A better way to put this IMHO is that automobile transportation just isn't scalable enough.
I came to this realization that this is really a scale problem a few months ago. I generally citibike (bike share) to work, and was feeling pretty smug about it for a period of time. I still use mass transit on occasion though, and one time I was on a packed train and some dude comes on with a bike. It took up the space of about 4, maybe 5 people. Biking, while far better than a car-which I also noticed that 5 bikes took up the space of 1 car, while taking up a space of approximately 5 cars, which clogged 3 lanes for almost the entirety of 2 rows, still just did not have anywhere near the density or scale of mass transit. Getting the entirety of the population on a bike isn't going to work either (at least not for megascale cities like NYC). A typical subway train can and does, fit about 1200 people on it during rush hour. Next time you are stuck in traffic (or just look on google images for pictures of traffic jams: https://daily.jstor.org/wp-con... ) try and count the number of cars you can see in front of you. For me it was about 200-250. At least when I was driving, it was safe to assume that most of these were only singly occupied, some had 2, a few had more in them, but the vast majority of these cars had only a driver. That means that this massively expensive piece of infrastructure, for as far as the eye could see, still did not have the capacity of a single NYC subway train. That blew my mind a bit.
I am not saying that we need to ban cars, or bikes, but we certainly need to encouraging and *improving* mass transit.
I was really hoping they would go to Newark, NJ. You still get access to the NYC talent market. Newark is a town that always seems on the cusp of being reborn, but its been rocky at best. If anything would push it forward, Amazon would. There is tons of under utilized space available there (though I heard they wanted 500k sq ft, newly built and ready to occupy ASAP). I don't know long term how many people would commute from NYC to Newark, but the trains are pretty much empty going in the opposite direction towards Newark out of NYC, this could help even out the commute and ease a bit of the stress on the transit system. Its much more accessible to those wanting to live in both an urban or suburban location with a major major airport nearby, with commercial and residential real estate being much cheaper.
Its unfortunate and I believe a real missed opportunity for both Amazon and the region. Instead you are cramming in more density on an already overcrowded transit infrastructure in a location that while slightly cheaper than Manhattan proper, is still very expensive and you are effectively land locked there.
I believe the key is that you didn't play it when it came out... Quake was the first truly 3d shooter, and came out 3 years later. Myst for the time was stunning visually, and personally I found the story/mysteriousness immersive. By 1998 it was definitely old hat and felt restrictive/primitive.
The puzzles... hard to say. I was 14 when I first played? I found them challenging without being so absurdly so that I had to resort to looking in a cheat manual in a game store or spend some of my few precious minutes of internet time looking them up on an aol board. After Myst I played a bunch of others in the genre, Phantasmagoria, Shivers, some had puzzles that were quite obscure, and that wasn't fun either.
This probably depends a lot on the airport, but I traveled over 100k miles last year, but checking vs not checking can be the difference of an hour+ in your total travel time. I have been forced to gate check going to Newark and literally waited an hour for the carousel to start moving. This is not insignifcant. This is (another) missed dinner with your family, or an hour of lost sleep or decompression time. This is an hour later to the office, which may be a big problem depending on whether you have an important meeting to attend.
Newark is an extreme case, but even waiting an extra 30 minutes when you are a frequent business traveler is really irritating (think about when you are doing 30+ flights a year... for work...)
I am teaching this curriculum right now (as a volunteer for https://www.tealsk12.org/ ) and I can tell you that 90% of our time is spent teaching coding, either HTML or CSS. We wrap the fluff stuff into the curriculum to break up the harder stuff.
The point of this class is to teach basic programming skills and computer concepts so that the regular AP CS class in Java (which I also teach) isn't such a culture shock. Kids drop out of that class at a very high rate and they are trying to give a sanctioned path to learning the real stuff. The regular class was very similar to my first semester CS class in content.
There are a lot of criticisms around Tesla as a company, or at least a for-profit company, but having bonds trade at "95 cents on the dollar" is not one of them. Bond prices fluctuate all the time just like the stock market, and frequently go above and below par (IE 100 cents on the dollar). Bond prices are based around risk for a given level of yield. GM's bonds have gone well above and below par as well: http://finra-markets.morningst...
