Nate Silver of fivethirtyeight became famous as he once predicted all 50 states in a presidential election right. ESPN gave him a ton of money but it now seems that the site is never right.
I was looking at it this morning.
He says there is a 46% chance the GOP takes the Senate. He says there is an 82% chance they win the House.
So what are the odds that the GOP also takes the House and the Senate?
He says it's also 46%. WHAT?
That makes no sense. I know the two aren't totally independent events but still that's crazy. It would have to be less than 46%.
The way I read that is he thinks there's a 100% chance the GOP takes the hosue but won't admit it.
If there is an 18% chance the Dems win the House and a 46% chance the GOP wins the senate, does it make any sense at all that the possibility of both happening is 0%?
You can think of it like this. Silver created a model that simulates what might happen in the real world. It's based on polling data, but also includes some random noise to reflect the fact that you never really know what's going to happen. As a result, every time you run the model you get a different outcome. So you run the model over and over and see which outcomes happen most often. This is called Monte Carlo simulation.
The numbers we're arguing about are basically the equivalent of what happens when Silver runs the model 100 times. He gets something like this:
- Dems win House and Senate: 18 times
- Dems win House, Reps win Senate: 0 times
- Reps win House, Dems win Senate: 36 times
- Reps win House and Senate: 46 times
Now, if you look at those numbers, you can ask: in how many of those runs did the Dems win the House? It's 18 + 0 = 18.
In how many did the Reps with the Senate? It's 46 + 0 = 46.
In how many of those runs did Dems win the House and Reps win the Senate? 0. There's no contradiction here. We're just adding up who won in each of the 100 runs.
If you read the original figure that Rojo posted, you'll notice it doesn't say "There's a 46 percent chance that the Republicans win the Senate." It says that Republicans win the Senate in 46 out of every 100 runs of their statistical model.
If there is an 18% chance the Dems win the House and a 46% chance the GOP wins the senate, does it make any sense at all that the possibility of both happening is 0%?
You can think of it like this. Silver created a model that simulates what might happen in the real world. It's based on polling data, but also includes some random noise to reflect the fact that you never really know what's going to happen. As a result, every time you run the model you get a different outcome. So you run the model over and over and see which outcomes happen most often. This is called Monte Carlo simulation.
The numbers we're arguing about are basically the equivalent of what happens when Silver runs the model 100 times. He gets something like this:
- Dems win House and Senate: 18 times
- Dems win House, Reps win Senate: 0 times
- Reps win House, Dems win Senate: 36 times
- Reps win House and Senate: 46 times
Now, if you look at those numbers, you can ask: in how many of those runs did the Dems win the House? It's 18 + 0 = 18.
In how many did the Reps with the Senate? It's 46 + 0 = 46.
In how many of those runs did Dems win the House and Reps win the Senate? 0. There's no contradiction here. We're just adding up who won in each of the 100 runs.
If you read the original figure that Rojo posted, you'll notice it doesn't say "There's a 46 percent chance that the Republicans win the Senate." It says that Republicans win the Senate in 46 out of every 100 runs of their statistical model.
Seriously? Do you really expect him to change his mind just because you put in the effort to patiently explain 538’s simulation process?
If there is an 18% chance the Dems win the House and a 46% chance the GOP wins the senate, does it make any sense at all that the possibility of both happening is 0%?
You can think of it like this. Silver created a model that simulates what might happen in the real world. It's based on polling data, but also includes some random noise to reflect the fact that you never really know what's going to happen. As a result, every time you run the model you get a different outcome. So you run the model over and over and see which outcomes happen most often. This is called Monte Carlo simulation.
The numbers we're arguing about are basically the equivalent of what happens when Silver runs the model 100 times. He gets something like this:
- Dems win House and Senate: 18 times
- Dems win House, Reps win Senate: 0 times
- Reps win House, Dems win Senate: 36 times
- Reps win House and Senate: 46 times
Now, if you look at those numbers, you can ask: in how many of those runs did the Dems win the House? It's 18 + 0 = 18.
In how many did the Reps with the Senate? It's 46 + 0 = 46.
In how many of those runs did Dems win the House and Reps win the Senate? 0. There's no contradiction here. We're just adding up who won in each of the 100 runs.
If you read the original figure that Rojo posted, you'll notice it doesn't say "There's a 46 percent chance that the Republicans win the Senate." It says that Republicans win the Senate in 46 out of every 100 runs of their statistical model.
