This is an interesting video. Notice the mention of using futures (managed futures) to target risk allocation rather than simply targeting asset allocation.
3/12/12
Do you own municipal bonds?
If so, or if someone is trying to talk you into buying them for the interest, you might want to read this: "City of Stockton will suspend bond payments to avoid bankruptcy".
Some quotes form the article:
It just seems that there are better ways to target risk for the potential of reward than buying municipal bonds. Remember Jefferson County, AL? Governments of all levels have shown time and time again that they are unable to properly assess risk.
Some quotes form the article:
"Hopeing to stave off the ignominy of becoming America's largest city to declare bankruptcy......"
"Deis said Stockton's $20 million budget deficit could double or worse if the city doesn't take dramatic actions."
"....the city effectively created an unsustainable 'ponzi sheme' when it approved benefits, including health care for life for city employees......"
"He said the city ran up as much as $320 million in municipal debt while wrongly counting on increased housing growth and resulting tax revenues....."
It just seems that there are better ways to target risk for the potential of reward than buying municipal bonds. Remember Jefferson County, AL? Governments of all levels have shown time and time again that they are unable to properly assess risk.
3/7/12
3/6/12
Kicking Out Losers
All trades eventually have to be exited. Which should be the first to go? Some schools of thought say take profits from winning trades in order to hold onto the losers until they pay off. If they pay off. Another school of thought is to remove the losers and let the winners run.
The second school of thought makes the most sense to me. I've learned to kick out the losers and to let the winners run. If you've seen Money Ball, you might remember the scene where Billy Bean is trading a handful of players. In response to Bean's sudden trading activity, the statistics master behind their system freaks out a little. Bean illustrates that numbers produced by these players make almost zero difference. At that point the stats master approves and sees the light.
That's when the team's trajectory turned around. Bean improved performance by getting rid of underperforming players. One of the immediate effects was to let the rest of the team know that this was serious business.
Of course the whole system of Moneyball was about buying runs, not individual players. Buying wins, not hype. It worked because most of baseball was stuck in the paradigm of seeing individual players as the most important factor for success. Money ball turned that paradigm on its head. It should be no surprise that a successful commodity trader such as John Henry recognized and adopted the immense value in Bean's system. According to stories told by Michael Covel, Henry's commodity trading strategy was basically to invest in systems that cut losses short and let winners run rather than to put a lot of hope in individual markets.
A loss sucks, but it's part of trading. We plan for it, limit it, and manage it. If you want to survive and thrive, you have to kick out the losing trades. Most investors kick out the winners. They hold losers, hoping they will come back. They exit winners happy to take the current profit. We are trying a different strategy. We systematically kick out losers. Doing so increases long-term absolute returns and improves the unrealized profit figure right away. It's nice to have a team of winners.
The second school of thought makes the most sense to me. I've learned to kick out the losers and to let the winners run. If you've seen Money Ball, you might remember the scene where Billy Bean is trading a handful of players. In response to Bean's sudden trading activity, the statistics master behind their system freaks out a little. Bean illustrates that numbers produced by these players make almost zero difference. At that point the stats master approves and sees the light.
That's when the team's trajectory turned around. Bean improved performance by getting rid of underperforming players. One of the immediate effects was to let the rest of the team know that this was serious business.
Of course the whole system of Moneyball was about buying runs, not individual players. Buying wins, not hype. It worked because most of baseball was stuck in the paradigm of seeing individual players as the most important factor for success. Money ball turned that paradigm on its head. It should be no surprise that a successful commodity trader such as John Henry recognized and adopted the immense value in Bean's system. According to stories told by Michael Covel, Henry's commodity trading strategy was basically to invest in systems that cut losses short and let winners run rather than to put a lot of hope in individual markets.
A loss sucks, but it's part of trading. We plan for it, limit it, and manage it. If you want to survive and thrive, you have to kick out the losing trades. Most investors kick out the winners. They hold losers, hoping they will come back. They exit winners happy to take the current profit. We are trying a different strategy. We systematically kick out losers. Doing so increases long-term absolute returns and improves the unrealized profit figure right away. It's nice to have a team of winners.
2/29/12
Speculators and Oil Prices
I keep hearing people blame speculators for high oil prices. None of us like paying more at the pump, but the argument that speculators cause the higher prices is incorrect. To illustrate a couple of points, I am going to use a very simple automated trend-following trading code to show trade signals in the Crude Oil market that are probably pretty similar to what a lot of speculators experienced.
Speculators don't always make money.
