Arbitrage opportunity on the presidential election

Currently Pinnacle has the following lines:

Obama: -322
Romney: +290

At Intrade you can buy Obama at 0.68, and sell at 0.675, which equates to these lines:

Obama: -212.5
Romney: +207.7

If you had an account at both places you could get Romney at +290 at Pinnacle and Obama at -212.5 at Intrade and have a HUGE middle. I've never seen such a big arbitrage opportunity before. Too bad I don't have an account at both places.

How could you make money on both candidates? The election is not a contest with a score that can win both sides, so I assume the books are putting the majority of their eggs in the Obama basket, and hoping to collect more from the Romney backers. If/when Romney loses, you may have hedged your Obama wager, but it will still be a loser. Isn't it better to pick one mount in this two horse race?

Quote: Ayecarumba

How could you make money on both candidates? The election is not a contest with a score that can win both sides, so I assume the books are putting the majority of their eggs in the Obama basket, and hoping to collect more from the Romney backers. If/when Romney loses, you may have hedged your Obama wager, but it will still be a loser. Isn't it better to pick one mount in this two horse race?

Bet Obama at -212.5
Bet Romney at +290
You win either way.

Right now, I'm taking action at -300/+250.
I wrote a few bets on Romney +170 awhile back.

Quote: Ayecarumba

How could you make money on both candidates?

Bet $1000 on Romney at Pinnacle and Lay $2650 on Obama at Intrade. Here is your net win under both outcomes:

Obama wins: -1000 + 2650/2.125 = +247.06
Romney wins: +2900 - 2650 = +250

Quote: Wizard

Bet $1000 on Romney at Pinnacle and Lay $2650 on Obama at Intrade. Here is your net win under both outcomes:

Obama wins: -1000 + 2650/2.125 = +247.06
Romney wins: +2900 - 2650 = +250

Is the fact that you are guaranteed (apologies to Ron Paul) a small win different than the "Thou Shalt Not Hedge" commandment #7? If you believe in one or the other to win, why not shoot the moon on the one, and collect $1,717.65 on Obama or $10,585 on Romney with the same $3,650 action?

Quote: Ayecarumba

Is the fact that you are guaranteed (apologies to Ron Paul) a small win different than the "Thou Shalt Not Hedge" commandment #7? If you believe in one or the other to win, why not shoot the moon on the one, and collect $1,717.65 on Obama or $10,585 on Romney with the same $3,650 action?

In the words of the late Henry Gordon " I don't want nothing to do with an even bet "

Quote: Ayecarumba

Is the fact that you are guaranteed (apologies to Ron Paul) a small win different than the "Thou Shalt Not Hedge" commandment #7?

Yes. Arbitrage > Hedging.

Quote: Ayecarumba

Is the fact that you are guaranteed (apologies to Ron Paul) a small win different than the "Thou Shalt Not Hedge" commandment #7? If you believe in one or the other to win, why not shoot the moon on the one, and collect $1,717.65 on Obama or $10,585 on Romney with the same $3,650 action?

This is not a hedge, it is arbitrage. You lock in a win the minute it is all placed. A hedge would be if you believed in Obama but bet some on Mitt as the polls for Obama fell, like they have recently from 50-47.

One guarantees a win, the other is minimizing possible loss.

Pretty sure "Thou Shalt Not Hedge" was written with table games in mind, where hedging will only cut out the variance from a negative EV - and that's assuming no further edge, which there usually is. (Even when the EV is currently positive, e.g., even money in blackjack, that will be the effect in the long run.) This is advantage gambling, where you want to minimize variance.

Quote: Ayecarumba

Is the fact that you are guaranteed (apologies to Ron Paul) a small win different than the "Thou Shalt Not Hedge" commandment #7?

No. Nobody knows what is going to happen in the election. Unless you can quantify the odds, as you can with card games, I say take the sure money.

Quote: Ayecarumba

How could you make money on both candidates?

Sounds like a math problem...

