Yeah, and the government still owes my family $400,000 plus interest from 1955 or thereabouts, for unpaid bills on that scrapped project. B*()^(&^(&#(*&%!!!!
Waves are exactly as predictable as stock market behavior, given the same level of information. If this were not so, the quants would have lost all their money. They _successfully_ applied physics models to markets.
Why do you think all the physics PhDs in the 1990s went to work on Wall Street? They could make 10x the money as they could teaching physics, _applying_physics_models_ to financial systems. Note also how neural networks have been successful as well. Neural networks happen to be good at modeling complex adaptive systems and dynamical systems.
It is essentially impossible to predict the next breaker is going to run up the beach, but it is relatively easy to predict how high the average breaker will be based on the wind and water velocities and the shape of the shore - but one has to look at the whole ocean to determine the sources of energy. A hurricane 1000 miles away can generate waves that will be the same height as a smaller wind close by - but shaped differently.
The rise of the quants and HFT formed a classic tech bubble, but now that both systems have been widely adopted, they are just part of the system, that has brought markets much more closely to the ideal 'efficient market'. For every HFT betting on +A there must be another HFT betting on -A (if not, the trade will not occur). The result is that the margin between +A and -A has become much smaller, and since there are other traders working at other time scales the curve of the flow rate between +A and -A has become much smoother - no more square wave (with its essential ringing).
Also in the past, market makers could make from 1/8 to 1/2 a dollar per share on the difference between bid and ask. Now that difference may be as small as 0.01 dollar (using US funds as an example). The effect in the long term, as this area matures, is that the market will act more and more smoothly to events in the real world.
But, of course, this has required the development of new rules to manage the flow velocities. For a very simple example, a rule that limits trading in a particular stock if the price changes by 5% in 10 minutes can be seen as forcing the flow to go through a smaller pipe, with predictable effects.
A savings account works a lot like a capacitor - or an inlet to the ocean (see how the rate of tidal flow varies with the different shapes of inlets, and the relative tidal force)
As a whole, economic systems are more easily viewed in information space, not physical space. But these two spaces have long ago been determined to be duals.
But the best way to look at economies is as complex adaptive systems. See the Santa Fe Institute for information on CAS. Of course, CAS are particular examples of dynamical systems.
If this were not true, the quants would have failed. Their singular advantage was recognizing (in a limited way) that economic systems do behave as complex systems. Just as one can not easily follow the motion of a single molecule in a gas cloud, but can model the behavior of the cloud via fluid dynamics, so also the behavior of a single financial instrument can not be fully predicted but the overall behavior of the system can be predicted to some extent - like the weather - and so the behavior of that instrument can be predicted within some range of probability. The predictability of the weather decreases as the physical scale is reduced, or the time scale is increased. But you can still say with 40% certainty that it will rain here on Thursday.
Sigh. An economic system is an implementation of fluid dynamics, or if you prefer complex adaptive systems (which amount to much the same thing in the limit.) What you are suggesting is tantamount to saying the ocean should be made of ice cubes, in order to avoid waves getting too big. Now that the tech bubbles around quants and HFT have been popped, and (important!) there are equally talented traders and programmers on both/all sides of the system, these are valuable tools for making the market more truly follow the essential physics.
Wrong X 2. While HFT and Quant analysis both generated tech bubbles in the financial markets, now that the bubble has passed they are merely tools that decrease the granularity (in quantities, prices and time-frames) of the efficient market. These technologies are turning the markets into much cleaner implementations of the essential fluid dynamics of the economic system.
Just for a very simple example, until very recently market makers could hold the difference between bid and ask prices a minimum of 1/8 of a dollar apart, making profits on that difference - you sell for $1.25, I buy for $1.375 and the difference goes to the market maker. Now the difference is as small as $0.01 - in fact (I haven't researched this) it may be less than that for high volume trades. And for every computer that wants to buy there has to be a computer that wants to sell. So those folks have to keep running just to stay where they are.
Indeed, the quants made a boatload of money (that's a separate group from the mortgage scammers), but eventually all that money will get recycled into the system. In fact, as they invest in real business they've probably already created more jobs directly and indirectly than the government. What they spend will go to making real things as well.