Silver is a fraud. If he's not willing to put his reputation on the line based on his stats, then he is the definition of a FRAUD. He should be confident enough in his work to say there is a 100% chance of this result happening.
If he thinks the dems win 54 out of 100 simulations, then his prediction is the dems win. He should come out and say there is a 100% chance the dems will win, but he doesn't have the balls.
Silver is a fraud. If he's not willing to put his reputation on the line based on his stats, then he is the definition of a FRAUD. He should be confident enough in his work to say there is a 100% chance of this result happening.
If he thinks the dems win 54 out of 100 simulations, then his prediction is the dems win. He should come out and say there is a 100% chance the dems will win, but he doesn't have the balls.
I challenge Silver or any of the others (Cook, etc.) to make a call on Monday November 7. Tell us the number of seats that will be won by each party in both the House and the Senate.
They are all frauds who got 2016 completely wrong and have been trying to hedge their way back to some semblance of credibility ever since.
If there is an 18% chance the Dems win the House and a 46% chance the GOP wins the senate, does it make any sense at all that the possibility of both happening is 0%?
You can think of it like this. Silver created a model that simulates what might happen in the real world. It's based on polling data, but also includes some random noise to reflect the fact that you never really know what's going to happen. As a result, every time you run the model you get a different outcome. So you run the model over and over and see which outcomes happen most often. This is called Monte Carlo simulation.
The numbers we're arguing about are basically the equivalent of what happens when Silver runs the model 100 times. He gets something like this:
- Dems win House and Senate: 18 times
- Dems win House, Reps win Senate: 0 times
- Reps win House, Dems win Senate: 36 times
- Reps win House and Senate: 46 times
Now, if you look at those numbers, you can ask: in how many of those runs did the Dems win the House? It's 18 + 0 = 18.
In how many did the Reps with the Senate? It's 46 + 0 = 46.
In how many of those runs did Dems win the House and Reps win the Senate? 0. There's no contradiction here. We're just adding up who won in each of the 100 runs.
If you read the original figure that Rojo posted, you'll notice it doesn't say "There's a 46 percent chance that the Republicans win the Senate." It says that Republicans win the Senate in 46 out of every 100 runs of their statistical model.
1. You just made up the numbers.
2. "..it doesn't say "There's a 46 percent chance that the Republicans win the Senate." It says that Republicans win the Senate in 46 out of every 100 runs of their statistical model." It's the same damn thing
If there is an 18% chance the Dems win the House and a 46% chance the GOP wins the senate, does it make any sense at all that the possibility of both happening is 0%?
You can think of it like this. Silver created a model that simulates what might happen in the real world. It's based on polling data, but also includes some random noise to reflect the fact that you never really know what's going to happen. As a result, every time you run the model you get a different outcome. So you run the model over and over and see which outcomes happen most often. This is called Monte Carlo simulation.
The numbers we're arguing about are basically the equivalent of what happens when Silver runs the model 100 times. He gets something like this:
- Dems win House and Senate: 18 times
- Dems win House, Reps win Senate: 0 times
- Reps win House, Dems win Senate: 36 times
- Reps win House and Senate: 46 times
Now, if you look at those numbers, you can ask: in how many of those runs did the Dems win the House? It's 18 + 0 = 18.
In how many did the Reps with the Senate? It's 46 + 0 = 46.
In how many of those runs did Dems win the House and Reps win the Senate? 0. There's no contradiction here. We're just adding up who won in each of the 100 runs.
If you read the original figure that Rojo posted, you'll notice it doesn't say "There's a 46 percent chance that the Republicans win the Senate." It says that Republicans win the Senate in 46 out of every 100 runs of their statistical model.
Silver gave us the 18% and 46% odds.
How do you arrive at your numbers?
Simple. You made them up. To fit the conclusion you wish to reach.
Or perhaps you can show the calculations you used to get these numbers and prove me wrong? You know, with math instead of made up numbers and gobbledy gook
Nate Silver of fivethirtyeight became famous as he once predicted all 50 states in a presidential election right. ESPN gave him a ton of money but it now seems that the site is never right.
I was looking at it this morning.
He says there is a 46% chance the GOP takes the Senate. He says there is an 82% chance they win the House.
So what are the odds that the GOP also takes the House and the Senate?
He says it's also 46%. WHAT?
That makes no sense. I know the two aren't totally independent events but still that's crazy. It would have to be less than 46%.
The way I read that is he thinks there's a 100% chance the GOP takes the hosue but won't admit it.