Let D be the money bet on Obama (Democrat) with Intrade, and R the money bet on Romney with Pinnacle.

If Romney wins, you gain 2.9 R on the Romney bet, and lose D on the Obama bet.
If Obama wins, you gain 0.47 D (100 / 212.5) on the Obama bet, and lose R on the Romney bet

The two returns are equal when 2.9 R - D = 0.47 D - R, or 3.9 R = 1.47 D
If you bet $147 on Romney and $390 on Obama, your profit is:
Romney win: 2.9 x 147 - 390 = 36.3
Obama win: 0.47 x 390 - 147 = 36.3
You make $36.30 for every $537 you bet regardless of who wins - that's about 6.75%.

Quote: Ayecarumba

Is the fact that you are guaranteed (apologies to Ron Paul) a small win different than the "Thou Shalt Not Hedge" commandment #7?

Technically there is not much difference between a hedge and an arbitrage. In both types you eliminate variance completely by the cost of reduced EV.
The difference between a hedge and an arbitrage is, that this reduced EV is still positive.

If you would know some true probability of the election, your EV per atake would be higher making a straight bet, but with increased variance your Kelly bet size will be limited. With reduced (still positive) EV the variance is zero, and you can bet your whole bankroll.
You should bet or arbitrage depending on the total (kelly determined) EV, whichever is larger - but this relies on an unbiased probability estimate. If you dont have such an estimate (or you are much more risk averse), the arbitrage is the way to go.

To expand on what Mango correctly said, the odds at Pinnacle suggest Romney is a fair bet at a 25.64% chance of winning. A Intrade, Romney is fair at 32.50%. If Romney's actual probability is anywhere between those two points then both sides of the arbitrage are good bets. If not, just one side is good.

Let's say I somehow felt that Romney's actual chance was 1 in 3. Then I should be betting on Obama at Intrade only. However, at some point, if I could bet enough, then it would be the correct Kelly strategy to also hedge at Pinnacle. As I've always said, if life changing amounts of money are involved, and/or the hedge bet is close to fair, then hedging can be the mathematically and economically optimal strategy.

Quote: Wizard

To expand on what Mango correctly said, the odds at Pinnacle suggest Romney is a fair bet at a 25.64% chance of winning. A Intrade, Romney is fair at 32.50%. If Romney's actual probability is anywhere between those two points then both sides of the arbitrage are good bets. If not, just one side is good.

This was the missing piece for me. Each individual punter's opinion and confidence of the candidates chances will vary; and some may feel certain one way or the other. Thank you all for the helpful information!

Quote: Ayecarumba

Each individual punter's opinion and confidence of the candidates chances will vary; and some may feel certain one way or the other.

The beauty of the arbitrage is: *you* don't need an opinion what a fair price will be, in fact you don't even care who wins the election. In effect you are just making a classic deal: buy low, and sell high. But you are not dealing with physical goods, you are dealing with "futures".

The only reason such arbitrage opportunity exists: the market is inefficient in finding a common price across different platforms. In fact arbitrage is the very mechanism to find a common price across different platforms: the low selling side will raise its prices with the strong demand from arbitrage bettors, and the high buying side can enforce lower prices from the large supply - until both prices are the same (and no arbitrage exists).

While this sounds good in theory, when this particular example is looked at closer, it's hardly worth the effort, and could even be a disaster for 2 reasons.

Reason #1, Intrade has 700 Obama shares at 7.20 ( -257). So you can buy $5040 to win $1960. Romney is currently +321 at Pinnacle, and you can buy anywhere up to 5k without worrying about a price adjustment, so you can bet $1663 to win $5338. So you are shelling out $6703 to win $297 or $298. Now, you have to get your money to and fro, so depending on the methods, you may incur some bank wire fees, or check fees from somewhere. So, lets just say it will end up costing you $30 in fees. Now you're looking at a little time and effort, and tying up close to 7k for close to 2 months for $250 and then there is always the possibility of reason #2 occurring.