The effective yield on TSLA bonds is 5.5%, this is nowhere near any sort of panic level. And "wiping out their yield" doesn't make sense either- the yield is the yield and based on what you purchased the bond for. If you are saying that the bonds lost 5% of their value according to their resale price, that is still only one years worth of yield, but otherwise this is just a blip in the market for buy and hold purchasers of bonds.
Remember when "friends" and "foes" on slashdot were a thing people paid attention to? We have been friends for somewhere around ~15 years now and I have no idea why:)
There was very little market for that kind of power as well. I was doing lots of IOT stuff 2-3 years back (before I ran out of potential projects), and to me the ideal way to go was to use super cheap chips like the esp8266 to gather sensor data and do the simple logic stuff (Ex: If soil moisture x, write to pin 12 for 30 seconds, which turns on a solenoid that lets waterflow), but then send that data to a "real" computer, probably in the cloud, to do any real processing on it. The only place I saw use for a Galileo type product was for when there was no internet available, yet you were in a location that you could not or would not want to set up a laptop. Drone projects that have to process gyroscopical, GPS, and perhaps video data in near real time for example- but the applications were relatively few.
They aren't the first to miss the mark here- For awhile I was playing with a Tessel, which was also about $80, which runs node.js natively, and it was processing data and serving a web interface that was visualizing the sensor data, but it was slow, and had other drawbacks like dealing with dynamic IPs and the wifi on the first gen was so flakey it was almost unusable, but really in the end what landed it in the dust bin was that it made tons more sense to just buy super cheap chips to collect that sensor data and store it somewhere- the whole system was a lot more robust and usable that way.
Not sure why you think that? The Very Large Company I work for pays myself and my team very well to do node.js. Maybe they are rare in that they pay based on the type of work they do (building highly scalable services) instead of just by the technology they work in.
I really asked myself why I was even bothering with the series when I decided I was just going to read online chapter summaries for books 7-10, and even then I found the pace to be glacially slow.
You seem to have a lot of animosity- did you post in this thread to just have a circlejerk about how obviously bad HFT is? Or have a reasoned and informed discussion?
Maybe my reading comprehension is off- here is your second sentence: And none of them benefited society in any respectable proportion to what they earned. So why should society infrastructure be modified to suit them (exclusive order types on exchanges regulated of necessity) ?
You don't even know how much they earned, so how can you really comment on it? Virtu Financial recently filed an S-1 to go public: https://www.sec.gov/Archives/e... Total revenue: $664 million. Net revenue: $182 million. Not bad, but not exactly killing it either.
What are you talking about with "society infrastructure", and particularly "exchanges regulated of necessity?" What does that even mean? Did you know that NYSE, amongst many other liquidity venues, is now a publicly traded company? The exchanges have provided these order types of their own volition, this isn't an "HFT" problem, its an exchange problem if anything- they are trying to attract the HFT flow to their exchanges!
As for "off the shelf" and "more difficult", it is nuanced. There are many components that are required to build a trading engine, major pieces of them can now be bought- low latency market data (Exegy), low latency network cards (Mellanox/Solarflare), and exchange connectivity, for example. There is now a critical mass of developers who can build this stuff for you, as opposed to this being arcane research type stuff. However, its all rather expensive. Co-location itself will cost you $10k/month per rack last I looked into it. Hence the "high barrier to entry" and "off the shelf" go together. 15-20 years ago, a boiler-room type phones and brokers operation might only have startup overhead (outside of employees) in total of $10k per month. The costs of some of this stuff will come down as it becomes commoditized, but bandwidth and datacenter space are likely to remain a sparse resource and remain costly.
As for whether this new technology is benefitting anyone, I would argue this is just a luddite argument that has been made many times before whenever there has been a disruptive new technology. Do you think the guys on the floor of the NYSE used to pay hundreds of thousands of dollars for their seat because they liked going to Champs Deli or because they could wear a funny colored blazer? They did it so they could trade on information first. Telephones disrupted the bucket shops, SOES bandits disrupted the floor traders, later electronic trading completely disrupted floor trading, and now we have a bunch of guys who realized that they could build much faster infra and make money off it, and they did, forcing others to beef up their systems to keep up. And most have.
I am not spouting off anything- I spent the last ten years building this stuff on both the HFT and Agency side. You read a few articles, and maybe the entire Flash Boys book? Good for you. I am trying to give you the rest of the story.
