Is Donald Trump Done for?

[Chose]

They are. All you are doing in that case is creating a different model. Taking multiple other models and combining them somehow to spit out a probability is still doing the same thing. Adjusting the output based on a variable you don't think the model is accounting for is just adjusting the model. You're still just analyzing past events.

Yeah, and then when you're done analyzing you make a prediction about a future event. Forecasts and predictions are statements about future events. Nate Silver's probabilistic forecasts are predictions, which is what started this whole thing. As I showed, he even calls them that. It's like... just what they are.

The only reason you guys started saying a forecast like "There is a 70% chance of rain tomorrow" isn't a statement about a future event in the first place was because you didn't want to admit it was a prediction, when it clearly is, and you both know it. 538 website is filled with probabilistic predictions called just that. So let's just move on.

 

[Lexicus]

You're just using them as one set of data points in your model.

The "events" concerned here are the outputs of the model, not the inputs.

 

[rah]

based on the inputs. Ya gotta start somewhere.

A different model is just one with a different methodology or one with different starting assumptions (inputs)

 

[Chose]

The "events" concerned here are the outputs of the model, not the inputs.

Like just go read about how this stuff works. Probability forecasting summarizes what is known about, or opinions about, future events. That is the definition. You cannot say it's about past events, if it is, then it isn't a probabilistic forecast.

 

[Lexicus]

Like just go read about how this stuff works. Probability forecasting summarizes what is known about, or opinions about, future events. That is the definition. You cannot say it's about past events, if it is, then it isn't a probabilistic forecast.

By now I realize you have nothing interesting to say on the subject of probability theory or the nature of "chance" itself, so let's take your advice and just move on.

 

[Chose]

By now I realize you have nothing interesting to say on the subject of probability theory or the nature of "chance" itself, so let's take your advice and just move on.

Let me get this straight. Because you repeated make trivially incorrect claims about probability, causing me to continue to try to explain the fundamentals to you, you then concluded that I must not have anything interesting to say about it. That says a lot about you.

Like think about your claim for a second that these two statements are literally the same:

1.) "There is a 70% chance it will rain tomorrow"
2.) "it rained in 70% of our computer trial runs."

Right away, you should have realized that (2) could be true while (1) is false, and vise versa, since they are literally about different things. One is about a past trial, the other, tomorrow. This fact alone reveals that the content of the statements are in fact different. Interestingly enough it's almost certainly the case that statement (1) isn't perfectly true, and we will never even know if it was true, whether it rains or not. Maybe it was close, maybe it wasn't. The second statement could easily be factually true and is easily verifiable.

So you're wrong about this. These statements are totally different. I seriously would like to agree with you but then we'd both be wrong.

So like, it makes me wonder, why are you saying something so obviously false? You go on about people arguing in bad faith and being dishonest and all that stuff in political discussions. But you won't even concede what essentially amounts to a tautology. How could we ever have a meaningful discussion about anything actually interesting if we cannot agree on this? I'm thinking we can't. We are bound to talk past each other and misunderstand each other.

Like any actually interesting discussion is going to be full of these kinds of basic things that must be understood. When I see someone so willing to dig in their heals about something so obviously wrong, I know it's just not going to work.

 

[Lexicus]

Like any actually interesting discussion is going to be full of these kinds of basic things that must be understood. When I see someone so willing to dig in their heals about something so obviously wrong, I know it's just not going to work.

On the contrary, the most interesting discussions occur when these kinds of "basic things that must be understood" are challenged, or deconstructed, or just looked at in a new and different way. But that can't really happen when people are so hung up on being "right" that they are literally unable to think originally and instead bombard the other people in the discussion with dictionary definitions just to prove how "right" they are.

 

[Chose]

On the contrary, the most interesting discussions occur when these kinds of "basic things that must be understood" are challenged, or deconstructed, or just looked at in a new and different way. But that can't really happen when people are so hung up on being "right" that they are literally unable to think originally and instead bombard the other people in the discussion with dictionary definitions just to prove how "right" they are.

Dude I have no problem with challenging the fundamentals of anything, but I still need to know what the hell you are talking about. You don't challenge the fundamentals of probability theory by saying a prediction isn't about a future event. At that point we were quite explicitly talking within the framework.

If you want to move outside the framework and challenge it then great, but you have to say that is what you are doing. I don't just say , yo guys 1+3=3, and then expect everyone to realize I'm making a philosophical challenge to the foundations of knowledge itself. More than likely people will just think I don't know how to do arithmetic.

