Romney, Obama, Rasmussen

RandFan said:
Yeah, once money is on the line things change. Anyone who has played poker knows that money has a way of focusing one's intellect to more accurately calculate odds of success.
Click to expand...

Indeed. Right now I peg the odds of Obama's re-election squarely at 70.4%... in other words, 538's forecast of the chances of Obama winning Ohio.

Anyone who's ever played backgammon knows what a 70-30 split looks like -- it's the chance of a single stone coming in off the bar against a five-point board. And anyone who's ever played backgammon has been on both sides of that particular split many, many, many times, and has seen strings of 5 or more 70s hit in a row, and 5 or more 30s hit in a row. But in the long run, the 70 comes up more than twice as often as the 30.

Could Romney find a miracle somewhere -- a real October Surprise? Sure, 30% ain't zero. But the trend line for Ohio polling at RCP doesn't indicate that it's likely to happen. This is where understanding how probability works has a way of providing clarity and calm in situations like this (as long as you only bet from your bankroll and not your rent money, of course :D).
 
RandFan said:
Yeah, once money is on the line things change. Anyone who has played poker knows that money has a way of focusing one's intellect to more accurately calculate odds of success.
Click to expand...

Anyone who has played penny-ante poker knows people play differently when there's not much on the line.

Didn't Obama raise insane amounts of money online in '08? Do you suppose all those ones who regularly contributed to his campaign might think it a legitimate 'contribution' (and lots more fun!) to keep the Intrade 'indicating' an Obama win, especially as they know political campaign-watchers in the media are following it? If Obama wins, they get their money back plus some, if Obama loses they don't lose anything they probably wouldn't have just given to his campaign anyway.

One thing I do recall from my peripheral following of the '08 election was that Andrew Sullivan was amongst those who first brought Intrade up and promoted its predictive capabilities....
 
mikedenk said:
I will hand it to you, you've put yourself right out there on the edge. I would say that among posters, only UnlovedRebel, who vowed never to post again on JREF if Romney wins, has invested more of their reputation as a prognosticator on the outcome of this election. There are also people who seem to be throwing money around like Amarillo Slim, but they don't match you for sheer insistence.
Click to expand...

So in other words, nothing important? :D

I won't disclose at this time whether I'm in action on InTrade or not, or the contents of any hypothetical portfolio, or what sort of buying/selling I might be doing -- other than to point out that 'buy long and hold' and 'sell short and wait' are the two simplest approaches to those markets. There are day-traders in there who are in motion constantly; I only wish I had that sort of time in my day. :)

In the end, it's just math, though; see prior message WRT backgammon example. When you've got better than 2 to 1 pointed your way and an effectively even-odds proposition (saying 'I think Obama will win' or 'I think Romney will win'), you take the 2 every time, and you sit back and watch. There's nothing I can actually do to affect the outcome of this election; I live in Georgia, and because of God, Guns, and Gays, it will go Republican for the rest of my life.
 
Last edited:
mikedenk said:
Kudos to Nate Silver as well, who has definitely put himself on the chopping block. It would be quite the "walk of shame" for him to have to slink back to DailyKos after a Romney win.
Click to expand...
I've no idea how that even follows. If I say that something is unlikely to happen and it happens it doesn't mean that I was wrong. If I say something is "impossible" and it happens, well, that's another thing.

When the meteorologist says there is a 30% chance of rain and it rains the meteorologist is just doing his job. Your rhetoric is absurd and hyperbolic.
 
ANTPogo said:
No, he hasn't.
Click to expand...

If Silver still has Obama at 68% or so on election day, and Romney wins, it will certainly cost him some credibility, particularly if the markets have moved in Mitt's direction decisively and early.
 
Brainster said:
If Silver still has Obama at 68% or so on election day, and Romney wins, it will certainly cost him some credibility, particularly if the markets have moved in Mitt's direction decisively and early.
Click to expand...

As I write this, Mittens is down $0.27 at InTrade for today, and at 42.7% total.

