Romney on intrade

[SkepticalDrew]

I'm not too familiar with how intrade works or what accounts for their seemingly schizo fluctuations in the last 24 hours, but the new swing state polls released today in Colorado, North Carolina, and Virginia certainly didn't help Romney.

Probably a combination of that and the Indiana thing. Just curious - is anyone familiar with how the intrade percentage works?

 

[JoeTheJuggler]

I went to sign up and they wanted actual money.



Etrade trades actual stocks, do they not?

Yes? You don't think there are investors who are essentially gambling?

I think Intrade is at least as much like investing as it is like making a bet.

You can buy and sell shares, unlike regular betting where you essentially only buy shares once and must sit on them until the outcome.

ETA: I think the University of Iowa prediction markets still use play money, but I'm not certain.

 

[JoeTheJuggler]

Probably a combination of that and the Indiana thing. Just curious - is anyone familiar with how the intrade percentage works?

The price of shares in a proposition (such as Obama to win the election) fluctuate in a market in a range between $0 and $10 per share. The share price is interpreted as a probability. $5 represents 50% probability of that proposition being true. $10 represents 100% probability.

 

[SkepticalDrew]

The price of shares in a proposition (such as Obama to win the election) fluctuate in a market in a range between $0 and $10 per share. The share price is interpreted as a probability. $5 represents 50% probability of that proposition being true. $10 represents 100% probability.

That's... a lot simpler than I imagined. Thanks.

 

[JoeTheJuggler]

Bloody hell, signed up when I saw the price for Obama was around $5.50. But by the time they approved my ID etc the price has jumped to $6.20. Not worth the trouble now, as it's the same as using a gambling site in Australia.

Funny, but a pretty similar thing happened to me a few years back when I decided to buy $100 worth of Ford stock on etrade. I signed up, but learned they had waiting period (I think it was a week?) for really low-price shares. I missed the "bottom" by that waiting period. I think I bought at $3, but the bottom was a little bit lower than that.

I don't claim any shrewd insight though. It was dumb luck. I tripled my money and had to sell (in order to pay my federal taxes year before last).

 

[JoeTheJuggler]

That's... a lot simpler than I imagined. Thanks.

I probably over-simplified things. When the event happens, the market is settled. If the proposition was true, shares are settled at $10. If not, they're settled at $0.

ETA: This is why the buy/sell prices correspond to probability. If you do a typical even money bet, it would be like buying a share at $5. If you win, you get $10, but if you lose, you get nothing. If you buy in at say $7, your potential payoff is less than even money. You get $10 at settlement if you win, but you had to risk $7.

It's complicated, though by the fact that you can sell shares too. So you could take a sure profit earlier, or cut your losses by selling for something rather than waiting to get nothing.

ETA: I also didn't realize that you can short sell: you can sell shares you don't yet own.

You'll probably do better to rely on the real info than my version of it:

http://www.intrade.com/v4/misc/howItWorks/theBasics.jsp

 

[RandFan]

Probably a combination of that and the Indiana thing. Just curious - is anyone familiar with how the intrade percentage works?

The price of shares in a proposition (such as Obama to win the election) fluctuate in a market in a range between $0 and $10 per share. The share price is interpreted as a probability. $5 represents 50% probability of that proposition being true. $10 represents 100% probability.

Thanks.

 

[remirol]

Moved from "Rasmussen" thread since this is largely InTrade related.

This has been happening since the 18th. The article that keeps getting getting cited talks about spikes that last for minutes. When those types of spikes occur, people jump on them and the market self corrects. Something that lasts for days, like this is not a market manipulation issue:

Did you look at your own chart?

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.

Well, yes... see... that's how trend lines work, and how more reality-based people avoid getting over-excited about short-term fluctuations that don't actually mean anything

Well, if you think it isn't just a short-term fluctuation, there's plenty of time for you to get in the markets, or in the betting thread. I think Matt and Ben have been looking for someone to put money down on Romney's side for awhile now -- be sure to bet big!

Bumping this so there's a more appropriate place for the InTrade discussion after today's closing price gets posted.

Still waiting to hear how this isn't an expected and predicted regression to the mean -- aka "around 40", and is instead somehow an indicator of Romney's upward momentum.