It might be hard to see, but if you look closely you will see red dotted lines and blue dotted lines in the chart above. They connect trade signals. "Buy" means the trader entered long (bullish). "Sell" means the trader is exiting a long position. "Short" means the trader entered the market by selling it short (bearish). "Cover" means the trader covered the short position.
The red lines (look closely) indicate loosing trades. Yes, contrary to what the talking heads are saying, speculators lose money in the oil market, and every other market for that matter. Draw-downs and extended draw-down periods are a part of the business. The blue lines indicate winning trades.
Notice how many losing trades there are! This particular strategy only wins 36% of the time in the Oil market. That's actually a pretty normal winning percentage for a trading strategy. Most trades are losers for every trader I know of. Speculators make money over a long period time by managing risk, not by manipulating a market.
Speculators participate in shorting the oil market just as much as they participate in buying the oil market.
The same automated system is used on the chart above. Look at how those evil speculators suppressed the price of oil by shorting it from $200 all the way to $68! Shame on them! And they weren't simply happy with profits from the first short sale. They did it three more times after that! One of the four shorts was a loser.
How come we don't hear about speculators bringing prices down on the news? Maybe its not entertaining enough. Perhaps they are clueless to reality. It could be that they want to mislead the public. I don't know. I'm just sick of hearing it.
Are speculators long Oil right now? Sure. Some of them are. Here's a picture of the same system from today:
The chart above is current as of the morning of 2/28/12. It shows that speculators who are following the price trend are likely to be long this market at the moment. On the far right, you will see "Stop". That's the exit level for the current long position.
Now look to the left of the chart. Those trades are short sales! Some of them worked and some of them did not. Same as with the longs on the right. Clearly, the argument that speculators are causing the oil price to rise is simply garbage. Besides, there is more than one type of speculator.
Different types of speculators.
Trend-followers are likely to behave in a similar way as shown in the charts above. They simply follow price trends as they breakout. Most of their trades are losers, and they make money by managing risk on the losers until they finally get a winner, which will often cover the losses of previous losers and then some. Many of these types use automated systems. Trend-followers do not forecast price action. They are tactical in their approach and simply follow.
Discretionary technical traders look for chart patterns and other indicators that are usually some type of derivative of price action, volume, and sentiment to guide them. A lot of times, they will end up in the same major trends as the systematic trend-followers. They differ from trend-followers in several ways, one of which is that they will often attempt to forecast market action. Some of these differences lead them to occasionally be on the opposite side of the market from the trend-followers.
Fundamentalist traders view price action from a Newtonian mindset, believing specific outside forces will push it one way or the other. Although they will often have a different perspective than the technical trader on what might move the market, the fundamentalist is similar to the discretionary technical trader in that he uses forecasts of market action as the main road map. Forecasts about the effects of outside economic forces guide this trader into speculative positions that can be long or short, and do not have to be in harmony with the prevailing trend.
Contrarian traders look for markets they believe are overextended and therefore providing a chance to take the opposite side of the trend that has been prevailing. To be fair, many times the discretionary or fundamentalist traders will fit into this group. There are probably not that many pure contrarian traders, or ones that use such a strategy exclusively, but maybe I just don't know of them. Trend-followers never fit into this group.
Hedgers represent the producers or major consumers of commodities. Farmers, airlines, processed food producers, etc.... they take positions not for the purpose of making money on the position, but for the purpose of hedging the price of the product they produce or use in great quantities. They are looking to lock in prices at certain levels. Hedgers are not really speculators, but I included them here because they are part of the market of buyers and sellers. Hedgers need speculators in the market so they will have someone to take the other side of their hedge and so that there is ample liquidity in the market.
As you can see, all of these groups are typically competing with each other and at the same time working together voluntarily in a process of price discovery.
Price discovery is important in free-markets
The Soviet system had no price discovery. The North Korean system does not have proper price discovery. People starved under the Soviet system. People Starve under the North Korean system. Both the Soviet system and the North Korean one were designed as supposed solutions to the presumed problems of a market economy, or one with price discovery.
Prices communicate the message of supply and demand to producers, entrepreneurs, and consumers. Generally speaking, rising prices signal producers and risk taking entrepreneurs that there is possible reward in bringing more product to the market to serve the demand. Suppliers compete to serve those on the demand side, which then brings prices down as supply and demand converge. If prices are lower than the cost of production, then suppliers and entrepreneurs leave or reduce exposure to that market in order to divert resources to operations that can meet other areas of demand. Without price, no one really knows what to produce and what quantities to produce it in.