 

[metalhead]

Either it rains tomorrow, or it doesn't. Saying there is a 70% chance of rain doesn't actually articulate a possible outcome. It doesn't tell me whether it is going to rain.

How can you say something is a prediction when it doesn't actually say what the outcome will be? How can you say something is a prediction when your statement is not testable against what happens tomorrow?

 

[brennan]

So a prediction cannot be a prediction unless it turns out to be correct? An odd use of language.

 

[Chose]

Either it rains tomorrow, or it doesn't. Saying there is a 70% chance of rain doesn't actually articulate a possible outcome. It doesn't tell me whether it is going to rain.

How can you say something is a prediction when it doesn't actually say what the outcome will be? How can you say something is a prediction when your statement is not testable against what happens tomorrow?

I'm just using the generally accepted definition of a prediction, which is, most basically, a statement about a future event. This statement may be probabilistic or deterministic in nature. By this commonly used and accepted definition, both would be considered predictions.

Just again to prove I'm not just making up this definition, I've used wiki, and 538 to repeated reinforce this.

https://projects.fivethirtyeight.com/2018-college-football-predictions/?ex_cid=rrpromo
https://projects.fivethirtyeight.com/2018-nfl-predictions/?ex_cid=rrpromo

EDIT: And again, this all started with you claiming that the 538 election forecasts were in fact not predictions. This wasn't the main point you were making, I did say in my post that I agreed with the main point, which was that just because trump won doesn't mean the 538 forecast was wrong, this is absolutely true, but I had an issue with your use of the word prediction.

 

[Lexicus]

So a prediction cannot be a prediction unless it turns out to be correct? An odd use of language.

I say "there is a 70% chance it will rain tomorrow." It rains. Was I correct?

 

[Timsup2nothin]

I say "there is a 70% chance it will rain tomorrow." It rains. Was I correct?

Maybe. If the weather, through widely accepted methods, was more accurately predicted by "there is a ten percent chance of rain tomorrow" and it rained then I'd say that you just got lucky.

Sorry, I'm guessing that you weren't asking me.

Carry on.

 

[metalhead]

I'm just using the generally accepted definition of a prediction, which is, most basically, a statement about a future event.

Generally accepted by whom? Every definition I've seen or used with regard to the term "prediction" involves a claim about an event or outcome.

Under your definition, "Tomorrow is Friday" would be a prediction. But that's stupid, because nobody uses the word "prediction" to simply mean a description.

 

[rah]

If you hear on the weather channel that there's an 80% chance of rain tomorrow, do you bring an umbrella?

 

[Timsup2nothin]

If you hear on the weather channel that there's an 80% chance of rain tomorrow, do you bring an umbrella?

I don't even take an umbrella if I'm on my way out the door and it is currently raining. In fact, I do not own an umbrella to take.

 

[rah]

I would try to come up with an equally clever comeback, but I don't either.

 

[Chose]

Generally accepted by whom? Every definition I've seen or used with regard to the term "prediction" involves a claim about an event or outcome.

Under your definition, "Tomorrow is Friday" would be a prediction. But that's stupid, because nobody uses the word "prediction" to simply mean a description.

Yeah, a claim about a future event or outcome, right, I said event. Which includes probabilistic predictions. "Tomorrow will be Friday" I suppose is a kind of generic prediction, silly of course and outside of probability theory.

I mean, I gave you the wiki and more importantly we have the original site in question, which happens to be one of the most popular forecast sites around, 538, calling his probabilistic forecasts predictions... like, what more do you want?

 

[Lexicus]

which happens to be one of the most popular forecast sites around, 538, calling his probabilistic forecasts predictions..

We are way past the semantic issue. Nate Silver would be the first to agree that the predictions on that website are actually just reporting the outcomes of statistical trial events that already happened.

 

[Chose]

We are way past the semantic issue. Nate Silver would be the first to agree that the predictions on that website are actually just reporting the outcomes of statistical trial events that already happened.

Okay well the question was asked so I answered it. And no that isn't the case. For example, for each head to head game in the NFL on there, the predictions are actually based entirely on current Elo ratings, which requires no trial events at all and instead uses a few pieces of historical data and mathematical formulas to rate each team, the rating difference will tell you the % victory chance, not any trial event.

He then uses that to simulate the entire season over and over again, in actual trial events and create predictions of the teams likelihood to win the division, super-bowl, all that good stuff, which yes those numbers would be the same as the outcomes of the most recent trial events. But the nature of the claim is of course different.

So basically some predictions would just use the trial data, some use something entirely different. The thing that makes it a prediction is that it is about an event in the future, it isn't about the source of the prediction, which may or may not be trials from some model.