Reason #2, This is not a true arbitrage opportunity because Intrade's price is for Obama ONLY, and not the Democratic candidate. What if Obama were to be assassinated, or even a freak accident occurred and he tripped over one of Michelle's tomatoes and fatally injured himself in the fall? Now, if they still ran the election with a fill in like Hillary, and the democrats won, you'd be out close to 7k. If the election is cancelled, you'd get refunded at Pinnacle, but what would happen to the Intrade bet? Since Obama wasn't re-elected, they could classify it as a loser.

With the return so small, and the possibility (albeit remote) of losing your whole stake, I see no point.

Additionally, there is the small possibility of a third-party win. This is a remote possibility when viewed over the global historical spectrum, but less remote when you consider the state of the economy and the massive contempt politicians are held in generally.

"Arbitrage" bets are rarely truly lock-solid. It is more like risking your bet for, say, a 2-3% return with a 1% chance of losing, analagous to a favourable short-priced bet.

Quote: GBV

Additionally, there is the small possibility of a third-party win.

Really? And which party are you referring to?

Somebody else brought up the possibility of an assassination causing somebody besides Romney and Obama winning. There are 42 days to go until the election. Since 1776 there have been four successful assassinations. That would make the mean number in a 42 day period 0.0004876. The probability of an assassination would be 1 in 2051. I would argue even less, due to increased security and improvements in medical care. Thus, I think the third party and assassination possibilities are negligible.

Update: Let me double that figure, because Romney could be assassinated too. So 1 in 1025.

Quote: Wizard

Really? And which party are you referring to?

Somebody else brought up the possibility of an assassination causing somebody besides Romney and Obama winning. There are 42 days to go until the election. Since 1776 there have been four successful assassinations. That would make the mean number in a 42 day period 0.0004876. The probability of an assassination would be 1 in 2051. I would argue even less, due to increased security and improvements in medical care. Thus, I think the third party and assassination possibilities are negligible.

In addition to Lincoln, Garfield, McKinley, and JFK, RFK was also killed while campaigning for the presidency. I don't know if there have been more challengers assassinated since 1776, but I don't think 4 is the right starting point. Yes, I realize that this only trivially changes the math...

Quote: rdw4potus

but I don't think 4 is the right starting point. Yes, I realize that this only trivially changes the math...

Bringing up RFK makes me realize that I forgot to account for the possibility that Romney is assassinated. So, let me adjust my figure to 1 in 1025.

By the way, I hope Secret Service web crawlers don't find this thread and then I get a knock on my door.

Quote: Wizard

Bringing up RFK makes me realize that I forgot to account for the possibility that Romney is assassinated. So, let me adjust my figure to 1 in 1025.

Mathematical dweeb here.

It isn't really good statistics to base probabilities on timelines. The idea of x attempts in y years results in a probability of and attempt in a given number of days is not sound because amount of time passing is not a relevant factor. It is equivalent to a betting system where you say the dealer hasn't busted in 15 hands, ...

It is notoriously difficult to quantify remote possibilities.

We had a discussion with a Zoologist on Spain. He concluded that whales in the Canary islands mass stranded whenever a Spanish navy exercise was being conducted. There was never a mass stranding when the Spanish navy was not conducting an exercise. The argument was when he tried to attach a probability to that occurring randomly.

It's much like the oft repeated observation that all 7 president's elected on a year that ended in zero from 1840 to 1960 died in office. There was only one president that died in office that was NOT elected in a zero year. Many people ask what is the probability of that? Some attribute this coincidence to a curse placed by Tecumseh, the Indian chief, in the battle of Tippecanoe. You may remember that Harrison used the campaign slogan, Tippecanoe and Tyler Too in the 1840 campaign.

Quote: pacomartin

It isn't really good statistics to base probabilities on timelines. The idea of x attempts in y years results in a probability of and attempt in a given number of days is not sound because amount of time passing is not a relevant factor. It is equivalent to a betting system where you say the dealer hasn't busted in 15 hands, ...

