Any set of rules will have a way to play. What's the fundamental difference? The timing gap? Or the fact that the whole curve is there?
The fact that
a) people have to explicitly commit to every price they could end up having to trade at (i.e. no "sell these shares 'at the market'... holy shit! A flash crash! No, no, reverse the trade! Waaaah!")
b) no one gains an advantage from an order being placed sooner or later than another's
c) no one gains an advantage from knowing what others are currently offering
People out there will spoof the curve, getting other people to think there's real demand at this or that point
You can't spoof the curve: no one sees the curve for the current auction until it's resolved, and they never see anyone's individual curve. Plus, they have to honor any fake supply/demand curve they put in. "Oh, you were just "kidding" about being willing to buy at $1000/share? Too bad, it was accepted since it found buyers."
As long as people only submit bids for what they think it's worth, decoupled from other estimates on the current auction, they never have to take an offer they didn't explicitly make themselves. The fact people who run with the herd can be deceived is a feature, not a bug.
And they'll guess that the cost of being called on their bluff is lower than what they make.
Like above, if they "bluff" that they'd buy at $1,000 when the last trade was at $100, then when the computer collects all bids on the current auction, it sees someone willing to buy for an insanely high price. Since many sellers are willing to sell for more than $100, the computer necessarily matches the offer with *some* other buyer, it's only an issue of how the overpaid share money is distributed between sellers, and what price (between $100 and $1,000) the algorithm declares for the trade.
Yep, people game systems that are different in fundamental ways from the one I'm describing. Still not responsive to the issue at hand.
BTW, how exactly do you enter your whole supply/demand curve? How do you keep it honest?
That is how Swedish utilities buy power, so it's definitely possible. Also, limit orders already do this: they say, e.g., "I will buy up to 40 shares at $100/share or any lower price; nothing otherwise."
You keep it honest because you are required to honor any offer within your curve. So you don't *really* want to buy any shares at $100? Then put a zero at that point (and all higher prices), or else someone can take the offer and force you to buy at that price.
~*sweet!*~ I get to put a whole $5000 each year, and I only have to wait until my life is almost over to touch it! And things only have go right for another 30 years or so. That is SO FUCKING AWESOME!
Say, know anyone who will let me split that across five mutual funds having $1000 each without having low-balance fees that defeat the whole purpose of it?
So how did that end up ripping off people who entered clear supply/demand curves for the securities, again?
Oh, you mean there was another significant divergence from what you're describing and the auction system I described? Well, try again when you can do a little better than pattern-matching for analysis.
That's not necessary -- you can just remove the margin to be faster altogether. Just collect all the bids placed before the fixed point, without telling anyone what the other bids are. (i.e., everyone sends in their supply/demand curve for a given security) Then, the computer could look at just those bids (any further bids would apply to the *next* fixed time), resolve the trades by an algorithm, and spit out the results when it's done.
In that case, everyone who got their bet in before the fixed time is on equal footing. You don't know any more or less, nor get any better or worse treatment as a result of being able to place a bid one pico/nano-second before the fixed time, and the race to compete on insanely fast execution of trades is over.
If the hackers are netting themselves a bunch of money by out-trading the other high-frequency-traders... good for them. It's not my money they're taking...
That's what I thought, too -- until Fall '08 hit, and I found out that if one of the big players lose to these guys, the government bails them out (at which point it *is* my money they're taking), revealing as a sham this whole idea that the big guys nobly make risky bets. No, if you're going to be bailed out on the downside, you weren't taking a risk to begin with -- ever.
In theory, you're right -- but let's bring back the concept of "failing when you're wrong" to Wall Street before blithely dismissing the harm these guys can cause.
And seriously -- is the tiny bit of extra liquidity REALLY worth the billions these guys sink into HFT?
And police don't have a right to demand that your girlfriend have sex with them to prove that she likes sex -- because if she doesn't like sex then you must be a rapist. (You think such things have never happened?)
Yep, I'll stick my neck out here: I don't think such things have ever happened. The fact that my girlfriend is (or isn't) willing to have sex with random police officers isn't very helpful in settling the matter of whether I'm a rapist, so I'm skeptical such an "acid (!) test" has ever been used, at least in the West.