You don't understand why dark pools were created in the first place. They originally came about so that institutional players, mutual funds, pensions, hedge funds, etc, could move large amounts of stock without tipping their hand. If Vanguard has its hand tipped that its selling all of it shares of say IBM in a large fund, that stock is going to plummet, and Vanguard is going to take a loss. This is called market impact. So dark pools were created where these large buyers and sellers could come together and trade large blocks of stock at once, and in the process they would also eliminate exchange fees and trade at the midpoint instead of eating the spread.
This was loved by the big players as they could easily move into and out of big positions- meaning that your pensions and mutual funds are now getting better prices and aren't getting taken advantage of in the lit markets.
The "price" is still the price for everyone, and you can not trade at a worse price than the NBBO- National Best Bid Offer. An HFT participant would have a field day if they were able to sniff out that large movements were going down. Many traders did this form or "tape reading" in the Bad Old Days you seem to be want to go back to.
This is not true, at least not for B/D owned dark pools. You could not buy real colocated access to those venues. You can buy other products to get low latency links into brokers, but you can't buy a direct route into a dark pool, at least not one owned by any of the b/d's I worked at, but I am unaware of any other B/Ds doing this either.
They just transferred profits from one group of guys that were making fairly easy money, collapsed the profit margin and concentrated that money into a fairly smaller group of people- so I think that they actually did benefit society. Though I would ask you why all actions have to benefit society? Does a gambler going to a casino benefit society? Does someone who goes to a restauarant? How about about someone who takes a weekend trip away from the city?
I am not sure what you are getting at about raising the barriers to entry. Technology has in general raised the barriers to entry for opening up a brokerage firm. It used to be that you needed phones, sales reps, a clearing firm and someone down on the floor to open a brokerage, and that was about it. These days everything is electronic and it is more difficult to do so. That technology needs to be reasonably good as well. To be competitive these days, the technology is commoditized, you can buy many pieces off the shelf.
The window for the glory days of HFT has closed. People caught on to them, and caught up. I wouldn't get too caught up in the storyline that a few geniuses almost took over the world. In 2008, I heard of several firms doing really stupid slow things- database accesses of security information in the critical path for instance.
This reminds me of when I was building a new stock trading system from scratch. We wanted to provide a web interface so you could enter orders, view your trades, etc. This wasn't considered a core thing, so it was outsourced to some guys in Brazil. These guys were smart, but knew nothing about our industry and clearly hadn't ever traded a stock. We gave them some mockups and specs, and what they gave us back looked just like what we asked for, but it was a usability nightmare.
The order and execution screens were in a random order, and there was no way to sort the records. Basic validations were not done- entering a negative number of shares or price was not caught for example. Summaries of numbers of shares, value executed, etc, were not there. There were just a bunch of things like this, that made the app a complete disaster to use.
But I realized after talking with these guys, it was entirely our fault. They really had no idea what they were building, how it was used, or who would use it. But they were amazing soldiers, I could have told them to walk in a straight line, and watched those guys just get up and walk, straight into a wall, and continue to do so until I told them to stop. Things like being able to sort the records, were so obvious to us that we didn't even think to put them in the spec.
You completely misunderstood Flash Boys when you read it. Diverting an order into a dark pool is meant to HIDE information from HFT guys. You don't know where the order came from, or who your counterparty is. You can't view the order book, its much more hidden (hence the name "dark" pool, as opposed to the "lit" exchanges). The only way to see if there is any activity in a given name is to actually send an order in and hope it gets executed.
What do you mean that this gives HFT guys more information that others don't have? They can try pinging dark pools and try to make guesses about the size that is actually there, but this not really information that others don't have. Around 2008-2009, it was becoming known that some participants were trying to sniff out size in dark pools, and most pools put in mechanisms to prevent this.
Dark pools exist as a way for brokers to make themselves money by keeping the exchange fees.
Source: I have built dark pools and routers that route flow into them.
Are you just some fatass in a chair bitching? Or do you have a real solid, informed complaint about HFT?
I ask this, as I worked in the algo trading industry for ten years on both the prop (HFT) side and the agency side, where I built tools to counteract HFT players. I recently left as the money dried up, while the hours didn't, and to be quite honest I just lost a lot of interest in the business.