Considering that $billions are being spent in order to shorten the lag for data transfer by microseconds, I'll bet the kernel has been pruned down and tweaked to the maximum extent.
A new data pipeline has just been completed at a cost of $100s of millions, going literally as straight as possible from Chicago to New York, in order to save a few milliseconds (or microseconds - I've heard both). For perspective, each microsecond is on the order of 200 meters of fiber optic cable (Wikipedia sez the speed of light in optical fiber is about 200,000 km/s). This in order to speed the access to the New York trading computers from Chicago, to try to stay as current as possible with the servers they are building into new facilities within a couple of blocks from NYSE computers. Just having your server farm uptown instead of downtown costs numerous microseconds, giving the ones closer a significant advantage.
Of course, now that we're past the bubble, this is just technology implementing the curve smoothing to higher and higher precision, making the market ever closer to the pure physics of the Efficient Market - which (IMHO) is basically an ocean in a large number of dimensions, with Brownian motion, ripples, waves bigger waves all generated by events large and small globally and interacting according to the physics of fluid dynamics.
Funny you should mention this. The very first hack that I ever heard of, back in the early 1970s (yes, I'm that old) was purportedly done to what is now Citibank in the mid-late 1960s. This was when they had one of the first IBM 360 machines - the standard mainframe (about the same power as a PC/AT). The guy who wrote their loan computation program wrote self-modifying code. When the program ran, it stored various constants in various places in memory, that happened (when these constants were all strung together in sequential memory) to be a hack. Then all he had to do was jump to the start of that routine, and it modified the interest computation function used by every part of the system and some other things, and... You have to realize that self-modifying code was still quite common at that time so nothing he did looked out of the ordinary.
The result of all this was that when interest payments were computed on every account (savings, checking, loans, etc), instead of rounding at 1/10 mil (1/1000 of a dollar), it truncated, and took the difference and put it into an account that only existed inside the machine. At the end of the day, the money in that account was quietly pushed into a transfer that sent the money by wire to a Swiss account, and the account disappeared.
The programmer, having finished his job, went on to other things. This hack supposedly went unnoticed for a couple of years, until for unrelated reasons the bank had to do a line-by-line audit of their accounts, and came up several million dollars short. They spent months trying to figure out where the money went but couldn't find it. Finally they went to the guy and said, "you can keep the money and we won't prosecute. Just tell us what you did so we can prevent it happening again." By then he had made something under $10 million.
I have no idea if this is true, or a very early urban legend. But I heard it when I was in school in the early 1970s, so it's at least that old.
I always thought they were Demoplicans. (or was it Demorons??)
In all fairness, the folks who succeed in becoming high level politicians got there by signing on to the establishment myths that prevailed when they were in their formative years - few of them got there by thinking for themselves. Those myths still drive their reasoning, and those myths still control most of the people in most of the think tanks and lobbying groups. So it's understandable that they continue to try the same things that their predecessors did. I expect a grand total of 1% of Washington insiders of all stripes have ever actually read Keynes in the original - where, among other things, he pointed out that deficit spending could be successful *for a short time* as a measure to get money flowing again - he never implied that it should be a permanent fixture in the economic landscape.
Highly recommended, these two books (get them as audiobooks if you want hours of driving amusement, education and horror). Just for perspective (one that is not in these books, or anywhere else that I know of - I thunk it up with my own little brain), this whole scenario can be seen as a classic tech bubble - the technology being the application of algorithms and various other forms of computer technology to technical stock trading. This bubble began back in the early 1970s, and happened to hit critical mass just when everything else was happening - the housing bubble, etc. Since this tech bubble lived in the second order space of the financial economy that sits on top of the first-order "real-world" economy, it got boosted out of sight by the multiplier of the two.
This. Also, "Your website has a bug." "What platform and browser are you running?" "Ubuntu 11.04, Firefox - I don't know what version, it's the latest." "Thanks, we'll get right on that - after the next millenium. Wanker."
Actually I don't like Chrome, and I am hoping that things just keep working in FF without constant fiddling. And I _really_ don't like things breaking the several add-ons that I make use of constantly.