Hi rojo, I'm a mathematician, and think it is very likely that both you and Nate are right. Let me explain. Lets assume the numbers for his two events are correct, and we want to look at the combined event. That is the chance the republicans win the senate is 46% and the house 82%. What Nate's model is saying is that the chance republicans win the senate but lose the house is 0%. This actually seems quite reasonable. If republicans lost the house, meaning the democrats were winning way more local races than expected, like tons more, because it was unlikely for them to win the house, the chance that democrats lose more senate races than expected is close to zero.
Why are you both right then. Well I suspect it isn't exactly zero. It is probably 0 point something, and Nate is just presenting round numbers for convenience. So you could be right in that respect.
Similarly, does his model put the democrats chance of winning both at around 18%? If so, it is the same thing going on, there is an 18% chance of them taking the house, but if they took the house, they will almost surely take the easier senate, so the senate similarly doesn't affect the calculation. When events are highly correlated as they are here, winning both is mostly if not entirely affected by the harder win.
As far as not predicting the first trump election, his models had trump winning with something like a 30% probability, so he was closer than most of the main pundentsm, who had trumps chances closer to 1%. A 30% chance of winning is like 1 in 3 times, so when Trump won its hard to say Nate was wrong. I have a few issues with things Nate does, but he does a much better job at data sci models than most data scientists.
Nate Silver is like Malcom Gladwell. Neither are stupid. Neither are real intellectuals. Both are pseudo-intellectuals that used some fortunate episodes in their lives - paired with great personal branding - to build an aurora of genius/high level expertise.
Nate Silver of fivethirtyeight became famous as he once predicted all 50 states in a presidential election right. ESPN gave him a ton of money but it now seems that the site is never right.
I was looking at it this morning.
He says there is a 46% chance the GOP takes the Senate. He says there is an 82% chance they win the House.
So what are the odds that the GOP also takes the House and the Senate?
He says it's also 46%. WHAT?
That makes no sense. I know the two aren't totally independent events but still that's crazy. It would have to be less than 46%.
The way I read that is he thinks there's a 100% chance the GOP takes the hosue but won't admit it.
Hi rojo, I'm a mathematician, and think it is very likely that both you and Nate are right. Let me explain. Lets assume the numbers for his two events are correct, and we want to look at the combined event. That is the chance the republicans win the senate is 46% and the house 82%. What Nate's model is saying is that the chance republicans win the senate but lose the house is 0%. This actually seems quite reasonable. If republicans lost the house, meaning the democrats were winning way more local races than expected, like tons more, because it was unlikely for them to win the house, the chance that democrats lose more senate races than expected is close to zero.
Why are you both right then. Well I suspect it isn't exactly zero. It is probably 0 point something, and Nate is just presenting round numbers for convenience. So you could be right in that respect.
Similarly, does his model put the democrats chance of winning both at around 18%? If so, it is the same thing going on, there is an 18% chance of them taking the house, but if they took the house, they will almost surely take the easier senate, so the senate similarly doesn't affect the calculation. When events are highly correlated as they are here, winning both is mostly if not entirely affected by the harder win.
As far as not predicting the first trump election, his models had trump winning with something like a 30% probability, so he was closer than most of the main pundentsm, who had trumps chances closer to 1%. A 30% chance of winning is like 1 in 3 times, so when Trump won its hard to say Nate was wrong. I have a few issues with things Nate does, but he does a much better job at data sci models than most data scientists.
Silver is such a hack. He claimed he predicted Obama's win perfectly, but he actually didn't do anything. He just averaged out polls that other people did. There's nothing scientific about that.
If he was so smart, he would have been able to sense that the polls in 2016 were obviously heavier biased towards Clinton.
Nate Silver is like Malcom Gladwell. Neither are stupid. Neither are real intellectuals. Both are pseudo-intellectuals that used some fortunate episodes in their lives - paired with great personal branding - to build an aurora of genius/high level expertise.
The two are nothing alike. Silver works with data and runs a business developing data-driven predictive models. He doesn’t claim to be a genius. Nor does Gladwell who writes books for entertainment.
As far as not predicting the first trump election, his models had trump winning with something like a 30% probability, so he was closer than most of the main pundentsm, who had trumps chances closer to 1%. A 30% chance of winning is like 1 in 3 times, so when Trump won its hard to say Nate was wrong. I have a few issues with things Nate does, but he does a much better job at data sci models than most data scientists.
