Romney: 47% of Americans are losers, don't care about 'em

[alfred russel]

I thought I was giving Yi a ridiculous discount.  At this point, I think what holds Romney's price above 0% is a chance of some kind of disaster for Obama, like finally admitting that he's a Muslim, or the effect of voter suppression in some key states.

What if he just gets his ass kicked in the debates?

The polls have been consistently close--over a month away I don't think a 4% lead or whatever it is should be considered all that secure. I think the tendancy is to give Romney less chance than the polling would normally indicate because he has shown no life as a candidate. Which I think is fair--if he is a crap candidate for the first years of his campaign, why should we expect a turnaround in the last month?

[DGuller]

I thought I was giving Yi a ridiculous discount.  At this point, I think what holds Romney's price above 0% is a chance of some kind of disaster for Obama, like finally admitting that he's a Muslim, or the effect of voter suppression in some key states.

What if he just gets his ass kicked in the debates?

The polls have been consistently close--over a month away I don't think a 4% lead or whatever it is should be considered all that secure. I think the tendancy is to give Romney less chance than the polling would normally indicate because he has shown no life as a candidate. Which I think is fair--if he is a crap candidate for the first years of his campaign, why should we expect a turnaround in the last month?

I don't think the polls have been that close.  Going by 538, the "now-cast" is 98% chance of Obama victory.  National polls or poll aggregations are meaningless, since most of the nation is meaningless when it comes to voting.

[Valmy]

since most of the nation is meaningless when it comes to voting.

Like me!  People in central Asia who have never heard of the US have roughly an equal impact in this election as I do.

[PDH]

since most of the nation is meaningless when it comes to voting.

Like me!  People in central Asia who have never heard of the US have roughly an equal impact in this election as I do.

Wyoming is like Texas only with seasons.

[Eddie Teach]

Those being Winter and Winter+

[alfred russel]

I don't think the polls have been that close.  Going by 538, the "now-cast" is 98% chance of Obama victory.  National polls or poll aggregations are meaningless, since most of the nation is meaningless when it comes to voting.

I don't get why you like the 538 blog so much.

The national polls would be meaningless, except the electoral vote does actually match the national polling (in terms of the winner at least). Yeah, Al Gore won the popular vote while losing (or "losing") the electoral vote, but the popular vote was about 50-50. The national polls have Obama up by about 4%, and he is up by about that much in most of the swing states. If there was a 5% swing to Romney nationally and included the swing states, Obama would be in real trouble. The national polls are also the most reliable we have (and also easier to discuss than the myriad of possibilities in the state polling).

[merithyn]

since most of the nation is meaningless when it comes to voting.

Like me!  People in central Asia who have never heard of the US have roughly an equal impact in this election as I do.

Same. I'll be voting for the Green Party just because it won't matter in the grand scheme, and I like supporting the smaller parties. Obama has Illinois locked up, so why not?

[alfred russel]

Same. I'll be voting for the Green Party just because it won't matter in the grand scheme, and I like supporting the smaller parties. Obama has Illinois locked up, so why not?

Because in a contested election the popular vote can have a persuasive effect. Al Gore may not have been able to contest Florida to the end if the popular vote didn't go his way. Also, the margin of victory can have an effect on the magnitude and length of a "honeymoon" period for a new president.

[DGuller]

I don't get why you like the 538 blog so much.

I like it because it is a statistically intelligent aggregator of poll results.  Being a professional statistician of sorts, I am aware of the need to adjust the data in multiple ways before even beginning to analyze it, and that's something that only 538 understands.  Others just follow one or two tracking polls, or a mindless average like RCP, which gives a shitload of noise that they misinterpret as signal.  Another reason I like it is because it understand that elections are won state by state, and thus all its analysis is done as a sum of 57 state results.

The national polls would be meaningless, except the electoral vote does actually match the national polling (in terms of the winner at least). Yeah, Al Gore won the popular vote while losing (or "losing") the electoral vote, but the popular vote was about 50-50. The national polls have Obama up by about 4%, and he is up by about that much in most of the swing states. If there was a 5% swing to Romney nationally and included the swing states, Obama would be in real trouble. The national polls are also the most reliable we have (and also easier to discuss than the myriad of possibilities in the state polling).

Again, the problem here is noise, and lack of precision.  Tracking polls seem to reverse themselves far more frequently than is warranted.

[alfred russel]

DGuller, I read 538 briefly after the Ryan selection, and was rather disappointed in the analysis. The basic concept was that Romney's boost on Intrade after the selection was unwarranted (from ~35% -> ~40%) because of historical poll data showing larger bounces were the norm, and these faded. All of which is a fine argument, but it doesn't address the possibility that Romney's choices were more  constrained than normal and coming off of the Palin fiasco was going to result in greater scrutiny--and these unique factors were priced into intrade. Thus I think there is a good case to be made that Romney making a choice that made the base happy while moderately boosting his poll numbers was a net positive for his election chances.

[DGuller]

What do you think would be a more correct approach:  adjust the prediction for the anticipated short term nature of the bounce, and run the risk of over- or under-adjusting it, or take the bounce at face value as the new reality of the race?

[alfred russel]

What do you think would be a more correct approach:  adjust the prediction for the anticipated short term nature of the bounce, and run the risk of over- or under-adjusting it, or take the bounce at face value as the new reality of the race?

Obviously you need to adjust for the short term nature of the bounce, but if you are really going to handicap the race you need to do more than use current and historical data. You also need to reflect the historical trends against he current reality to project the degree to which the historical trends will be followed. This is something a market theoretically does well, so it was annoying to read a critique of a market based largely on historical and current data.

[DGuller]

When it comes to data adjustments, the perfect is the enemy of the good.  You can always refine your adjustments, and I'm sure 538 does plenty of that behind the scenes, but even an imperfect adjustment is better than just taking raw numbers and going with them.

[crazy canuck]

since most of the nation is meaningless when it comes to voting.

Like me!  People in central Asia who have never heard of the US have roughly an equal impact in this election as I do.

Same. I'll be voting for the Green Party just because it won't matter in the grand scheme, and I like supporting the smaller parties. Obama has Illinois locked up, so why not?

Because you are electing the most powerful person on the planet and if enough people think the way you do Romney will be put into that position.

I dont know about you but that possibility would make me vote for Obama if I could.

[merithyn]

Same. I'll be voting for the Green Party just because it won't matter in the grand scheme, and I like supporting the smaller parties. Obama has Illinois locked up, so why not?

Because in a contested election the popular vote can have a persuasive effect. Al Gore may not have been able to contest Florida to the end if the popular vote didn't go his way. Also, the margin of victory can have an effect on the magnitude and length of a "honeymoon" period for a new president.

I'm not sure you understand where I live. Illinois, at the best of times, goes Democrat, and now we're the incumbant's home state. My vote is completely irrelevant in this state. I expect it to be 65/35 Obama, and that's pretty conservative. I wouldn't be the least surprised to see it go 70/30.

And given how loud the extreme views of politics in this country are, no matter who wins, there will be no "honeymoon" period after this election. The extremists will make sure of it. So if I cast my vote for a third party, I'm casting my vote against the extremists in the two major parties - and against the two-party system - which makes me happy. Quite frankly, it's the only way I feel like my vote will matter in the least.