If we assume that the number of assassinations in a given period of time follows a Poisson distribution (in other words fixed probability at any given moment) then my method is perfectly valid and maintains a memory-less property. Let's use an example that is easier to sink our teeth into.

In the 236 years since independence there have been been 4 assassinations. Given that fact, and assuming no changes in that probability, what is the probability of at least one assassination in the next 100 years?

Answer:

Mean number of assassinations in 100 years = 100*(4/236) = 1.6949.

Probability of 1 or more assassinations in 100 years = 1-exp(-1.6949) = 81.64%.

If you differ with me answer, where is the flaw in my math?

Quote: Wizard

If we assume that the number of assassinations in a given period of time follows a Poisson distribution (in other words fixed probability at any given moment) then my method is perfectly valid and maintains a memory-less property. Let's use an example that is easier to sink our teeth into.

In the 236 years since independence there have been been 4 assassinations. Given that fact, and assuming no changes in that probability, what is the probability of at least one assassination in the next 100 years?

Answer:

Mean number of assassinations in 100 years = 100*(4/236) = 1.6949.

Probability of 1 or more assassinations in 100 years = 1-exp(-1.6949) = 81.64%.

If you differ with me answer, where is the flaw in my math?

Ford was shot at, Reagan was hit and almost died, and it is possible one of the 9/11 planes was headed for the White House. Do attempts count too, or only instances where the victim expires? Even a serious injury, not resulting in death could be enough to derail an election.

It's really irrelevant. Each assassination was based in the circumstances of the time, so you can't work it out like you'd analyze a die's bias. In any case, I'd say it's more probable than a third-party victory, but still far less than 1%.

I don't dispute that the probability is more, or less, now than any other time. However, Paco is challenging my math based on no change in that probability. I am defending myself under that assumption.

Pinnaclesports is actually Republican or Democrat, not Romney or Obama, so could still win if your side dies over the next month

When McCain and Dole were running I thought the chance of a candidate dieing during the campaign was at least slightly elevated

Quote: TomG

When McCain and Dole were running I thought the chance of a candidate dieing during the campaign was at least slightly elevated

The actuarial tables would certainly back you up.

Quote: Wizard

I don't dispute that the probability is more, or less, now than any other time. However, Paco is challenging my math based on no change in that probability. I am defending myself under that assumption.

I really didn't say that. I was challenging the assumption that there is no change in probability (which is inherent in just using timelines). Statisticians don't respect using time as the independent variable and saying "with all the variables, the probability density is uniform". In general, I think that they would discount assigning any number at all to the probability of an assassination happening.

The observation that president's die in office when they are elected on a zero year (starting with Harrison) has been speculated on since 1931 when it first received widespread notoriety. People who attempt to correlate it with statistics are making the same assumption that you are doing, the probability of the president dying is uniform over time.

The curse was put on Harrison during the Battle of Tippecanoe ( Nov 7, 1811). It was put on by "The Prophet" who was Tecumseh's brother after Tecumseh was killed. The curse began with Harrison being elected. President Harrison caught a cold during his inauguration speech because he thought wearing a hat in the rain made him look weak. He died in 31 days.

elect	#	President	Death Date	Party	Cause of Death
1800	3	Thomas Jefferson
1820	5	James Monroe
1840	9	William Henry Harrison	Apr 4, 1841	Democratic	Disease
1848	12	Zachary Taylor	Jul 9, 1850	Whig	Disease
1860	16	Abraham Lincoln	Apr 15, 1865	Republican	Assassination
1880	20	James A. Garfield	Sep 19, 1881	Republican	Assassination
1900	25	William McKinley	Sep 14, 1901	Republican	Assassination
1920	29	Warren G. Harding	Aug 2, 1923	Republican	Disease
1940	32	Franklin D. Roosevelt	Apr 12, 1945	Democratic	Disease
1960	35	John F. Kennedy	Nov 22, 1963	Democratic	Assassination
1980	40	Ronald Reagan	Mar 30, 1981	Republican	Assassination attempt
2000	43	George W. Bush	May 10, 2005	Republican	Assassination attempt

The exception to the curse is Zachary Taylor who was not elected in a zero year. Note that the zero year election does not have to be the first time elected. In the case of FDR it was his 3rd election. Also the president does not have to die within 4 years of the cursed election.

