are about to find out. I just hope that all the BS about ballot fraud and voter suppression was seen through by most of the population and whoever wins, their lackeys chill the Fuck up.
Well, meteorologists give probablitiies (and are paid to do so), but they do also make what would be considered predictions, “it’s going to be a wet one out there today*” and similar things to give a bit of human relatability to their statistics.
*when there’s a 90% chance of rain
Two separate issues that deserve two separate assessments.
7 Likes
