- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
re: Nate trying to get his mojo back
Posted on 4/24/17 at 12:43 pm to therick711
Posted on 4/24/17 at 12:43 pm to therick711
quote:This Buzzfeed article lists a couple (they include PredictWise which I think is just a betting market and not a model)
Link them. Let's not assume he was better than some, as of yet, undisclosed predictive models.
1
Posted on 4/24/17 at 12:43 pm to buckeye_vol
quote:
No it's not, because if he didn't do that, then his most would have had a Trump at like 5% like the other models instead of 28.6%. It gave Trump almost 6 times the likelihood winning.
You're talking about a different issue now.
Posted on 4/24/17 at 12:45 pm to Iosh
"But those are all liberal! This somehow disproves your premise, which I will continue to sort of vaguely snark at and not furnish the single counter-example which would suffice to disprove it!"
Posted on 4/24/17 at 12:51 pm to Iosh
Asking for data is a crime in your book apparently. Or, in the alternative, I just wanted to see the other predictive models.
Posted on 4/24/17 at 12:56 pm to therick711
quote:Well I'm talking about the probability calculations that FoxMulder incorrectly made, which is why his predictions gave Trump a much better chance than the other models:
You're talking about a different issue now.
This post was edited on 4/24/17 at 12:57 pm
Posted on 4/24/17 at 12:58 pm to buckeye_vol
quote:
His model accounts for polling errors between states and overall nationally. So of Michigan has a polling error, then that error is correlated between similar states, like Wisconsin and Pennsylvania.
=
quote:
oh but he was wrong on this one and knowing that he would give a higher probability in the other states
and the answer is still duh. you don't discover a polling error until the results are in. 'A' for effort.
i'm not even sure what you're arguing. he is getting laughed out of the room for a reason. he put out a prediction...you don't contest that. he said it was a 70% chance (which is high) he's high profile and has a legend built around him and he was wrong...it's really not rocket science. that's the business he's in.
Posted on 4/24/17 at 1:03 pm to DelU249
quote:No, but a model can account for it. You treated the 3 states as independent and came to that calculation of <1% when his model probably had that 10+ times that.
don't discover a polling error until the results are in.
So you're making criticisms about a model, but those criticisms are based on some ignorance of the methodology.
quote:It's high, but it's not that high. 3 in 10 times is more likely then a whole bunch of things that happen everyday. As a Saints fan, that is almost the equivalent of Brees throwing an incompletion in 2011, when he had a career high 71.2% completion percentage.
he said it was a 70% chance (which is high)
And based on the polling data, what would you have put it at then?
This post was edited on 4/24/17 at 1:08 pm
Posted on 4/24/17 at 1:11 pm to buckeye_vol
quote:I love the way you dress up things that are just common sense. i'm not impressed that given the results in a state ahead of time his model would reflect better probabilities in similar states...no shite.
, but a model can account for it.
man, you're really stretching here. I know what you're saying but practically speaking you're saying "he would've done a better job if he had some of the results beforehand"
but he didn't. he gave the probabilities anyway and the probability of a and b and c is simple multiplication. he gave trump a less than 1% chance to pull those 3 states. that is a matter of irrefutable fact.
Posted on 4/24/17 at 1:18 pm to DelU249
quote:Then why did you make those calculations that didn't include that? You said it his model gave him a 1% chance of winning those 3, which was your criticism, and that was incorrect.
I love the way you dress up things that are just common sense. i'm not impressed that given the results in a state ahead of time his model would reflect better probabilities in similar states...no shite. man, you're really stretching here.
quote:But that's incorrect multiplication, which is what I'm trying telling you. When he wins Wisconsin the 16.5% of the simulations, the conditional probability of Winning Pennsylvania wasn't 23.0%, it was much closer to 100%.
but he didn't. he gave the probabilities anyway and the probability of a and b and c is simple multiplication
quote:No that is a matter of incorrect probability assumptions. You're assuming complete independence, which his model does not. The events (state results) are dependent, therefore your calculations are incorrect.
that is a matter of irrefutable fact.