I just imagined my car being on a rapid-release cycle, and the various kinds of bad things that could happen on the way home from work as the latest update accidentally disabled the brakes, or set the radio to play rap at full volume (not sure which is worse, actually). Eeek!
I've thought about this (back when I rode the bus to work) - in fact I considered finding out the total amount of fuel used by the buses in Portland OR (where I was riding), and the total ridership. This seems to me to be a simple computation - divide the fuel by the number of passengers, and you get the gallons per trip. The agency probably also has a good idea of the length of the average trip (but that is also complicated by the fact that the bus does not go from the passenger's point A to point B - it goes along its route, which may twice as long as the direct route.)
The reason I got interested was that the buses in Portland had stickers on the back saying, "This bus takes 152 cars off the road", which I figured was probably a self-serving and specious statement. IIRC the average bus gets something like 3 MPG in actual use, and the average ridership is on the order of 10 (obviously commute-time trips are more efficient, midnight runs with empty buses cost a lot.) So this means the bus is getting about 30 passenger miles per gallon average. Considering the routing inefficiency, things look worse.
To make things even worse, an impromptu experiment some years ago occurred due to a bus strike in London. The average speed of traffic almost doubled.
The ultimate solution would be to re-institute the 'trolley towns' that used to grow up around each trolley (light rail) station, before everyone had cars. Then trolleys could run at optimal efficiency with minimum stops. Trolleys running on their own lanes or rails are the only way that mass transit can improve their speed relative to driving.
It's fission. I haven't read the article but several folks who did said that it mentions 'subcritical' fission - IOW the reaction does not sustain itself, it must be fed extra neutrons, which are generated by a laser-driven particle accelerator.
Oddly enough, my dad built the engine plants for the Atomic Airplane project, back in the 1950s. The project was snakebit from the start. The buildings are now part of the Idaho National Laboratory. He got stiffed for $400,000 by the US Government - in 1954 that was not pocket change..
IIRC the output of proposed 1000 MW Liquid Thorium Fluoride Reactors (power-station-scale) is non-radioactive nickel and boron, or some such. The amount of radioactive waste for such a plant is expected to be less than a teacup per year. None of it is bomb-making material. As a continuous ambient-pressure liquid system, it's much safer and easier to control than uranium-fission plants. So for power plants, it's almost ideal. IMHO if we had gone to LTFR back in the late 1960s (as we could have, but AEC lost interest because LTFRs don't produce bomb materials!) by now the prospect of switching from 'safe, efficient, scalable LTFRs to uranium power plants would be unthinkable.
I don't know about the system in the article, but the LTFR (Liquid Thorium Fluoride Reactor - a power station-scale reactor technology) runs at ambient pressure. And the amount of nuclear waste is something like a teacup per year, for a 1000 MW station.
I haven't read the article (probably won't) but there is legitimate research into using a laser to drive a particle accelerator, which provides the slow neutrons necessary to trigger the Thorium fission chain reaction.
I haven't read the article (and probably won't) but it's worth your time to look up Liquid Fluoride Thorium Reactors. (I always get LFTR and LTFR confused - sorry). For actual power plants I think LTFR are probably the best solution to all the problems with uranium-fueled reactors. The biggest force preventing them may turn out to be the present big suppliers of uranium fuel rods - a very lucrative industry that would be destroyed. Switching power to LTFR would provide sufficient power to do almost anything we want to do for perhaps 10,000 years, by which time I would hope that we would have figured out how to tap the vacuum energy!:D
Indeed - container ships would be a great application. I've read that ocean shipping is now producing some huge fraction of all the CO2. And their existing fuel tanks and engines are probably larger than an LTFR reactor that would produce the required 200 KW. So, good idea.
Slashdot Mirror
Yeah, and the government still owes my family $400,000 plus interest from 1955 or thereabouts, for unpaid bills on that scrapped project. B*()^(&^(&#(*&%!!!!
Waves are exactly as predictable as stock market behavior, given the same level of information. If this were not so, the quants would have lost all their money. They _successfully_ applied physics models to markets.