What threshold do the so called "forecasters" need to be below to make it easy to say they're wrong ? Trump at 30% means Silver's not wrong when he had Clinton at 70% ? That's BS.
I predicted a Trump victory in 2016. 5 states wrong, but I had it at Trump 275 electoral votes. Far below the final total, but superior to a charlatan like Silver.
Yes, you're smarter than both Nate Silver and Malcolm Gladwell. Take the rest of the day off!
The probability Malcolm Gladwell is smart is approaching zero. The probability someone who thinks he is smart has even a erage intelligence is also very low. He uses big words and sounds intelligent to unintelligent people but he puts nothing original out 9ntobtje universe that is a good idea.
Nate Silver of fivethirtyeight became famous as he once predicted all 50 states in a presidential election right. ESPN gave him a ton of money but it now seems that the site is never right.
I was looking at it this morning.
He says there is a 46% chance the GOP takes the Senate. He says there is an 82% chance they win the House.
So what are the odds that the GOP also takes the House and the Senate?
He says it's also 46%. WHAT?
That makes no sense. I know the two aren't totally independent events but still that's crazy. It would have to be less than 46%.
The way I read that is he thinks there's a 100% chance the GOP takes the hosue but won't admit it.
Hi rojo, I'm a mathematician, and think it is very likely that both you and Nate are right. Let me explain. Lets assume the numbers for his two events are correct, and we want to look at the combined event. That is the chance the republicans win the senate is 46% and the house 82%. What Nate's model is saying is that the chance republicans win the senate but lose the house is 0%. This actually seems quite reasonable. If republicans lost the house, meaning the democrats were winning way more local races than expected, like tons more, because it was unlikely for them to win the house, the chance that democrats lose more senate races than expected is close to zero.
Why are you both right then. Well I suspect it isn't exactly zero. It is probably 0 point something, and Nate is just presenting round numbers for convenience. So you could be right in that respect.
Similarly, does his model put the democrats chance of winning both at around 18%? If so, it is the same thing going on, there is an 18% chance of them taking the house, but if they took the house, they will almost surely take the easier senate, so the senate similarly doesn't affect the calculation. When events are highly correlated as they are here, winning both is mostly if not entirely affected by the harder win.
As far as not predicting the first trump election, his models had trump winning with something like a 30% probability, so he was closer than most of the main pundentsm, who had trumps chances closer to 1%. A 30% chance of winning is like 1 in 3 times, so when Trump won its hard to say Nate was wrong. I have a few issues with things Nate does, but he does a much better job at data sci models than most data scientists.
As far as not predicting the first trump election, his models had trump winning with something like a 30% probability, so he was closer than most of the main pundentsm, who had trumps chances closer to 1%. A 30% chance of winning is like 1 in 3 times, so when Trump won its hard to say Nate was wrong. I have a few issues with things Nate does, but he does a much better job at data sci models than most data scientists.
What threshold do the so called "forecasters" need to be below to make it easy to say they're wrong ? Trump at 30% means Silver's not wrong when he had Clinton at 70% ? That's BS.
I predicted a Trump victory in 2016. 5 states wrong, but I had it at Trump 275 electoral votes. Far below the final total, but superior to a charlatan like Silver.
One way is to think how different his predictions were than others. All other major newspapers had Clinton at 90% or greater. So he predicted Trump was 3 times more likely to win than all major papers. So he was more correct than most other modelers. What was the reason why he favored Trump more than others, correlations between states. His models factored in the chance that if the polls are wrong in one state they are more likely to be wrong in other states in the same direction. I'm not saying Nate is amazing, he's just factoring in basic probability. It's not like he is a magician or a super predictor or something.
Silver is a farce. Take a look at his most recent article where he has a full conversation with himself pretending to be a republican. I don't know what's more pathetic - the way he imagines republicans talk or the way SILVER actually talks himself.
Republicans reached a milestone on Tuesday, surpassing a 50 percent chance of a Senate takeover in our Deluxe forecast for the first time since July 25. Still, …
Silver is a farce. Take a look at his most recent article where he has a full conversation with himself pretending to be a republican. I don't know what's more pathetic - the way he imagines republicans talk or the way SILVER actually talks himself.
I don't know about this Silver guy, but I can state for sure that nobody thinks Rojo is smart.
Yes, a thread title about Rojo being smart would be seen as satire. Based on his writing as well as his opinions and analysis, I generously assign him an estimated IQ of 91. Some point to his 'success', but realistically that has been quite modest for someone born on third base.