Supposedly Tecumseh's curse is getting weaker, because the last two victims survived the attempt. Reagan would have died from his wounds had he received that injury in 1811. Modern surveillance prevented the grenade attack on Bush's life. The curse may not have adapted to changes in technology.

Sometimes you just have to make a rough and dirty estimate of something. I make no apologies for it. I also specifically deny that an assassination ever becomes due, by virtue of my Poisson assumption.

By the way, I don't think that attempt on George W Bush should make your list. To be honest, I didn't even know about it until you brought it up. This source says the grenade landed 30 feet away from Bush, well outside the kill zone, even if it had gone off. If you count that one then there are a host of other amateurish attempts on other presidents that should also be counted.

Quote: Wizard

Sometimes you just have to make a rough and dirty estimate of something. I make no apologies for it. I also specifically deny that an assassination ever becomes due, by virtue of my Poisson assumption.

By the way, I don't think that attempt on George W Bush should make your list. To be honest, I didn't even know about it until you brought it up. This source says the grenade landed 30 feet away from Bush, well outside the kill zone, even if it had gone off. If you count that one then there are a host of other amateurish attempts on other presidents that should also be counted.

People use the GW Bush assassination attempt with the grenade just to complete the argument. To be perfectly clear, I consider the whole zero year election thing to be a coincidence and of no statistical validity. But it is an example of a time based pseudo statistics argument that people use. BTW the US government convicted him in absentia, so if Georgia ever lets him out of jail, the US government has the option to petition to have him sent to the USA to serve time here.

I wouldn't know how to come up with a probability number without assuming what you did. There is no other way to come up with a number unless you can make some reasonable correlation. I simply said that true statistician won't accept that argument.

The example from work is reproduced in the quote below. The zoologist did the same thing you did, he assumed a uniform probability density. Even if the conclusion is acceptable, the number was not accepted in peer review. I should point out that "Nature" is one of the most respected publications in the world, and they did publish this piece.

Quote: Does acoustic testing strand whales? Nature, Volume 392, Issue 6671, pp. 29 (1998).

We know that LFAS was used in the Kyparissiakos Gulf. We also know that no other LFAS tests or mass strandings have occurred in the Greek Ionian Sea since 1981. Taking the past 16.5-year period into account, the probability of a mass stranding occurring for other reasons during the period of the LFAS tests is less than 0.07%. Although pure coincidence cannot be excluded, it seems improbable that the two events were independent. Little is known about whales’ reactions to LFAS; to obtain definitive answers, more information needs to be gathered. But unfortunately, most of the data about the use of LFAS are subject to military secrecy.

I still don't understand what you're trying to say. If the time between assassinations does not follow an exponential distribution then what does it follow? Your argument seems void of any substance. Keep in mind you are speaking with somebody with a degree in math and member of the Society of Actuaries. So when it comes to statistics I would like to think I know a thing or two.

I'm not going to let this go. When it comes to math you have publicly challenged my honor. I demand you back up your argument with more facts than just a "true statistician wouldn't accept that argument." I also don't see what your graph about the whales has to do with it.

If I haven't made myself clear you have deeply offended and insulted me so I think I'm at least entitled to hear more substance to your argument.

This thread got weird fast

Time for a duel.

Quote: Wizard

If I haven't made myself clear you have deeply offended and insulted me so I think I'm at least entitled to hear more substance to your argument.

That was not my intention at all. I seem to have worded my response poorly, because I didn't even think you would have that reaction.

What I was trying to say is that the problem of trying to assign probabilities to events that coincide in time is extremely difficult for all scientists. Especially if you don't have an independent variable that you trust.

