Trading is pushed faster and faster by competition. Nothing sinister or complex. The slower trading company either stops making or loses money. Natural market forces.
In over 10 years of HFT, is that the only evidence you have? Because this is more evidence that HFT is beneficial.
1. There are stops below AND above the price - some people have short positions that they are protecting. 2. Stops only trigger once - if this was someone pushing the price around to make money from stop orders, why did they keep pushing it across the same prices over and over? 3. If this was a single entity pushing the price around the SEC would have discovered that in their investigation.
This was actually triggered by $4billion worth of E-Minis being sold in a very short time. That is a huge amount and will push the price down. But the market makers that took the other side of those trades are now losing money on them as the price drops. Now the market makers are trading furiously to offload those positions without losing too much money. The increased trade volume cause NYSE quote data to get behind, but the timestamps indicated the data were fresh, so some were trading on delayed prices. Some traders couldn't tell that something was wrong and tried to arbitrate between NYSE and the other exchanges. This caused an increase in trading, which causes an increase in market data volume, which caused NYSE to get further behind, exacerbating the problem. After a short time, most HFT market making algos were pushed outside their risk parameters and pulled out, reducing liquidity. This pushed spreads wider, which also exacerbated the problem. It's this trading from delayed prices that causes the saw-tooth repeating pattern.
All trading in E-Minis affects the major components, hence a wide range of stocks were also affected by the same issue.
Without NYSE's delayed quotes the system would have remained stable.
Selling $4billion worth of E-Minis is going to cause trouble unless it's done over a long enough duration. Even then it will have some effect.
Without HFT you would see that cause heavy downward momentum, and people desperate to avoid losses as the prices tank start selling at lower and lower prices.
Before HFT the spreads were 25c. That's a cost to the investor. Because of HFT the spreads are 1c, saving the investor transaction costs. The amount HFT takes our of the trade is a tiny fraction of what your broker charges. The evidence I've seen points to malformed delayed quotes coming out of NYSE as the proximate cause of the 2010 flash crash. Many HFT algos stopped trading because of the illogical info and increased risk. A transaction tax would hurt the investor, not the market maker, because the spreads would increase, since it would be too costly for any participant to keep them less than the tax. That would reduce liquidity and increase volatility.
People react to the news anyway, and were very effective at crashing markets. HFT does two things in such a situation: 1 combines much more information into the price 2 provides liquidity Combining all the information simply prices more accurately and quickly. This resists momentum, not the opposite. Providing liquidity gives people more confidence. A contributing factor to crashes is the lack of liquidity. When someone wants to sell, but there is no buyer, they lower the price to get rid of it before it drops even more. That really does create momentum. HFT provides the buyer, which reduces momentum. Fact: the biggest crashes over the past couple of years have occurred when HFT pulled out due to bad information and impossible quotes. Fact: studies has concluded that HFT is overall good for the marketplace.
I don't think any exchanges are matching subpenny anymore. Even crossfinder. Yes: the dark pool! The price levels that are hidden due to crossing the nbbo just cause trouble. They exist because of stupid rules that attempt to protect the customer, by forcing a broker to shop around to find the best price. These days the markets do that naturally, because any discrepency is captured by Arbitrageurs. I agree that those prices should be displayed.
But that's exactly what I'm saying. The algos submit buys and sells to the exchange. Even if some hacker had complete control and replaced the algo, about the worst it could do would be sell short in a loop. In a very short time, the stock would drop in value. But that triggers a circuit breaker in the exchange halting trading in that symbol. Then an investigation would identify the bad trades and break them. It would be bad, but not even a fraction of the mortgage CDO fiasco.
The only reason you feel this way is because of bad press. It's clear from your post that you know nothing about exchange trading, which traces its roots to before 1800. HFT does, among other things, market making. Market making is almost as old as exchange trading. It used to be that that a specialist would stand ready to buy or sell and would maintain a liquid and orderly market. It's a risky position: they are required to keep orders on both sides of the market within 3 ticks or best bid and best ask. When news hits, the value of a company changes rapidly, and traders sweep the market makers orders costing them a great deal of money. So market making is good for the market, but costly. As a result the market makers used to keep a minimum of 25c between their bid and ask. The resulting spread meant more profit to the market maker at a cost to all other participants, including investors. Since HFT, computers can move the bids and asks much quicker in response to value changes, and therefore reduce risk. This reduced risk means they can make a profit on reduced spreads. Directly due to HFT, spreads are now 1c on liquid products. This significantly reduced the cost to investors. From 25c to 1c! When you buy a stock, who do you think is selling it to you? When your holding in AMD tanked, and you had stop limit orders in place to protect you from losses, who do you honk bought that falling stock? The HFT market maker.