There are problems with HFT activity, but I believe that overall, they have benefitted the retail investor. Since the rise of HFT and electronic markets, spreads have collapsed to be an insignificant cost of trading. HFT guys ate the lunch of market makers who used to have cushy little businesses and traders getting mid 6 to 7 figure bonuses. My first job involved automating those guys out of a job. Those guys used to legitimately front-run orders, anyone talking about HFT front running is either redefining the term, or doesn't know what they are talking about.
Guys with a speed advantage have always used that advantage to make money in the stock market. Whether it be guys with faster horses in the pre-railroad/telegraph era (supposedly the rothschilds made their fortune this way, buying up english bonds as they had news that a war had ended first), telephones ripping off bucket shops in the 1900's, SOES bandits in the 1980s, and now HFT today, this has always existed. All those guys who actually used to sit on the floor of the NYSE- why do you think they were there?- So they could trade on the news first (one quote from the book Market Wizards: "First its the floor traders, the next day its the dentists, then after that comes Joe Schmoe.")
The games that HFT guys are playing is generally sniping a penny here and there. As a retail investor who is buying and holding, their game has nothing to do with yours, and they have eaten the lunches of the market makers and brokers who used to rip you off.
Are there problems with HFT? Yeah- mainly that exchanges are developing order types exclusively for their use. The fact that they are acting like market makers by providing liquidity and squeezing the legit market makers, but once things start looking weird, pull out immediately (though after the flash crash, many of these guys started becoming legit market makers).
Net/net though, these guys are good for retail traders. If you disagree, come up with a good, specific, informed reason on how they are hurting you and your orders in the market. If you look at some of the major detractors of HFT like Joe Saluzzi, they are almost always from smaller niche firms who can not afford technology to adequately compete in an electronic world, and are getting squeezed out by the bulge bracket guys.
The HFT business is drying up as it is though. The arms race has put enough players on equal footing that the low hanging fruit is gone. The major banks have invested enough in their infra that they can't just be picked off by these guys anymore. This is good for the industry in my opinion, maybe the focus can go back to trading smarter, not just saving off ten microseconds on the slice time.
You can trade in London, because it is based in London. Just like you can buy a car from a german car company, but once it is on US soil, it needs to comply with US standards.
InTrade is acting as a futures market in the US, so it needs to be regulated by the SEC if it wants to continue to do so. I am not intimately familiar w/ the situation, though I have used InTrade in the past, but I don't see what is stopping them from becoming a legit exchange. I think it will help their business a great deal.
Imagine a convenience store that initially sells some cans of lighter fluid. Then it starts selling some gasoline in soda-like cans intended for lawn mowers. Then it starts selling larger cans that are primarily used to put into car tanks. The government is saying "Hey now, you are really starting to look like a gas station, and need to follow the rules and regulations that go along with being a gas station, and you can't sell any more gas until you do."
This isn't mission control, we are talking about a few blogs here. The security policies I am avoiding are intended to prevent data being sent out of the firm undetected, and I am not doing that, and I don't have the ability to do that in any reasonable way.
I don't consider loading up a web browser from a friend's house or wherever I am in an outage situation to log into my server "beyond stupid." I wouldn't even consider it mildly stupid, and its certainly not something to flame someone about.
It is a lot easier to ask for forgiveness than permission.
I run a few websites that generate a small but growing amount of traffic (and hopefully one day, revenue), so its also nice to know that from any computer, I can log in to my jump server, and take a look if something goes down, with nothing more than a web browser. It could be a locked down computer only offering a web browser, or a friend's computer I don't want to be fumbling around and installing putty on, etc.
Yeah there are. Most are pretty crappy. GateOne just performs far better, to the point where you can actually use it for useful work. I used to use ajaxterm, it worked, but had weird bugs. Later I moved to webshell, which was a big improvement. GateOne is an improvement further still. I use this on my personal server, so unless they are inspecting packets to see that this is an app, which currently my very paranoid company does not do, this won't get blocked.
No, its not. It runs in a browser and is thus cross platform. The main point is that the ssh session runs on the server side, so you are essentially tunneling ssh through http. Very nice if you need ssh access in an environment that won't let you ssh out of the network.
Slashdot Mirror
A better way to put this IMHO is that automobile transportation just isn't scalable enough.