Let me use the Dr. Frantzis paper. He was looking at 16.5 years of data in the Canary islands, and there was a single mass stranding of whales in that entire time that occurred over a 4 day period. At the same four day period, there was a sonar test. He admitted that might be a coincidence, but he wanted to attach a firm number to the possibility of a coincidence.

Now the open ocean has millions of variables that you can't hope to deal with. There could be predators that come and go, there could be disease, etc. Given all those unknown variables, Dr. Frantzis chose to do a straightforward uniform probability density function and assume that all four day periods in that 16.5 years were independent. There are 1505 four day periods. So he felt that the probability of the two events occurring simultaneously was 1/1505.6 or < .07%.

Nobody was arguing with the fact that the two events happening at the same time was important. The argument was only with the calculation. Correlation never provide evidence for causation. A peer reviewed scientific paper should not use a statistical metric without first doing an experiment with an appropriate design for statistical analysis of conditional probability.

-------------------
Now, let's look at the presidents who died in office. The observation that 7 of the 8 times that it happened, the president was elected in a zero year. That is a correlation. It is not the result of a statistical test. If someone asks me "What is the probability of that occurring by chance" I would have to say that the question is unanswerable. As this observation has been widely circulated for 80 years, there is a lot of people who believe it is based on some reality. The believers struggle to explain the exceptions with the failed assassination attempts.

==============
My contention is that there is absolutely no way to use the number of successful assassinations to derive a probability function to assign future probabilities for a given length of time. From a scientific view, time is not a credible independent variable. It would only be a credible independent variable if there was some buildup of "public anger", similar to stress in a the San Andreas fault.

It was not my intention to challenge your mathematics, it is the age old problem of how do you analyze a problem without a good independent variable.


If I haven't made myself clear, I am talking about a universal problem here. Nothing to do with your mathematics.

Quote: 1BB

Time for a duel.

Actually, this problem is very common. We tested 40 units of a prototype, and we had three failures in a row. The Commander who was in charge insisted that the probability of this happening was so unlikely, that it had to be a failure in the instrumentation. In his mind if all events were independent, then the probability of 3 in a row out of 40 was 38 /combin(40,3)= 0.003846.
My argument was that I had verified that the failures were real, and the fact that they occurred in a row could be the result of some underlying unknown relationship (like the person who assembled the units). I also had concrete evidence that the instrumentation did not fail since part of the acoustic signature could not have occurred without it being a real life situation.

Naturally, they threw out the results, and the units failed so badly during real life testing that it ended up costing millions of dollars.

Quote: pacomartin

What I was trying to say is that the problem of trying to assign probabilities to events that coincide in time is extremely difficult for all scientists.

Sometimes you just have to make a quick "back of the envelope" estimate if you need something in a hurry and "it is too complicated" just won't do as an answer. It is easy for other to nit-pick such rough estimates, as you are doing. Perhaps your position is that if you can't provide an accurate answer, don't provide one at all. That isn't me. I will do the best I can with what I have in a limited amount of time. Even I admitted my estimate was probably high, due to increased Secret Service security.

My advice is that if you are going to challenge somebody in their area of expertise, do so carefully and respectfully.

Quote: Wizard

Perhaps your position is that if you can't provide an accurate answer, don't provide one at all.
My advice is that if you are going to challenge somebody in their area of expertise, do so carefully and respectfully.

It sounded a little different in my head, I didn't even think that I was challenging your expertise.

Don't forget that I have a master's in mathematics as well, and I was working with some of the most respected scientists in the world.

When we were dealing with this Nature paper, I had the same question. Given the lack of any other way to come up with a number, should we accept this calculation? However we had a group from MIT, Wood's Hole Institute and Cornell peer review the paper. There conclusion was precisely since an accurate answer can't be provided, no answer should be provided at all. They rejected any number and it was written up that way in the legal mandated impact statement for worldwide distribution.

So, a little more carefully and respectfully, I will paraphrase the language agreed to by all the scientists with regard to such rare events.