There isn't that much havoc a single HFT shop can do. With an enormous amount of effort you might be able to push some prices around. But before those trades settle the fraud would be detected and the trades broken. There would be a cost, but it's not like NYSE wouldn't open the following day, or the dollar would be worthless.
You have no idea what you are talking about. HFT doesn't push the price around. Supply and demand, profit, news, similar company price changes - the unseen hand - push the price around. HFT just makes it quicker and more efficient. Even if the value of a company stays steady, the bids will be below the value and he asks will be above. The market maker makes 1c per pair of trades acrid he spread per share. When companies price fluctuates it's mostly due to fluctuations in other instruments. Commodities, futures, options, other stocks, bonds, currencies, etc... When the price of milk futures changes, it affects the price of Kraft Foods because their costs change, and therefore their profit. HFT does not change price discovery (very basic economics) it makes it happen quicker. HFT results directly from competition, and has made the markets much more efficient.
Hacking into a single HFT shop will maybe get you some source code, and maybe cost the shop serious money. Even put it out of business, perhaps. But the overall effect on the marketplace as a whole? Little more than negligible.
There really is a queue at the exchange. What's important is the time to get into the queue. That would still be the case even if the queue weren't processed until the end of a delay. BTW, they are orders. The trades are the result not the input.
Markets are correlated with or without HFT. Destroying HFT would not change than, only increase inefficiency. Stocks aren't priced in a bubble. Companies have an effect on each others prices. HFT makes the price discovery process quicker, and for less profit. You are also wrong about HFT gaming the system. HFT plays an important role in the market place. Without it prices would swing more violently, risk would be less manageable,the cost of trading would skyrocket. Portfolios would cost more to rebalance, so they would be less profitable. Growth would decline or reverse.
Also, you are wrong about a transaction tax. A tax on each transaction would result in wider spreads increasing the cost to the investor, not the market maker. This is because the spreads are as tight as they are only because there is a low cost to market makers.
Removing intraday trading would be insane. A replacement system would be even more complex - with more inefficiency and more opportunity to game the system. Today people can offset risk as it arises. If you can only trade once a day you have to give risk a higher cost resulting in higher cost to investing and slower growth.
Caps don't make it true. HFT is directly responsible for a reduction in spreads from 25c (or more) to 1c in liquid stocks. This takes money from the original market makers and gives it to the investors. If a tax of, say 25c per stock, were imposed, HFT wouldn't bear the cost. The investor would. This is because the spread takes into account the cost of trading. No company can market make at a loss, and since they make less than 1c the spreads would necessarily widen. QED
Based on what? Just increases in volume? Grab your pitchfork! I'd recommend waiting for a little more evidence. So far all research into the subject has concluded the HFT is good for markets. But don't let facts dissuade you.
The HFT shops I've worked at are all for transparency. Dark pools are considered inefficient. When you trade on an exchange all the quotes, prices, volumes and trades are public. What else would make things more fair?
This is what you get when the uninformed do journalism. This article has only HFT volumes to go by, and yet draws some dark conclusions about the future. It's all nonsense. I've been in HFT since before the beginning of the animation. I can say that the trades are mostly very simple. They are written with great care and attention to risk. HFT shops have no unfair advantage - like any trading company they put up risk and capital in order to attempt to make a profit. And when things go wrong they can lose bug and fast. The reason for growth in HFT is partly due to how easy it really is. More and more trading companies are trying to enter the game. HFT is like Target. They cut margins to next to nothing and make it up with volume. Knight is really a good example. They had the worst case scenario: heir robots were buying and selling on bad information and their risk systems didn't detect it. Bad for Knight. But did it crash the whole market? Did it have the same effect as the mortgage CDO fiasco? Not at all. HFT shops aren't leveraged. If they put up $1 they could lose $1. And often they do.