I came to this realization that this is really a scale problem a few months ago. I generally citibike (bike share) to work, and was feeling pretty smug about it for a period of time. I still use mass transit on occasion though, and one time I was on a packed train and some dude comes on with a bike. It took up the space of about 4, maybe 5 people. Biking, while far better than a car-which I also noticed that 5 bikes took up the space of 1 car, while taking up a space of approximately 5 cars, which clogged 3 lanes for almost the entirety of 2 rows, still just did not have anywhere near the density or scale of mass transit. Getting the entirety of the population on a bike isn't going to work either (at least not for megascale cities like NYC). A typical subway train can and does, fit about 1200 people on it during rush hour. Next time you are stuck in traffic (or just look on google images for pictures of traffic jams: https://daily.jstor.org/wp-con... ) try and count the number of cars you can see in front of you. For me it was about 200-250. At least when I was driving, it was safe to assume that most of these were only singly occupied, some had 2, a few had more in them, but the vast majority of these cars had only a driver. That means that this massively expensive piece of infrastructure, for as far as the eye could see, still did not have the capacity of a single NYC subway train. That blew my mind a bit.
I am not saying that we need to ban cars, or bikes, but we certainly need to encouraging and *improving* mass transit.
I was really hoping they would go to Newark, NJ. You still get access to the NYC talent market. Newark is a town that always seems on the cusp of being reborn, but its been rocky at best. If anything would push it forward, Amazon would. There is tons of under utilized space available there (though I heard they wanted 500k sq ft, newly built and ready to occupy ASAP). I don't know long term how many people would commute from NYC to Newark, but the trains are pretty much empty going in the opposite direction towards Newark out of NYC, this could help even out the commute and ease a bit of the stress on the transit system. Its much more accessible to those wanting to live in both an urban or suburban location with a major major airport nearby, with commercial and residential real estate being much cheaper.
Its unfortunate and I believe a real missed opportunity for both Amazon and the region. Instead you are cramming in more density on an already overcrowded transit infrastructure in a location that while slightly cheaper than Manhattan proper, is still very expensive and you are effectively land locked there.
I believe the key is that you didn't play it when it came out... Quake was the first truly 3d shooter, and came out 3 years later. Myst for the time was stunning visually, and personally I found the story/mysteriousness immersive. By 1998 it was definitely old hat and felt restrictive/primitive.
The puzzles... hard to say. I was 14 when I first played? I found them challenging without being so absurdly so that I had to resort to looking in a cheat manual in a game store or spend some of my few precious minutes of internet time looking them up on an aol board. After Myst I played a bunch of others in the genre, Phantasmagoria, Shivers, some had puzzles that were quite obscure, and that wasn't fun either.
This probably depends a lot on the airport, but I traveled over 100k miles last year, but checking vs not checking can be the difference of an hour+ in your total travel time. I have been forced to gate check going to Newark and literally waited an hour for the carousel to start moving. This is not insignifcant. This is (another) missed dinner with your family, or an hour of lost sleep or decompression time. This is an hour later to the office, which may be a big problem depending on whether you have an important meeting to attend.
Newark is an extreme case, but even waiting an extra 30 minutes when you are a frequent business traveler is really irritating (think about when you are doing 30+ flights a year... for work...)
I am teaching this curriculum right now (as a volunteer for https://www.tealsk12.org/ ) and I can tell you that 90% of our time is spent teaching coding, either HTML or CSS. We wrap the fluff stuff into the curriculum to break up the harder stuff.
The point of this class is to teach basic programming skills and computer concepts so that the regular AP CS class in Java (which I also teach) isn't such a culture shock. Kids drop out of that class at a very high rate and they are trying to give a sanctioned path to learning the real stuff. The regular class was very similar to my first semester CS class in content.
There are a lot of criticisms around Tesla as a company, or at least a for-profit company, but having bonds trade at "95 cents on the dollar" is not one of them. Bond prices fluctuate all the time just like the stock market, and frequently go above and below par (IE 100 cents on the dollar). Bond prices are based around risk for a given level of yield. GM's bonds have gone well above and below par as well: http://finra-markets.morningst...
The effective yield on TSLA bonds is 5.5%, this is nowhere near any sort of panic level. And "wiping out their yield" doesn't make sense either- the yield is the yield and based on what you purchased the bond for. If you are saying that the bonds lost 5% of their value according to their resale price, that is still only one years worth of yield, but otherwise this is just a blip in the market for buy and hold purchasers of bonds.