ETA: And as I edit this five minutes later, it's changed to down $0.10. :D That's pretty humorous overall that things are so volatile right now, but I think perhaps sticking with closing prices for each day is a more productive way to measure :D
 
Last edited:
Brainster said:
If Silver still has Obama at 68% or so on election day, and Romney wins, it will certainly cost him some credibility,
Click to expand...

No it won't, because Silver isn't predicting the eventual winner, but listing statistical probabilities. He won't lose any more credibility than a baseball statistician loses credibility if this batter ends up striking out on any given at-bat.

particularly if the markets have moved in Mitt's direction decisively and early.
Click to expand...

"The markets" have precisely zip to do with Silver's analysis one way or the other.
 
remirol said:
Did you look at your own chart? :rolleyes:

Plot the trend line between the 11th and 23rd. Same ol', same ol', especially taken in context of my own diagram indicating that InTrade has been hovering around 40% for the past 5+ months.

The spike from "around 40" (normal) to "near 46" between yesterday and today is exactly what the market manipulation article describes -- one Romney loon with a fat wallet and poor impulse control. I realize acceptance of facts isn't your strong suit, but come on, dude, it's right there in black and white in front of you.
Click to expand...

I think Remirol wins the competition of the charts here. A graph without zeroes intentionally magnifies the apparent significance of small changes. Also the shorter time range does much the same.

I buy Remirol's argument (so to speak). The big picture shows a relatively uneventful history.
 
Brainster said:
If Silver still has Obama at 68% or so on election day, and Romney wins, it will certainly cost him some credibility, particularly if the markets have moved in Mitt's direction decisively and early.
Click to expand...

Is this the 68% really means 100% argument?

I suspect even if what you say happens but Silver still comes out correct in most of his electoral college predictions, he would lose very little if any credibility.

Remember, 68% reflects a probability of Obama winning and says nothing about the margin of victory. It's going to be a pretty close election--close enough that Silver could be wrong on just a few key states and Romney would win.

But I think 68% is about right.
 
Brainster said:
If Silver still has Obama at 68% or so on election day, and Romney wins, it will certainly cost him some credibility, particularly if the markets have moved in Mitt's direction decisively and early.
Click to expand...
No. If Silver had Obama at 100% you would have a point. 68% means that Romney could win. I'm not sure why you are having difficulty understanding that 68% != 100%. When silver says 68% it does not mean 100%, do you understand the difference between 68% and 100%? Do you understand statistics and probability? 68% for Obama means Romney "could win". If Silver says Romney "could win" and Romney does win how does that affect Silver. Help me out here because I'm not sure you are getting it.
 
ANTPogo said:
No it won't, because Silver isn't predicting the eventual winner, but listing statistical probabilities. He won't lose any more credibility than a baseball statistician loses credibility if this batter ends up striking out on any given at-bat.
Click to expand...
That's what I was trying to say. I was thinking of it in terms of 3 Card Monte, though. If the house has a 2 in 3 chance of winning, a customer winning doesn't mean that those odds are wrong. (In enough trials, such outcomes are even expected or, if you like, "predicted"!)


"The markets" have precisely zip to do with Silver's analysis one way or the other.
Click to expand...
I didn't realize that. I thought he incorporated prediction markets into his analysis, but I see it isn't so:
http://fivethirtyeight.blogs.nytimes.com/methodology/
 
BTW:

A few weeks ago weather.com had a 10% chance of rain in Lake LA. It rained. Question: Brainster, did the weatherman get it wrong?

Does 10% = impossible?
 
JoeTheJuggler said:
That's what I was trying to say. I was thinking of it in terms of 3 Card Monte, though. If the house has a 2 in 3 chance of winning, a customer winning doesn't mean that those odds are wrong. (In enough trials, such outcomes are even expected or, if you like, "predicted"!)/
Click to expand...
See Law of Large numbers. Hint, Vegas survives on this law. You can beat the house but you are very unlikely to beat the house in the long run.

Those who are not getting it, buy this book Innumeracy: Mathematical Illiteracy and Its Consequences. Buy it now.
 