Pump and dump usually uses news and similar to push up the price. And it's investors that buy into the news. No one really speculates anymore anyway. It's just as likely to lose money as make it. In the long term that means losing everything due to fees. Traders always hedge these days. Usually even arbitrageurs. Buying a stock simply because it's seen recent gains is just as likely to lose you money as make it. Facebook tanked exactly because investors DID look at the fundamentals, and they saw a company that has already captialized on its market and doesn't seem likely to appreciate much future growth. It's evidence AGAINST your position. Also, nothing you've said has anything to do with HFT or Knight.
At 250 fps a frame is generated every 4ms. If you can generate a frame in 4ms, you can do other processing for 12+ms and still have the next frame ready to achieve 60fps. Not sure what you mean by linear in this context.
You just restated your opinion. Speculators buy and sell equally - their effect on value is zero. It takes a large disparity between aggressive buys and sells to move a price. How large depends on how liquid the stock is.
Tearing a company apart to make a quick buck is short term thinking. Speculating could be called that too. That doesn't mean they are equally bad. When someone buys a stock and holds it for a short time then sells it, it only returns to the company if they buy back - a deliberate action. Usually it's another market participant who buys it. Why is it bad for the stock to change ownership? What makes you think speculation is bad? Is it better not to buy the stock in the first place? The government doesn't print money in response to some company's stock tanking. It might do it to cover obligations (bonds) that it has to repay, if tax didn't bring in enough revenue. Which it hasn't been. And given how raising taxes is political suicide the deficit will only increase. But I digress.
Liquidity is provided by market makers - participants who keep both buy and sell orders in the market. Such activity is risky - when an event occurs another participant will attempt to cross those buys or sells and pick them off for a quick profit. The market maker has to react very quickly to avoid losses. They gain a very small amount on most trades, but these picks and lose vastly more. The competition leads to increased reliance on low latency. Also, market makers compete with each other, also forcing latency down. Note that this reduces he spreads too - before HFT they were at 25c at least. Liquidity providing and flash trading are very different things.
Slashdot Mirror
Batching doesn't change the number of transactions.
Trading is pushed faster and faster by competition. Nothing sinister or complex. The slower trading company either stops making or loses money.
Natural market forces.
In over 10 years of HFT, is that the only evidence you have? Because this is more evidence that HFT is beneficial.
1. There are stops below AND above the price - some people have short positions that they are protecting.
2. Stops only trigger once - if this was someone pushing the price around to make money from stop orders, why did they keep pushing it across the same prices over and over?
3. If this was a single entity pushing the price around the SEC would have discovered that in their investigation.
This was actually triggered by $4billion worth of E-Minis being sold in a very short time. That is a huge amount and will push the price down. But the market makers that took the other side of those trades are now losing money on them as the price drops. Now the market makers are trading furiously to offload those positions without losing too much money. The increased trade volume cause NYSE quote data to get behind, but the timestamps indicated the data were fresh, so some were trading on delayed prices. Some traders couldn't tell that something was wrong and tried to arbitrate between NYSE and the other exchanges. This caused an increase in trading, which causes an increase in market data volume, which caused NYSE to get further behind, exacerbating the problem. After a short time, most HFT market making algos were pushed outside their risk parameters and pulled out, reducing liquidity. This pushed spreads wider, which also exacerbated the problem. It's this trading from delayed prices that causes the saw-tooth repeating pattern.
All trading in E-Minis affects the major components, hence a wide range of stocks were also affected by the same issue.
Without NYSE's delayed quotes the system would have remained stable.
Selling $4billion worth of E-Minis is going to cause trouble unless it's done over a long enough duration. Even then it will have some effect.
Without HFT you would see that cause heavy downward momentum, and people desperate to avoid losses as the prices tank start selling at lower and lower prices.
Before HFT the spreads were 25c. That's a cost to the investor.
Because of HFT the spreads are 1c, saving the investor transaction costs. The amount HFT takes our of the trade is a tiny fraction of what your broker charges.
The evidence I've seen points to malformed delayed quotes coming out of NYSE as the proximate cause of the 2010 flash crash. Many HFT algos stopped trading because of the illogical info and increased risk.
A transaction tax would hurt the investor, not the market maker, because the spreads would increase, since it would be too costly for any participant to keep them less than the tax.
That would reduce liquidity and increase volatility.
People react to the news anyway, and were very effective at crashing markets.