Remember when "friends" and "foes" on slashdot were a thing people paid attention to? We have been friends for somewhere around ~15 years now and I have no idea why :)
There was very little market for that kind of power as well. I was doing lots of IOT stuff 2-3 years back (before I ran out of potential projects), and to me the ideal way to go was to use super cheap chips like the esp8266 to gather sensor data and do the simple logic stuff (Ex: If soil moisture x, write to pin 12 for 30 seconds, which turns on a solenoid that lets waterflow), but then send that data to a "real" computer, probably in the cloud, to do any real processing on it. The only place I saw use for a Galileo type product was for when there was no internet available, yet you were in a location that you could not or would not want to set up a laptop. Drone projects that have to process gyroscopical, GPS, and perhaps video data in near real time for example- but the applications were relatively few.
They aren't the first to miss the mark here- For awhile I was playing with a Tessel, which was also about $80, which runs node.js natively, and it was processing data and serving a web interface that was visualizing the sensor data, but it was slow, and had other drawbacks like dealing with dynamic IPs and the wifi on the first gen was so flakey it was almost unusable, but really in the end what landed it in the dust bin was that it made tons more sense to just buy super cheap chips to collect that sensor data and store it somewhere- the whole system was a lot more robust and usable that way.
Not sure why you think that? The Very Large Company I work for pays myself and my team very well to do node.js. Maybe they are rare in that they pay based on the type of work they do (building highly scalable services) instead of just by the technology they work in.
There is in my company. Seriously. A very very big e-commerce player. With all the politics and careers at stake that you would expect.
I really asked myself why I was even bothering with the series when I decided I was just going to read online chapter summaries for books 7-10, and even then I found the pace to be glacially slow.
You seem to have a lot of animosity- did you post in this thread to just have a circlejerk about how obviously bad HFT is? Or have a reasoned and informed discussion?
Maybe my reading comprehension is off- here is your second sentence: And none of them benefited society in any respectable proportion to what they earned. So why should society infrastructure be modified to suit them (exclusive order types on exchanges regulated of necessity) ?
You don't even know how much they earned, so how can you really comment on it? Virtu Financial recently filed an S-1 to go public: https://www.sec.gov/Archives/e... Total revenue: $664 million. Net revenue: $182 million. Not bad, but not exactly killing it either.
What are you talking about with "society infrastructure", and particularly "exchanges regulated of necessity?" What does that even mean? Did you know that NYSE, amongst many other liquidity venues, is now a publicly traded company? The exchanges have provided these order types of their own volition, this isn't an "HFT" problem, its an exchange problem if anything- they are trying to attract the HFT flow to their exchanges!
As for "off the shelf" and "more difficult", it is nuanced. There are many components that are required to build a trading engine, major pieces of them can now be bought- low latency market data (Exegy), low latency network cards (Mellanox/Solarflare), and exchange connectivity, for example. There is now a critical mass of developers who can build this stuff for you, as opposed to this being arcane research type stuff. However, its all rather expensive. Co-location itself will cost you $10k/month per rack last I looked into it. Hence the "high barrier to entry" and "off the shelf" go together. 15-20 years ago, a boiler-room type phones and brokers operation might only have startup overhead (outside of employees) in total of $10k per month. The costs of some of this stuff will come down as it becomes commoditized, but bandwidth and datacenter space are likely to remain a sparse resource and remain costly.
As for whether this new technology is benefitting anyone, I would argue this is just a luddite argument that has been made many times before whenever there has been a disruptive new technology. Do you think the guys on the floor of the NYSE used to pay hundreds of thousands of dollars for their seat because they liked going to Champs Deli or because they could wear a funny colored blazer? They did it so they could trade on information first. Telephones disrupted the bucket shops, SOES bandits disrupted the floor traders, later electronic trading completely disrupted floor trading, and now we have a bunch of guys who realized that they could build much faster infra and make money off it, and they did, forcing others to beef up their systems to keep up. And most have.
I am not spouting off anything- I spent the last ten years building this stuff on both the HFT and Agency side. You read a few articles, and maybe the entire Flash Boys book? Good for you. I am trying to give you the rest of the story.