Why do you think all the physics PhDs in the 1990s went to work on Wall Street? They could make 10x the money as they could teaching physics, _applying_physics_models_ to financial systems. Note also how neural networks have been successful as well. Neural networks happen to be good at modeling complex adaptive systems and dynamical systems.
It is essentially impossible to predict the next breaker is going to run up the beach, but it is relatively easy to predict how high the average breaker will be based on the wind and water velocities and the shape of the shore - but one has to look at the whole ocean to determine the sources of energy. A hurricane 1000 miles away can generate waves that will be the same height as a smaller wind close by - but shaped differently.
The rise of the quants and HFT formed a classic tech bubble, but now that both systems have been widely adopted, they are just part of the system, that has brought markets much more closely to the ideal 'efficient market'. For every HFT betting on +A there must be another HFT betting on -A (if not, the trade will not occur). The result is that the margin between +A and -A has become much smaller, and since there are other traders working at other time scales the curve of the flow rate between +A and -A has become much smoother - no more square wave (with its essential ringing).
Also in the past, market makers could make from 1/8 to 1/2 a dollar per share on the difference between bid and ask. Now that difference may be as small as 0.01 dollar (using US funds as an example). The effect in the long term, as this area matures, is that the market will act more and more smoothly to events in the real world.
But, of course, this has required the development of new rules to manage the flow velocities. For a very simple example, a rule that limits trading in a particular stock if the price changes by 5% in 10 minutes can be seen as forcing the flow to go through a smaller pipe, with predictable effects.
Velocity of Money
A savings account works a lot like a capacitor - or an inlet to the ocean (see how the rate of tidal flow varies with the different shapes of inlets, and the relative tidal force)
As a whole, economic systems are more easily viewed in information space, not physical space. But these two spaces have long ago been determined to be duals.
But the best way to look at economies is as complex adaptive systems. See the Santa Fe Institute for information on CAS. Of course, CAS are particular examples of dynamical systems.
If this were not true, the quants would have failed. Their singular advantage was recognizing (in a limited way) that economic systems do behave as complex systems. Just as one can not easily follow the motion of a single molecule in a gas cloud, but can model the behavior of the cloud via fluid dynamics, so also the behavior of a single financial instrument can not be fully predicted but the overall behavior of the system can be predicted to some extent - like the weather - and so the behavior of that instrument can be predicted within some range of probability. The predictability of the weather decreases as the physical scale is reduced, or the time scale is increased. But you can still say with 40% certainty that it will rain here on Thursday.
Sigh. An economic system is an implementation of fluid dynamics, or if you prefer complex adaptive systems (which amount to much the same thing in the limit.) What you are suggesting is tantamount to saying the ocean should be made of ice cubes, in order to avoid waves getting too big. Now that the tech bubbles around quants and HFT have been popped, and (important!) there are equally talented traders and programmers on both/all sides of the system, these are valuable tools for making the market more truly follow the essential physics.
Wrong X 2. While HFT and Quant analysis both generated tech bubbles in the financial markets, now that the bubble has passed they are merely tools that decrease the granularity (in quantities, prices and time-frames) of the efficient market. These technologies are turning the markets into much cleaner implementations of the essential fluid dynamics of the economic system.
Just for a very simple example, until very recently market makers could hold the difference between bid and ask prices a minimum of 1/8 of a dollar apart, making profits on that difference - you sell for $1.25, I buy for $1.375 and the difference goes to the market maker. Now the difference is as small as $0.01 - in fact (I haven't researched this) it may be less than that for high volume trades. And for every computer that wants to buy there has to be a computer that wants to sell. So those folks have to keep running just to stay where they are.
Indeed, the quants made a boatload of money (that's a separate group from the mortgage scammers), but eventually all that money will get recycled into the system. In fact, as they invest in real business they've probably already created more jobs directly and indirectly than the government. What they spend will go to making real things as well.
Considering that $billions are being spent in order to shorten the lag for data transfer by microseconds, I'll bet the kernel has been pruned down and tweaked to the maximum extent.