HFT does two things in such a situation:
1 combines much more information into the price
2 provides liquidity
Combining all the information simply prices more accurately and quickly. This resists momentum, not the opposite.
Providing liquidity gives people more confidence. A contributing factor to crashes is the lack of liquidity. When someone wants to sell, but there is no buyer, they lower the price to get rid of it before it drops even more. That really does create momentum.
HFT provides the buyer, which reduces momentum.
Fact: the biggest crashes over the past couple of years have occurred when HFT pulled out due to bad information and impossible quotes.
Fact: studies has concluded that HFT is overall good for the marketplace.
I don't think any exchanges are matching subpenny anymore.
Even crossfinder. Yes: the dark pool!
The price levels that are hidden due to crossing the nbbo just cause trouble. They exist because of stupid rules that attempt to protect the customer, by forcing a broker to shop around to find the best price. These days the markets do that naturally, because any discrepency is captured by Arbitrageurs.
I agree that those prices should be displayed.
But that's exactly what I'm saying.
The algos submit buys and sells to the exchange.
Even if some hacker had complete control and replaced the algo, about the worst it could do would be sell short in a loop.
In a very short time, the stock would drop in value. But that triggers a circuit breaker in the exchange halting trading in that symbol.
Then an investigation would identify the bad trades and break them.
It would be bad, but not even a fraction of the mortgage CDO fiasco.
The only reason you feel this way is because of bad press. It's clear from your post that you know nothing about exchange trading, which traces its roots to before 1800.
HFT does, among other things, market making. Market making is almost as old as exchange trading. It used to be that that a specialist would stand ready to buy or sell and would maintain a liquid and orderly market. It's a risky position: they are required to keep orders on both sides of the market within 3 ticks or best bid and best ask. When news hits, the value of a company changes rapidly, and traders sweep the market makers orders costing them a great deal of money.
So market making is good for the market, but costly. As a result the market makers used to keep a minimum of 25c between their bid and ask. The resulting spread meant more profit to the market maker at a cost to all other participants, including investors.
Since HFT, computers can move the bids and asks much quicker in response to value changes, and therefore reduce risk. This reduced risk means they can make a profit on reduced spreads.
Directly due to HFT, spreads are now 1c on liquid products. This significantly reduced the cost to investors. From 25c to 1c!
When you buy a stock, who do you think is selling it to you?
When your holding in AMD tanked, and you had stop limit orders in place to protect you from losses, who do you honk bought that falling stock?
The HFT market maker.
There isn't that much havoc a single HFT shop can do.
With an enormous amount of effort you might be able to push some prices around. But before those trades settle the fraud would be detected and the trades broken.
There would be a cost, but it's not like NYSE wouldn't open the following day, or the dollar would be worthless.
You have no idea what you are talking about.
HFT doesn't push the price around. Supply and demand, profit, news, similar company price changes - the unseen hand - push the price around.
HFT just makes it quicker and more efficient.
Even if the value of a company stays steady, the bids will be below the value and he asks will be above. The market maker makes 1c per pair of trades acrid he spread per share.
When companies price fluctuates it's mostly due to fluctuations in other instruments. Commodities, futures, options, other stocks, bonds, currencies, etc...
When the price of milk futures changes, it affects the price of Kraft Foods because their costs change, and therefore their profit.
HFT does not change price discovery (very basic economics) it makes it happen quicker.
HFT results directly from competition, and has made the markets much more efficient.
Hacking into a single HFT shop will maybe get you some source code, and maybe cost the shop serious money. Even put it out of business, perhaps.
But the overall effect on the marketplace as a whole? Little more than negligible.
An investor only needs the best ask and best bid. They don't need the full microstructure.
Quoting without intention to trade the quote is illegal.
This applies equally to HFT.
There really is a queue at the exchange. What's important is the time to get into the queue. That would still be the case even if the queue weren't processed until the end of a delay.
BTW, they are orders. The trades are the result not the input.
Markets are correlated with or without HFT. Destroying HFT would not change than, only increase inefficiency.
Stocks aren't priced in a bubble. Companies have an effect on each others prices. HFT makes the price discovery process quicker, and for less profit. You are also wrong about HFT gaming the system. HFT plays an important role in the market place. Without it prices would swing more violently, risk would be less manageable,the cost of trading would skyrocket. Portfolios would cost more to rebalance, so they would be less profitable. Growth would decline or reverse.