You don't understand why dark pools were created in the first place. They originally came about so that institutional players, mutual funds, pensions, hedge funds, etc, could move large amounts of stock without tipping their hand. If Vanguard has its hand tipped that its selling all of it shares of say IBM in a large fund, that stock is going to plummet, and Vanguard is going to take a loss. This is called market impact. So dark pools were created where these large buyers and sellers could come together and trade large blocks of stock at once, and in the process they would also eliminate exchange fees and trade at the midpoint instead of eating the spread.
This was loved by the big players as they could easily move into and out of big positions- meaning that your pensions and mutual funds are now getting better prices and aren't getting taken advantage of in the lit markets.
The "price" is still the price for everyone, and you can not trade at a worse price than the NBBO- National Best Bid Offer. An HFT participant would have a field day if they were able to sniff out that large movements were going down. Many traders did this form or "tape reading" in the Bad Old Days you seem to be want to go back to.
This is not true, at least not for B/D owned dark pools. You could not buy real colocated access to those venues. You can buy other products to get low latency links into brokers, but you can't buy a direct route into a dark pool, at least not one owned by any of the b/d's I worked at, but I am unaware of any other B/Ds doing this either.
They just transferred profits from one group of guys that were making fairly easy money, collapsed the profit margin and concentrated that money into a fairly smaller group of people- so I think that they actually did benefit society. Though I would ask you why all actions have to benefit society? Does a gambler going to a casino benefit society? Does someone who goes to a restauarant? How about about someone who takes a weekend trip away from the city?
I am not sure what you are getting at about raising the barriers to entry. Technology has in general raised the barriers to entry for opening up a brokerage firm. It used to be that you needed phones, sales reps, a clearing firm and someone down on the floor to open a brokerage, and that was about it. These days everything is electronic and it is more difficult to do so. That technology needs to be reasonably good as well. To be competitive these days, the technology is commoditized, you can buy many pieces off the shelf.
The window for the glory days of HFT has closed. People caught on to them, and caught up. I wouldn't get too caught up in the storyline that a few geniuses almost took over the world. In 2008, I heard of several firms doing really stupid slow things- database accesses of security information in the critical path for instance.
This reminds me of when I was building a new stock trading system from scratch. We wanted to provide a web interface so you could enter orders, view your trades, etc. This wasn't considered a core thing, so it was outsourced to some guys in Brazil. These guys were smart, but knew nothing about our industry and clearly hadn't ever traded a stock. We gave them some mockups and specs, and what they gave us back looked just like what we asked for, but it was a usability nightmare.
The order and execution screens were in a random order, and there was no way to sort the records. Basic validations were not done- entering a negative number of shares or price was not caught for example. Summaries of numbers of shares, value executed, etc, were not there. There were just a bunch of things like this, that made the app a complete disaster to use.
But I realized after talking with these guys, it was entirely our fault. They really had no idea what they were building, how it was used, or who would use it. But they were amazing soldiers, I could have told them to walk in a straight line, and watched those guys just get up and walk, straight into a wall, and continue to do so until I told them to stop. Things like being able to sort the records, were so obvious to us that we didn't even think to put them in the spec.
You completely misunderstood Flash Boys when you read it. Diverting an order into a dark pool is meant to HIDE information from HFT guys. You don't know where the order came from, or who your counterparty is. You can't view the order book, its much more hidden (hence the name "dark" pool, as opposed to the "lit" exchanges). The only way to see if there is any activity in a given name is to actually send an order in and hope it gets executed.
What do you mean that this gives HFT guys more information that others don't have? They can try pinging dark pools and try to make guesses about the size that is actually there, but this not really information that others don't have. Around 2008-2009, it was becoming known that some participants were trying to sniff out size in dark pools, and most pools put in mechanisms to prevent this.
Dark pools exist as a way for brokers to make themselves money by keeping the exchange fees.
Source: I have built dark pools and routers that route flow into them.
Are you just some fatass in a chair bitching? Or do you have a real solid, informed complaint about HFT?
I ask this, as I worked in the algo trading industry for ten years on both the prop (HFT) side and the agency side, where I built tools to counteract HFT players. I recently left as the money dried up, while the hours didn't, and to be quite honest I just lost a lot of interest in the business.