Sounds like the military version of the TSA incident:
"What??? You're limiting what publicly-available classified material I can see, when I could already defect with an Air Force plane if I wanted... [/court-martial]"
Wait, wasn't there just a slashdot story showing how the pro-Wikileaks "hacktivists" can be easily identified? If so, it seems they actually are putting themselves at risk.
Yeah, good point. I could fund fantastically expensive projects if I could borrow huge amounts of money, get $50 billion in free government money, stiff bondholders, and screw pensioners on what I promised them.
That's where 99% of GM's brilliance lies. Anyone who could play with that much money could probably do something more impressive.
And have done so: even NASA, with all the bureaucratic stuff they have to put up with, can do a lot more with the same money.
I thought Linux was always 100% secure, completely unhackable, because they're the admins are experts and the software has built in security! You mean... it's not true?
Sorry, you stopped short of the part where you carefully explained how any of this makes the *buyer of the stock* a criminal in the sense that a shoplifter is, and where the justifications given for both are equally invalid rationalizations of knowingly immoral behavior.
Uh, wait, so shoplifting is morally equivalent to buying a portfolio of voting rights in productive enterprises (that those enterprises knowingly issued for that purpose) and to failing to speak up soon enough about your negative estimate of R&D success?
No. Just... just, no. That is about fifty different kinds of wrong.
Slashdot Mirror
Any set of rules will have a way to play. What's the fundamental difference? The timing gap? Or the fact that the whole curve is there?
The fact that
a) people have to explicitly commit to every price they could end up having to trade at (i.e. no "sell these shares 'at the market' ... holy shit! A flash crash! No, no, reverse the trade! Waaaah!")
b) no one gains an advantage from an order being placed sooner or later than another's
c) no one gains an advantage from knowing what others are currently offering
People out there will spoof the curve, getting other people to think there's real demand at this or that point
You can't spoof the curve: no one sees the curve for the current auction until it's resolved, and they never see anyone's individual curve. Plus, they have to honor any fake supply/demand curve they put in. "Oh, you were just "kidding" about being willing to buy at $1000/share? Too bad, it was accepted since it found buyers."
As long as people only submit bids for what they think it's worth, decoupled from other estimates on the current auction, they never have to take an offer they didn't explicitly make themselves. The fact people who run with the herd can be deceived is a feature, not a bug.
And they'll guess that the cost of being called on their bluff is lower than what they make.
Like above, if they "bluff" that they'd buy at $1,000 when the last trade was at $100, then when the computer collects all bids on the current auction, it sees someone willing to buy for an insanely high price. Since many sellers are willing to sell for more than $100, the computer necessarily matches the offer with *some* other buyer, it's only an issue of how the overpaid share money is distributed between sellers, and what price (between $100 and $1,000) the algorithm declares for the trade.
Dude, that's what people do.
Yep, people game systems that are different in fundamental ways from the one I'm describing. Still not responsive to the issue at hand.
BTW, how exactly do you enter your whole supply/demand curve? How do you keep it honest?
That is how Swedish utilities buy power, so it's definitely possible. Also, limit orders already do this: they say, e.g., "I will buy up to 40 shares at $100/share or any lower price; nothing otherwise."
You keep it honest because you are required to honor any offer within your curve. So you don't *really* want to buy any shares at $100? Then put a zero at that point (and all higher prices), or else someone can take the offer and force you to buy at that price.
~*sweet!*~ I get to put a whole $5000 each year, and I only have to wait until my life is almost over to touch it! And things only have go right for another 30 years or so. That is SO FUCKING AWESOME!
Say, know anyone who will let me split that across five mutual funds having $1000 each without having low-balance fees that defeat the whole purpose of it?
So how did that end up ripping off people who entered clear supply/demand curves for the securities, again?
Oh, you mean there was another significant divergence from what you're describing and the auction system I described? Well, try again when you can do a little better than pattern-matching for analysis.
Then would you mind being a bit more specific about how the auction type I described is traditionally gamed?
In auctions, people know the other bids, and can adapt the information they reveal as they learn this. Under what I proposed, they don't and can't.