A new data pipeline has just been completed at a cost of $100s of millions, going literally as straight as possible from Chicago to New York, in order to save a few milliseconds (or microseconds - I've heard both). For perspective, each microsecond is on the order of 200 meters of fiber optic cable (Wikipedia sez the speed of light in optical fiber is about 200,000 km/s). This in order to speed the access to the New York trading computers from Chicago, to try to stay as current as possible with the servers they are building into new facilities within a couple of blocks from NYSE computers. Just having your server farm uptown instead of downtown costs numerous microseconds, giving the ones closer a significant advantage.
Of course, now that we're past the bubble, this is just technology implementing the curve smoothing to higher and higher precision, making the market ever closer to the pure physics of the Efficient Market - which (IMHO) is basically an ocean in a large number of dimensions, with Brownian motion, ripples, waves bigger waves all generated by events large and small globally and interacting according to the physics of fluid dynamics.
Funny you should mention this. The very first hack that I ever heard of, back in the early 1970s (yes, I'm that old) was purportedly done to what is now Citibank in the mid-late 1960s. This was when they had one of the first IBM 360 machines - the standard mainframe (about the same power as a PC/AT). The guy who wrote their loan computation program wrote self-modifying code. When the program ran, it stored various constants in various places in memory, that happened (when these constants were all strung together in sequential memory) to be a hack. Then all he had to do was jump to the start of that routine, and it modified the interest computation function used by every part of the system and some other things, and ... You have to realize that self-modifying code was still quite common at that time so nothing he did looked out of the ordinary.
The result of all this was that when interest payments were computed on every account (savings, checking, loans, etc), instead of rounding at 1/10 mil (1/1000 of a dollar), it truncated, and took the difference and put it into an account that only existed inside the machine. At the end of the day, the money in that account was quietly pushed into a transfer that sent the money by wire to a Swiss account, and the account disappeared.
The programmer, having finished his job, went on to other things. This hack supposedly went unnoticed for a couple of years, until for unrelated reasons the bank had to do a line-by-line audit of their accounts, and came up several million dollars short. They spent months trying to figure out where the money went but couldn't find it. Finally they went to the guy and said, "you can keep the money and we won't prosecute. Just tell us what you did so we can prevent it happening again." By then he had made something under $10 million.
I have no idea if this is true, or a very early urban legend. But I heard it when I was in school in the early 1970s, so it's at least that old.
I always thought they were Demoplicans. (or was it Demorons??)
In all fairness, the folks who succeed in becoming high level politicians got there by signing on to the establishment myths that prevailed when they were in their formative years - few of them got there by thinking for themselves. Those myths still drive their reasoning, and those myths still control most of the people in most of the think tanks and lobbying groups. So it's understandable that they continue to try the same things that their predecessors did. I expect a grand total of 1% of Washington insiders of all stripes have ever actually read Keynes in the original - where, among other things, he pointed out that deficit spending could be successful *for a short time* as a measure to get money flowing again - he never implied that it should be a permanent fixture in the economic landscape.
Highly recommended, these two books (get them as audiobooks if you want hours of driving amusement, education and horror). Just for perspective (one that is not in these books, or anywhere else that I know of - I thunk it up with my own little brain), this whole scenario can be seen as a classic tech bubble - the technology being the application of algorithms and various other forms of computer technology to technical stock trading. This bubble began back in the early 1970s, and happened to hit critical mass just when everything else was happening - the housing bubble, etc. Since this tech bubble lived in the second order space of the financial economy that sits on top of the first-order "real-world" economy, it got boosted out of sight by the multiplier of the two.
The Quants: How a New Breed of Math Whizzes Conquered Wall Street and Nearly Destroyed It /. gets a nickel if you actually buy a copy from Amazon by linking from here: The Quants ... ...
In case
- it all started with the guy who figured out how to make card counting work in Vegas and Atlantic City
and
Too Big to Fail: The Inside Story of How Wall Street and Washington Fought to Save the Financial System---and Themselves
Too Big to Fail
This. Also,
"Your website has a bug."
"What platform and browser are you running?"
"Ubuntu 11.04, Firefox - I don't know what version, it's the latest."
"Thanks, we'll get right on that - after the next millenium. Wanker."
Actually I don't like Chrome, and I am hoping that things just keep working in FF without constant fiddling. And I _really_ don't like things breaking the several add-ons that I make use of constantly.