 

[JoeTheJuggler]

The overall trend looks to me like "around 38" except for two big upward spikes and two equally short-lived troughs--at least based on what is shown in this short a time span with a scale that exaggerates very slight changes around 40%.

If we saw the entire 0 to 100 scale, these spikes and troughs would look like tiny jig jags.

 

[JoeTheJuggler]

By the way, the green bars representing "volume"--that's just number of shares traded, right? Not number of transactions or number of traders?

 

[remirol]

Nothing has been happening, really. Romney's price has been largely consistent since right about mid-May, except for a sharp downward spike after the 47% video was released, and a rebound after the first debate. And it's sitting there hovering right around 40-ish.

The overall trend looks to me like "around 38" except for two big upward spikes and two equally short-lived troughs--at least based on what is shown in this short a time span with a scale that exaggerates very slight changes around 40%.

If we saw the entire 0 to 100 scale, these spikes and troughs would look like tiny jig jags.

This is where I'm getting the "around 40" from; back in the other thread, the first thing I did was pull the entire set of data and take a look at the overall trend since about mid-May or so -- the point at which most other Republican primary candidates had effectively conceded the race to Romney by stopping spending. Since then, he's ricocheted between 37.5 (ish) and 42.5 (ish), excluding the huge post-47% crash and post-debate rebound.

And yes, I believe the green bars simply represent number of shares traded.

 

[Meadmaker]

So it's a gambling website....

Yes.

There's an element of skill involved, though. It's not like a random number generator. If you understand politics, and are not biased by wishful thinking for your own team, you are more likely to make the correct prediction about the outcome, and thus more likely to make money buying and selling "shares" in Romney.

It's very much like sports betting. You use your knowledge of the sport to predict outcomes. If you are good at it, you win more often than you lose, and make money.

 

[Ampulla of Vater]

How is it affected if I were to buy 1000 shares of Romney (so to speak) just for the heck of it? Doesn't that then skew what you are relying on as some type of accurate indicator of what is really happening with the election probability?

 

[remirol]

How is it affected if I were to buy 1000 shares of Romney (so to speak) just for the heck of it? Doesn't that then skew what you are relying on as some type of accurate indicator of what is really happening with the election probability?

You can only buy shares that are being sold by someone, and the price you purchase them at is what matters (though volume is a good indicator of how 'serious' anyone is).

So, as of _right now_, Romney is trading at $3.70 a share. If you follow the link you'll see a tab for "unmatched predictions". What we see here is that people are willing to sell shares of Romney at:

Best (lowest) price members are selling at
Price per share Quantity
$3.70 36 shares
$3.71 105 shares
$3.73 36 shares
$3.80 5 shares
$3.82 16 shares
$3.83 1 share
$3.84 100 shares
$3.85 69 shares
$3.86 100 shares
$3.87 604 shares
$3.88 200 shares
$3.89 700 shares$3.90 401 shares
$3.91 2 shares
$3.92 1 share

So if you wanted to buy those 700 shares of Romney at $3.89, sure, that would change the last-traded price and Romney would be +0.9 on the day instead of -0.10. But you'd be an outlier, and while some people would come along with you, it wouldn't significantly change the general opinion that Romney is worth around $3.70 right now. The only difference is that you'd have about $2700 in action.

If Romney stayed at $3.70 EOD, then you'd be looking at (in a few short hours) having lost $133 -- assuming you can find someone to buy the 700 shares of Romney back at $3.70. If you look at who's _buying_, you get an entirely different picture...

$3.66 10 shares
$3.65 100 shares
$3.63 125 shares
$3.60 100 shares
$3.59 709 shares
$3.58 52 shares
$3.57 400 shares
$3.56 900 shares
$3.55 510 shares
$3.53 70 shares
$3.51 101 shares
$3.50 1581 shares
$3.46 10 shares
$3.43 20 shares
$3.41 1 share

...and you see that the highest anyone is willing to buy Romney shares for is $3.66. So really, it's even worse; you're closer to being out $150 simply by clicking the button on "Buy at $3.89".

Now think about what happens to your money if Romney gets caught eating a baby on live TV and his price drops to $2.50 a share... and you see why, although volatile and subject to longshot bias, a prediction market does tend to have _reasonably_ accurate data.