There are problems with HFT activity, but I believe that overall, they have benefitted the retail investor. Since the rise of HFT and electronic markets, spreads have collapsed to be an insignificant cost of trading. HFT guys ate the lunch of market makers who used to have cushy little businesses and traders getting mid 6 to 7 figure bonuses. My first job involved automating those guys out of a job. Those guys used to legitimately front-run orders, anyone talking about HFT front running is either redefining the term, or doesn't know what they are talking about.
Guys with a speed advantage have always used that advantage to make money in the stock market. Whether it be guys with faster horses in the pre-railroad/telegraph era (supposedly the rothschilds made their fortune this way, buying up english bonds as they had news that a war had ended first), telephones ripping off bucket shops in the 1900's, SOES bandits in the 1980s, and now HFT today, this has always existed. All those guys who actually used to sit on the floor of the NYSE- why do you think they were there?- So they could trade on the news first (one quote from the book Market Wizards: "First its the floor traders, the next day its the dentists, then after that comes Joe Schmoe.")
The games that HFT guys are playing is generally sniping a penny here and there. As a retail investor who is buying and holding, their game has nothing to do with yours, and they have eaten the lunches of the market makers and brokers who used to rip you off.
Are there problems with HFT? Yeah- mainly that exchanges are developing order types exclusively for their use. The fact that they are acting like market makers by providing liquidity and squeezing the legit market makers, but once things start looking weird, pull out immediately (though after the flash crash, many of these guys started becoming legit market makers).
Net/net though, these guys are good for retail traders. If you disagree, come up with a good, specific, informed reason on how they are hurting you and your orders in the market. If you look at some of the major detractors of HFT like Joe Saluzzi, they are almost always from smaller niche firms who can not afford technology to adequately compete in an electronic world, and are getting squeezed out by the bulge bracket guys.
The HFT business is drying up as it is though. The arms race has put enough players on equal footing that the low hanging fruit is gone. The major banks have invested enough in their infra that they can't just be picked off by these guys anymore. This is good for the industry in my opinion, maybe the focus can go back to trading smarter, not just saving off ten microseconds on the slice time.
You can trade in London, because it is based in London. Just like you can buy a car from a german car company, but once it is on US soil, it needs to comply with US standards.
InTrade is acting as a futures market in the US, so it needs to be regulated by the SEC if it wants to continue to do so. I am not intimately familiar w/ the situation, though I have used InTrade in the past, but I don't see what is stopping them from becoming a legit exchange. I think it will help their business a great deal.
Imagine a convenience store that initially sells some cans of lighter fluid. Then it starts selling some gasoline in soda-like cans intended for lawn mowers. Then it starts selling larger cans that are primarily used to put into car tanks. The government is saying "Hey now, you are really starting to look like a gas station, and need to follow the rules and regulations that go along with being a gas station, and you can't sell any more gas until you do."
Very good to know. Thanks!
This isn't mission control, we are talking about a few blogs here. The security policies I am avoiding are intended to prevent data being sent out of the firm undetected, and I am not doing that, and I don't have the ability to do that in any reasonable way.
I don't consider loading up a web browser from a friend's house or wherever I am in an outage situation to log into my server "beyond stupid." I wouldn't even consider it mildly stupid, and its certainly not something to flame someone about.
I hope your day gets better.
It is a lot easier to ask for forgiveness than permission.
I run a few websites that generate a small but growing amount of traffic (and hopefully one day, revenue), so its also nice to know that from any computer, I can log in to my jump server, and take a look if something goes down, with nothing more than a web browser. It could be a locked down computer only offering a web browser, or a friend's computer I don't want to be fumbling around and installing putty on, etc.
But yeah 98% of my use is avoiding IT policies.
It runs over https. For most users like myself, its about access, not security.
Yeah there are. Most are pretty crappy. GateOne just performs far better, to the point where you can actually use it for useful work. I used to use ajaxterm, it worked, but had weird bugs. Later I moved to webshell, which was a big improvement. GateOne is an improvement further still. I use this on my personal server, so unless they are inspecting packets to see that this is an app, which currently my very paranoid company does not do, this won't get blocked.
No, its not. It runs in a browser and is thus cross platform. The main point is that the ssh session runs on the server side, so you are essentially tunneling ssh through http. Very nice if you need ssh access in an environment that won't let you ssh out of the network.