Also, you are wrong about a transaction tax. A tax on each transaction would result in wider spreads increasing the cost to the investor, not the market maker. This is because the spreads are as tight as they are only because there is a low cost to market makers.
Removing intraday trading would be insane. A replacement system would be even more complex - with more inefficiency and more opportunity to game the system. Today people can offset risk as it arises. If you can only trade once a day you have to give risk a higher cost resulting in higher cost to investing and slower growth.
Caps don't make it true.
HFT is directly responsible for a reduction in spreads from 25c (or more) to 1c in liquid stocks. This takes money from the original market makers and gives it to the investors.
If a tax of, say 25c per stock, were imposed, HFT wouldn't bear the cost. The investor would. This is because the spread takes into account the cost of trading. No company can market make at a loss, and since they make less than 1c the spreads would necessarily widen. QED
Based on what? Just increases in volume?
Grab your pitchfork!
I'd recommend waiting for a little more evidence.
So far all research into the subject has concluded the HFT is good for markets. But don't let facts dissuade you.
The HFT shops I've worked at are all for transparency.
Dark pools are considered inefficient.
When you trade on an exchange all the quotes, prices, volumes and trades are public. What else would make things more fair?
This is what you get when the uninformed do journalism.
This article has only HFT volumes to go by, and yet draws some dark conclusions about the future.
It's all nonsense. I've been in HFT since before the beginning of the animation. I can say that the trades are mostly very simple. They are written with great care and attention to risk. HFT shops have no unfair advantage - like any trading company they put up risk and capital in order to attempt to make a profit. And when things go wrong they can lose bug and fast. The reason for growth in HFT is partly due to how easy it really is.
More and more trading companies are trying to enter the game.
HFT is like Target. They cut margins to next to nothing and make it up with volume.
Knight is really a good example. They had the worst case scenario: heir robots were buying and selling on bad information and their risk systems didn't detect it. Bad for Knight. But did it crash the whole market? Did it have the same effect as the mortgage CDO fiasco? Not at all. HFT shops aren't leveraged. If they put up $1 they could lose $1. And often they do.
Pump and dump usually uses news and similar to push up the price. And it's investors that buy into the news.
No one really speculates anymore anyway. It's just as likely to lose money as make it. In the long term that means losing everything due to fees. Traders always hedge these days. Usually even arbitrageurs.
Buying a stock simply because it's seen recent gains is just as likely to lose you money as make it.
Facebook tanked exactly because investors DID look at the fundamentals, and they saw a company that has already captialized on its market and doesn't seem likely to appreciate much future growth. It's evidence AGAINST your position.
Also, nothing you've said has anything to do with HFT or Knight.
At 250 fps a frame is generated every 4ms.
If you can generate a frame in 4ms, you can do other processing for 12+ms and still have the next frame ready to achieve 60fps.
Not sure what you mean by linear in this context.
You just restated your opinion.
Speculators buy and sell equally - their effect on value is zero. It takes a large disparity between aggressive buys and sells to move a price. How large depends on how liquid the stock is.
Tearing a company apart to make a quick buck is short term thinking. Speculating could be called that too.
That doesn't mean they are equally bad.
When someone buys a stock and holds it for a short time then sells it, it only returns to the company if they buy back - a deliberate action. Usually it's another market participant who buys it. Why is it bad for the stock to change ownership?
What makes you think speculation is bad? Is it better not to buy the stock in the first place?
The government doesn't print money in response to some company's stock tanking. It might do it to cover obligations (bonds) that it has to repay, if tax didn't bring in enough revenue. Which it hasn't been. And given how raising taxes is political suicide the deficit will only increase. But I digress.
None of those points have anything to do with HFT.
Liquidity is provided by market makers - participants who keep both buy and sell orders in the market. Such activity is risky - when an event occurs another participant will attempt to cross those buys or sells and pick them off for a quick profit. The market maker has to react very quickly to avoid losses. They gain a very small amount on most trades, but these picks and lose vastly more. The competition leads to increased reliance on low latency.
Also, market makers compete with each other, also forcing latency down.
Note that this reduces he spreads too - before HFT they were at 25c at least.
Liquidity providing and flash trading are very different things.