 

[Ampulla of Vater]

You can only buy shares that are being sold by someone, and the price you purchase them at is what matters (though volume is a good indicator of how 'serious' anyone is).

So, as of _right now_, Romney is trading at $3.70 a share. If you follow the link you'll see a tab for "unmatched predictions". What we see here is that people are willing to sell shares of Romney at:



So if you wanted to buy those 700 shares of Romney at $3.89, sure, that would change the last-traded price and Romney would be +0.9 on the day instead of -0.10. But you'd be an outlier, and while some people would come along with you, it wouldn't significantly change the general opinion that Romney is worth around $3.70 right now. The only difference is that you'd have about $2700 in action.

If Romney stayed at $3.70 EOD, then you'd be looking at (in a few short hours) having lost $133 -- assuming you can find someone to buy the 700 shares of Romney back at $3.70. If you look at who's _buying_, you get an entirely different picture...



...and you see that the highest anyone is willing to buy Romney shares for is $3.66. So really, it's even worse; you're closer to being out $150 simply by clicking the button on "Buy at $3.89".

Now think about what happens to your money if Romney gets caught eating a baby on live TV and his price drops to $2.50 a share... and you see why, although volatile and subject to longshot bias, a prediction market does tend to have _reasonably_ accurate data.

Thanks, that makes sense. Though I would rather fantasize about what would happen to my money if Obama got caught eating a baby on live TV.

 

[remirol]

Thanks, that makes sense. Though I would rather fantasize about what would happen to my money if Obama got caught eating a baby on live TV.

As would everyone in that hypothetical position (holding +700 of Romney/$3.89)... which is exactly where longshot bias comes from

It's just much easier for humans to think about risking a little bit of money to win a lot, rather than risking a lot to win a little.

 

[remirol]

Still waiting to hear how this isn't an expected and predicted regression to the mean -- aka "around 40", and is instead somehow an indicator of Romney's upward momentum.

Chart including Friday's closing price. I've included one of the handy InTrade tools you can use -- a simple moving average covering 7 days' worth of data, which is the brown line (look in the bottom left). I've also extended the chart out to 21 days so that the average itself gives a better picture of what's been happening: ie. nothing special.

Neally? Any thoughts on this? For your edification, a moving average is one of the ways you can more accurately assess what a particular trend looks like, rather than focusing on a particular cherry-picked range. And as we can see, the trend for Mitt is... well, literally nothing special. He goes up, he goes down, he hovers around 40 (well, actually, it _is_ 38 in this particular date range, but I'm willing to be generous as an Obama supporter).

It will be interesting to see if the people who bought Romney and/or sold short on Obama start a market run at some point, and if so when. I would generally say November 5th would be the earliest possible time that would happen (desperation holdout) but we might see rats jumping off the ship earlier to try to get in while the price is still tolerable.

 

[not_so_new]

Thanks, that makes sense. Though I would rather fantasize about what would happen to my money if Obama got caught eating a baby on live TV.

Which is exactly why prediction markets work. While we may wish for one outcome or another, we are often reluctant to place actual money on something that we don't have a certain level of confidence in. Everyone has a risk tolerance and that risk tolerance trumps emotional whims.

 

[remirol]

Which is exactly why prediction markets work. While we may wish for one outcome or another, we are often reluctant to place actual money on something that we don't have a certain level of confidence in. Everyone has a risk tolerance and that risk tolerance trumps emotional whims.

...At least sometimes. There's a miniwhale in the water over there again right now:

Best (highest) price members are buying at
Price per share Quantity
$3.64 16 shares (highest)
...
$3.60 2026 shares ($7,300)
...
$3.50 1567 shares ($5,500)

It's entirely possible that this is someone who sold Romney short at $4.54 and is now trying to make a quick couple of grand, of course, with the intent of immediately re-selling him short at $3.70. But it also could be someone who's decided "damn the statistics, full speed ahead".

 

[JoeTheJuggler]

But it also could be someone who's decided "damn the statistics, full speed ahead".

Yeah, it's hard to tell. With the greater press Intrade has been getting this election, it could even be someone trying to manipulate the market in lieu of making a similarly large campaign donation.