Quote: Consensus

A strong inference cannot be inferred (i.e. a hard number) without a test for the appropriate design for statistical analysis of conditional probability, and no such correlative study can provide evidence for causation.

This conclusion was subject to many questions. People kept asking why we were unable to provide a probability if we were rejecting the on by Dr. Frantzis. It helped that we had the consensus of all these internationally respected scientists.

I hope that statement convinces you that I didn't mean to challenge your expertise, even if I inadvertently made it sound that way.

Quote: pacomartin

What I was trying to say is that the problem of trying to assign probabilities to events that coincide in time is extremely difficult for all scientists. Especially if you don't have an independent variable that you trust.

The Wizard's assumption were (if I read it correctly), that the rate of the event in question is independent of time. Note that is quite the opposite of assigning probabilities that "coincide" with something. Anyway - the time-independence assumption naturally implies a Poisson distribution of events for any finite time interval. And if the assumption is true, the best estimate of the Poisson parameter is the observed event rate from past results.

Whether the initial assumption of time-indepenent event rate is true or not should not be a real dispute. Without assumptions, you cannot prove or calculate anything.

Quote: MangoJ

Whether the initial assumption of time-indepenent event rate is true or not should not be a real dispute. Without assumptions, you cannot prove or calculate anything.

Exactly.

If I could go back to my original post I should have reworded it radically.

In my experience doing a peer review of an article that made the time-independent event rate assumption, the group of scientists which included a Guggenheim Genius grant winner, and Phd's from Cornell, MIT, and Wood's Hole, they flat out rejected the assumption. Their conclusion was that no calculation should be made.

In contrast to Presidential history, while the office of Vice President has been empty for over 37 years out of 223, there has been no VP that has been assassinated (although 7 died in office of natural causes in the 19th century). Using the time-independent event rate assumption it would imply that Joe Biden has a zero percent possibility of being assassinated.

Quote: pacomartin

they flat out rejected the assumption. Their conclusion was that no calculation should be made.

Again, thats not the point. You can always reject any assumption you like, but then you are not looking at the same animal anymore.

I could say "(1) If every natural number is even and there exists a smallest natural number, then '2' is the smallest natural number".

Now you can say "Well I don't believe in even numbers, so I discard this assumption."
Then you would talk about the statement "(2) If there exists a smallest natural number, then '2' is the smallest natural number", which is false.

Am I wrong ? No. The original statement (1) is perfectly true - the prove is simple. By dropping an assumption you are talking about a totally different thing.

Bovada released some new lines on the presidential election today that I found interesting. Specifically, they are offering a couple different prop bets on who will win the states of Colorado, Nevada, Iowa, Wisconsin, Ohio, Florida, North Carolina, Virginia, New Hampshire. In addition to a specific money line on each state, a few on the group as a whole were listed.

Who will claim more of the above states?
Obama -1.5 -200 (Must win 6 or more to win)
Romney +1.5 +150 (Must win 4 or more to win)

How many of the above states will be won by the Democratic Party?
Over 5 Even
Under 5 -130

How many of the above states will be won by the Republican Party?
Over 4 +110
Under 4 -140

I feel like the best bet on this list is Romney +1.5 +150. North Carolina is locked up so he only needs three more. Florida is looking strong as well, thus Romney only needs two of Colorado, Nevada, Virginia, New Hampshire and Iowa to cash the bet. I noticed an arbitrage opportunity as either Romney +1.5 or Democratic Party Over 5 must occur. Additionally, I feel a scenario where Obama wins 5 of the 9 to be most likely allowing for a Win and a Push on the pair of bets. I wanted to see if anyone else had thoughts on this or if my logic is flawed.

Thought I might perform some necromancy to point out that, for those who don't have a Pinnacle account, Intrade is selling Obama at $6.60 a share, whereas Bovada is giving +275 on Romney (look under "Entertainment," disturbingly). Obviously, you'll have to act fast, but (and I don't know a whole lot about this, so I might be mixed up) that's a bigger middle than in the OP: a potential yield of 7.91% on your money, versus 6.79%.