That's not necessary -- you can just remove the margin to be faster altogether. Just collect all the bids placed before the fixed point, without telling anyone what the other bids are. (i.e., everyone sends in their supply/demand curve for a given security) Then, the computer could look at just those bids (any further bids would apply to the *next* fixed time), resolve the trades by an algorithm, and spit out the results when it's done.
In that case, everyone who got their bet in before the fixed time is on equal footing. You don't know any more or less, nor get any better or worse treatment as a result of being able to place a bid one pico/nano-second before the fixed time, and the race to compete on insanely fast execution of trades is over.
If the hackers are netting themselves a bunch of money by out-trading the other high-frequency-traders... good for them. It's not my money they're taking...
That's what I thought, too -- until Fall '08 hit, and I found out that if one of the big players lose to these guys, the government bails them out (at which point it *is* my money they're taking), revealing as a sham this whole idea that the big guys nobly make risky bets. No, if you're going to be bailed out on the downside, you weren't taking a risk to begin with -- ever.
In theory, you're right -- but let's bring back the concept of "failing when you're wrong" to Wall Street before blithely dismissing the harm these guys can cause.
And seriously -- is the tiny bit of extra liquidity REALLY worth the billions these guys sink into HFT?
Or when you swallow someone's spooge.
Confirmed. I can't find it on my Kindle either.
Good points, except for:
And police don't have a right to demand that your girlfriend have sex with them to prove that she likes sex -- because if she doesn't like sex then you must be a rapist. (You think such things have never happened?)
Yep, I'll stick my neck out here: I don't think such things have ever happened. The fact that my girlfriend is (or isn't) willing to have sex with random police officers isn't very helpful in settling the matter of whether I'm a rapist, so I'm skeptical such an "acid (!) test" has ever been used, at least in the West.
Climate shifts are due to the shifting magnetic properties of the earth's core?
~*SWEEEET!*~
Thanks for giving me my latest skeptical counter-theory to anthropogenic global warming!
And finally... Why the heck are our friends at Siemens selling systems to the Iranians?
I don't know, but I hope they shoot 'em an even bigger load next time [/couldn't resist].
Well, baking soda actually does work for some causes of abdominal pain because it neutralizes acid ... but your basic point is correct.
"Scientists use quasars like a drunk uses a lamppost: for support rather than illumination."
Yeah, well, energy-efficient technology always finds a way of looking phallic.
Sounds like the military version of the TSA incident:
"What??? You're limiting what publicly-available classified material I can see, when I could already defect with an Air Force plane if I wanted ... [/court-martial]"
Except that he'll edit all the records to make it look like the kids shot first ... oh, the irony.
Well, not anymore I guess. I don't have another family for you to threaten to kill!
Wait, wasn't there just a slashdot story showing how the pro-Wikileaks "hacktivists" can be easily identified? If so, it seems they actually are putting themselves at risk.
Yeah, good point. I could fund fantastically expensive projects if I could borrow huge amounts of money, get $50 billion in free government money, stiff bondholders, and screw pensioners on what I promised them.
That's where 99% of GM's brilliance lies. Anyone who could play with that much money could probably do something more impressive.
And have done so: even NASA, with all the bureaucratic stuff they have to put up with, can do a lot more with the same money.
Fuck GM. Fuck GM long, fuck GM hard.
I thought Linux was always 100% secure, completely unhackable, because they're the admins are experts and the software has built in security! You mean ... it's not true?
Didn't the IRA phone in warnings of their bombings in advance so people could evacuate the area?
Convince al-Qaeda to do that, and you have my vote.
Sorry, you stopped short of the part where you carefully explained how any of this makes the *buyer of the stock* a criminal in the sense that a shoplifter is, and where the justifications given for both are equally invalid rationalizations of knowingly immoral behavior.
Uh, wait, so shoplifting is morally equivalent to buying a portfolio of voting rights in productive enterprises (that those enterprises knowingly issued for that purpose) and to failing to speak up soon enough about your negative estimate of R&D success?
No. Just ... just, no. That is about fifty different kinds of wrong.