This post was edited on 4/24/17 at 1:20 pm
Posted on 4/24/17 at 1:24 pm to DelU249
From 10/24/16
Why Our Model Is More Bullish Than Others On Trump
Why Our Model Is More Bullish Than Others On Trump
quote:
State outcomes are highly correlated with one another, so polling errors in one state are likely to be replicated in other, similar states.
quote:
Basically, this means that you shouldn’t count on states to behave independently of one another, especially if they’re demographically similar. If Clinton loses Pennsylvania despite having a big lead in the polls there, for instance, she might also have problems in Michigan, North Carolina and other swing states. What seems like an impregnable firewall in the Electoral College may begin to collapse.
Posted on 4/24/17 at 1:27 pm to buckeye_vol
quote:because I only care what it says BEFORE the results
But that's incorrect multiplication, which is what I'm trying telling you. When he wins Wisconsin the 16.5% of the simulations, the conditional probability of Winning Pennsylvania wasn't 23.0%, it was much closer to 100%.
you don't deny, he's making predictions. once results are in, you don't get to change your predictions. Posted on 4/24/17 at 1:29 pm to buckeye_vol
i don't really need a nate silver disclaimer that if he fricks up Michigan or Wisconsin then he fricked up Pennsylvania too.
i get it, he was the least wrong. problem is, he's the most visible.
gotta pop into a meeting. my favorite buckeye reach ever
really impressive stuff
i get it, he was the least wrong. problem is, he's the most visible.
gotta pop into a meeting. my favorite buckeye reach ever
really impressive stuffPosted on 4/24/17 at 1:30 pm to DelU249
quote:Yes he made predictions, and he didn't change them. You're just using incorrect assumptions to make a criticism.
because I only care what it says BEFORE the results you don't deny, he's making predictions. once results are in, you don't get to change your predictions.
Your criticism is completely invalid in this case.
Posted on 4/24/17 at 1:59 pm to CorporateTiger
quote:
If you listened to their podcasts leading up to the election they actually spent a lot of time talking about how volatile the election was and how Trump's 33% chance to win was very real.
The idea he gave Trump no chance is a weird ex post facto creation:
Wrong, wrong, wrong, wrong, wrong
quote:
If Hillary Clinton wins Nevada, even Nate Silver thinks she has a greater than 90 percent chance of winning the White House.
90%. BTW, who won Nevada?
You don't get to take parts of what Nate says, and completely disregard others. And the only reason Nate gave Trump a bigger chance? Well, I'll let his friends tell you why . . . .
quote:
Nate Silver Is Unskewing Polls — All Of Them — In Trump’s Direction. The vaunted 538 election forecaster is putting his thumb on the scales.
quote:
The problem with FiveThirtyEight’s model is that Silver is “putting his thumb on the scale” to generate a closer race (and thus higher traffic to his site)
Nate was purposefully inflating his data in order to drive traffic. Not because he saw something others didn't. They all had the exact same poll numbers. Except Trafalgar, And Nate didn't even use them
Posted on 4/24/17 at 2:02 pm to Ag Zwin
So you basically don't understand statistics or what Silver does right?
Posted on 4/24/17 at 2:05 pm to BugAC
quote:
Will you say the same thing when the primary is finished? He'll probably get 38%, maybe.
The primary is finished he recieved 48%, so you were just a bit off.
Posted on 4/24/17 at 2:08 pm to montanagator
Damn libs get pissy when their sacred cows get prodded
Posted on 4/24/17 at 2:08 pm to League Champs
quote:Prove it.
Nate was purposefully inflating his data in order to drive traffic.
quote:It's included
They all had the exact same poll numbers. Except Trafalgar
quote:Because Trump had a better chance of winning than the other models were giving him.
And the only reason Nate gave Trump a bigger chance?
Posted on 4/24/17 at 2:10 pm to SirWinston
quote:
Trump literally NEVER LED in a single state poll in WI, MI, or PA that was released after the conventions.
Multiple studies show Comey basically swung the election 4 points because he's a conservative who applied a double standard to investigations.
Posted on 4/24/17 at 2:10 pm to League Champs
quote:
Wrong, wrong, wrong, wrong, wrong
What part of that is wrong? His prediction is really obvious and has been linked in this thread already. As to the discussions, go listen to the podcast. They are all available.
Page 6 of 8
Popular
Back to top
Follow TigerDroppings for LSU Football News