Quote: 1arrowheaddr

I feel like the best bet on this list is Romney +1.5 +150. North Carolina is locked up so he only needs three more. Florida is looking strong as well, thus Romney only needs two of Colorado, Nevada, Virginia, New Hampshire and Iowa to cash the bet.

Romney winning more than North Carolina and Florida will be difficult.

Yes, I imagine there will be opportunities as I doubt the various Prediction Markets are overlooked by campaign advisers who realize that there is money in being the favorite in a prediction market and that such money comes from winning the election not winning the prediction market's odds maker.

Quote: pacomartin

I really didn't say that. I was challenging the assumption that there is no change in probability (which is inherent in just using timelines). Statisticians don't respect using time as the independent variable and saying "with all the variables, the probability density is uniform". In general, I think that they would discount assigning any number at all to the probability of an assassination happening.

The observation that president's die in office when they are elected on a zero year (starting with Harrison) has been speculated on since 1931 when it first received widespread notoriety. People who attempt to correlate it with statistics are making the same assumption that you are doing, the probability of the president dying is uniform over time.

The curse was put on Harrison during the Battle of Tippecanoe ( Nov 7, 1811). It was put on by "The Prophet" who was Tecumseh's brother after Tecumseh was killed. The curse began with Harrison being elected. President Harrison caught a cold during his inauguration speech because he thought wearing a hat in the rain made him look weak. He died in 31 days.

elect	#	President	Death Date	Party	Cause of Death
1800	3	Thomas Jefferson
1820	5	James Monroe
1840	9	William Henry Harrison	Apr 4, 1841	Democratic	Disease
1848	12	Zachary Taylor	Jul 9, 1850	Whig	Disease
1860	16	Abraham Lincoln	Apr 15, 1865	Republican	Assassination
1880	20	James A. Garfield	Sep 19, 1881	Republican	Assassination
1900	25	William McKinley	Sep 14, 1901	Republican	Assassination
1920	29	Warren G. Harding	Aug 2, 1923	Republican	Disease
1940	32	Franklin D. Roosevelt	Apr 12, 1945	Democratic	Disease
1960	35	John F. Kennedy	Nov 22, 1963	Democratic	Assassination
1980	40	Ronald Reagan	Mar 30, 1981	Republican	Assassination attempt
2000	43	George W. Bush	May 10, 2005	Republican	Assassination attempt

Don't forget the shoe attempt!

Quote: pacomartin

Romney winning more than North Carolina and Florida will be difficult.

It was looking a lot better for Romney when I made that post 18 days ago. He led in Virginia and Colorado at the time. Bovada took that line down shortly after I made that post anyway.

I just bet Obama at -350 at Bovada, and -175 for Obama to win Ohio.

Between those, and many other bets, I think you'll see selling water bottles on the pedestrian bridge between the Bellagio and Caesars if Romney wins.

Quote: Wizard

I just bet Obama at -350 at Bovada, and -175 for Obama to win Ohio.

Between those, and many other bets, I think you'll see selling water bottles on the pedestrian bridge between the Bellagio and Caesars if Romney wins.

Be careful!

Quote: Wizard

I just bet Obama at -350 at Bovada, and -175 for Obama to win Ohio.

Between those, and many other bets, I think you'll see selling water bottles on the pedestrian bridge between the Bellagio and Caesars if Romney wins.

Fortunately, I don't think this will be a problem, which is good for me because I prefer the water salesmen on the Bally's-Bill's bridge anyway!

Quote: Wizard

I just bet Obama at -350 at Bovada, and -175 for Obama to win Ohio.

Between those, and many other bets, I think you'll see selling water bottles on the pedestrian bridge between the Bellagio and Caesars if Romney wins.

Meh, your boss will be so elated that he'll give out raises all around.

Quote: s2dbaker

Meh, your boss will be so elated that he'll give out raises all around.

The Wiz is his own boss again. That jerk who hates online poker got rid of his position.