I just imagined my car being on a rapid-release cycle, and the various kinds of bad things that could happen on the way home from work as the latest update accidentally disabled the brakes, or set the radio to play rap at full volume (not sure which is worse, actually). Eeek!
Almost as bad as 'massive' - a massively overused adjective these days.
Which makes spotting the dangerous ones an interesting problem, as relative motion is the best& easiest way to detect things.
I've thought about this (back when I rode the bus to work) - in fact I considered finding out the total amount of fuel used by the buses in Portland OR (where I was riding), and the total ridership. This seems to me to be a simple computation - divide the fuel by the number of passengers, and you get the gallons per trip. The agency probably also has a good idea of the length of the average trip (but that is also complicated by the fact that the bus does not go from the passenger's point A to point B - it goes along its route, which may twice as long as the direct route.)
The reason I got interested was that the buses in Portland had stickers on the back saying, "This bus takes 152 cars off the road", which I figured was probably a self-serving and specious statement. IIRC the average bus gets something like 3 MPG in actual use, and the average ridership is on the order of 10 (obviously commute-time trips are more efficient, midnight runs with empty buses cost a lot.) So this means the bus is getting about 30 passenger miles per gallon average. Considering the routing inefficiency, things look worse.
To make things even worse, an impromptu experiment some years ago occurred due to a bus strike in London. The average speed of traffic almost doubled.
The ultimate solution would be to re-institute the 'trolley towns' that used to grow up around each trolley (light rail) station, before everyone had cars. Then trolleys could run at optimal efficiency with minimum stops. Trolleys running on their own lanes or rails are the only way that mass transit can improve their speed relative to driving.
Thorium by itself can never go critical, no matter how much of it you put together.
It's fission. I haven't read the article but several folks who did said that it mentions 'subcritical' fission - IOW the reaction does not sustain itself, it must be fed extra neutrons, which are generated by a laser-driven particle accelerator.
Oddly enough, my dad built the engine plants for the Atomic Airplane project, back in the 1950s. The project was snakebit from the start. The buildings are now part of the Idaho National Laboratory. He got stiffed for $400,000 by the US Government - in 1954 that was not pocket change..
IIRC the output of proposed 1000 MW Liquid Thorium Fluoride Reactors (power-station-scale) is non-radioactive nickel and boron, or some such. The amount of radioactive waste for such a plant is expected to be less than a teacup per year. None of it is bomb-making material. As a continuous ambient-pressure liquid system, it's much safer and easier to control than uranium-fission plants. So for power plants, it's almost ideal. IMHO if we had gone to LTFR back in the late 1960s (as we could have, but AEC lost interest because LTFRs don't produce bomb materials!) by now the prospect of switching from 'safe, efficient, scalable LTFRs to uranium power plants would be unthinkable.
1 hp = 786 w iirc :) about .75 kw in other words - 30 kw is about equal to 40 hp.
I don't know about the system in the article, but the LTFR (Liquid Thorium Fluoride Reactor - a power station-scale reactor technology) runs at ambient pressure. And the amount of nuclear waste is something like a teacup per year, for a 1000 MW station.
I haven't read the article (probably won't) but there is legitimate research into using a laser to drive a particle accelerator, which provides the slow neutrons necessary to trigger the Thorium fission chain reaction.
I haven't read the article (and probably won't) but it's worth your time to look up Liquid Fluoride Thorium Reactors. (I always get LFTR and LTFR confused - sorry). For actual power plants I think LTFR are probably the best solution to all the problems with uranium-fueled reactors. The biggest force preventing them may turn out to be the present big suppliers of uranium fuel rods - a very lucrative industry that would be destroyed. Switching power to LTFR would provide sufficient power to do almost anything we want to do for perhaps 10,000 years, by which time I would hope that we would have figured out how to tap the vacuum energy! :D
AFAIK that's not true of diesels. Very large diesels can be much more efficient than small ones. But I'm too lazy to look it up. :)
Indeed - container ships would be a great application. I've read that ocean shipping is now producing some huge fraction of all the CO2. And their existing fuel tanks and engines are probably larger than an LTFR reactor that would produce the required 200 KW. So, good idea.