RandFan said:
No. If Silver had Obama at 100% you would have a point. 68% means that Romney could win. I'm not sure why you are having difficulty understanding that 68% != 100%. When silver says 68% it does not mean 100%, do you understand the difference between 68% and 100%? Do you understand statistics and probability? 68% for Obama means Romney "could win". If Silver says Romney "could win" and Romney does win how does that affect Silver. Help me out here because I'm not sure you are getting it.
Click to expand...

Oh oh, I know, I know! Because it makes Silver 68% wrong... Right? :D
Wait. This wasn't a quiz? :mad: Now I just look silly again.
 
JoeTheJuggler said:
I think Remirol wins the competition of the charts here. A graph without zeroes intentionally magnifies the apparent significance of small changes. Also the shorter time range does much the same. I buy Remirol's argument (so to speak).
Click to expand...
You "buy" it because it's what you want to purchase. There has been an undeniable shift in public sentiment since the first debate. This tells me that there is a significant number of people that haven't made up their minds beyond any persuasion. As long as you aren't looking at intraday data, end of day data from the past few days is a better indicator of current public sentiment than weeks old data. Your market manipulation hopes have been destroyed.
 
JoeTheJuggler said:
That's what I was trying to say. I was thinking of it in terms of 3 Card Monte, though. If the house has a 2 in 3 chance of winning, a customer winning doesn't mean that those odds are wrong. (In enough trials, such outcomes are even expected or, if you like, "predicted"!)



I didn't realize that. I thought he incorporated prediction markets into his analysis, but I see it isn't so:
http://fivethirtyeight.blogs.nytimes.com/methodology/
Click to expand...

Three-card Monty, properly done, the sucker wins only when you want him to, to make him think he can win the game before you go for the score.
 
RandFan said:
See Law of Large numbers. Hint, Vegas survives on this law. You can beat the house but you are very unlikely to beat the house in the long run.

Those who are not getting it, buy this book Innumeracy: Mathematical Illiteracy and Its Consequences. Buy it now.
Click to expand...

On a bit of a tangent: one of my favorite exhibits at the St. Louis Science Center (actually it might not still be there; it's been a while) is an illustration of this. It's got a sort of Pachinko game set up (for Price is Right fans, think Plinko). Marbles are dropped in the top center of a grid of pins. At each pin, each marble has a 50/50 chance of going left or right to the next lower row of pins. At the bottom are a series of collection chutes which collect the marbles in a vertical stack (looking something like a bar graph). When you push the button to start the thing, it carries about 800 marbles to the top and begins releasing them in the center top of the grid one-by-one. Where they will land, can only be described with probabilities. Highest probability is the center chute. Lowest probability are the two "tails" (the chutes at either extreme). Very often, in the first couple of minutes, you don't see much of a pattern. But by the time all 800 marbles have bounced their way to the bottom, you invariably get a very nice normal curve.

As you say, the casino industry relies on this predictability over large number of trials.

To come back to the topic, Silver's 68% probability no more says that Obama must win than does the observation that any given marble has a 68.1% chance of landing in one of the 3 central chutes (what describes, more or less one standard deviation to either side of the expected mean) means that any single marble cannot land outside those 3 chutes or the probability is wrong (or not credible).
 
Last edited:
BenBurch said:
Three-card Monty, properly done, the sucker wins only when you want him to, to make him think he can win the game before you go for the score.
Click to expand...

I was talking about a fair game where the choice was truly a guess--sort of a Platonic ideal, I suppose!

Or, as Randfan said, think of the casino industry where payouts that favor the house (relative to the actual probabilities) are all that's necessary, not cheating.
[ETA: And games/odds can be set--especially with the machines--such that you can pretty well predict the frequency of a win to keep the suckers interested.]

Heck, with the slots, they even brag that the "loosest" slots pay out on average less money than you put into them. As Monty Burns said, "I've discovered the perfect business: people swarm in, empty their pockets, and scuttle off."
 
Last edited:
You must log in or register to reply here.

Back
Top Bottom