As was already stated, anti-free-market systems such as that of the former Soviets and that of the current North Koreans have tried planning this from a top down approach, with people starving as a result. Prices tell producers what to do and how much to do it. It is spontaneous anti-authoritarian voluntary order between producers and consumers. In fact, the line of who is a producer and who is a consumer is often blurred. It's really just voluntary spontaneous order. And that is a good thing. Prices convey a vital message for voluntary interactions.
The real culprit is the Fed
I don't want to go on and on making this blog post supper long, but I do want to mention the Federal Reserve Bank. They have been destroying the value of the US Dollar since 1913. Prices, as we have seen, fluctuate up and down due to normal supply and demand and the price discovery process, but sometimes it is the value of the dollar going down rather than prices going up.
Just look at Gold and Silver over recent years as an example. A high quality suit has cost the same as one ounce of gold for generations. But look how much a suit costs in dollars over that same period of time. It has increased exponentially! The Fed and it's expansionary monetary policy distorts markets.
A quick Google search should pull up some charts of phenomenon for you. Below are a couple I got just from doing a simple search. I didn't even have to go to sites I normally visit, so this is not an endorsement of any particular site.
Source: http://mungowitzend.blogspot.com/2011/12/every-picture-tells-story_06.html
Source: http://seekingalpha.com/article/136589-dollar-s-purchasing-power-annihilated-the-chart-they-don-t-want-you-to-see
These last two charts were just the results of a quick Google search that you can do yourself. To be clear, the trading system charts are mine, and the system is a strategy I wrote the code for.
You can get some ultra long-term charts of oil from: http://www.thechartstore.com/. I found some posted around the internet, but noticed a bunch of copyright notices at the bottom of them. With the way IP battles are going these days, I'll just leave it up to you to look it up.
As you can see, speculators, which is just a derogatory way of saying entrepreneurs, are not the cause of oil prices going up. I don't know if the media intentionally wants to take your eye off the ball or not. They are possibly just clueless pretty faces meant to entertain. Who knows? All I know is that I'm sick of hearing about evil speculators, but it did give me something to get fired up about while riding the bike at the gym. I turned the TV at my house off a long time ago.
Speculators don't always make money.
It might be hard to see, but if you look closely you will see red dotted lines and blue dotted lines in the chart above. They connect trade signals. "Buy" means the trader entered long (bullish). "Sell" means the trader is exiting a long position. "Short" means the trader entered the market by selling it short (bearish). "Cover" means the trader covered the short position.
The red lines (look closely) indicate loosing trades. Yes, contrary to what the talking heads are saying, speculators lose money in the oil market, and every other market for that matter. Draw-downs and extended draw-down periods are a part of the business. The blue lines indicate winning trades.
Notice how many losing trades there are! This particular strategy only wins 36% of the time in the Oil market. That's actually a pretty normal winning percentage for a trading strategy. Most trades are losers for every trader I know of. Speculators make money over a long period time by managing risk, not by manipulating a market.
Speculators participate in shorting the oil market just as much as they participate in buying the oil market.
The same automated system is used on the chart above. Look at how those evil speculators suppressed the price of oil by shorting it from $200 all the way to $68! Shame on them! And they weren't simply happy with profits from the first short sale. They did it three more times after that! One of the four shorts was a loser.
How come we don't hear about speculators bringing prices down on the news? Maybe its not entertaining enough. Perhaps they are clueless to reality. It could be that they want to mislead the public. I don't know. I'm just sick of hearing it.
Are speculators long Oil right now? Sure. Some of them are. Here's a picture of the same system from today:
The chart above is current as of the morning of 2/28/12. It shows that speculators who are following the price trend are likely to be long this market at the moment. On the far right, you will see "Stop". That's the exit level for the current long position.
Now look to the left of the chart. Those trades are short sales! Some of them worked and some of them did not. Same as with the longs on the right. Clearly, the argument that speculators are causing the oil price to rise is simply garbage. Besides, there is more than one type of speculator.
Different types of speculators.
Trend-followers are likely to behave in a similar way as shown in the charts above. They simply follow price trends as they breakout. Most of their trades are losers, and they make money by managing risk on the losers until they finally get a winner, which will often cover the losses of previous losers and then some. Many of these types use automated systems. Trend-followers do not forecast price action. They are tactical in their approach and simply follow.
Discretionary technical traders look for chart patterns and other indicators that are usually some type of derivative of price action, volume, and sentiment to guide them. A lot of times, they will end up in the same major trends as the systematic trend-followers. They differ from trend-followers in several ways, one of which is that they will often attempt to forecast market action. Some of these differences lead them to occasionally be on the opposite side of the market from the trend-followers.
Fundamentalist traders view price action from a Newtonian mindset, believing specific outside forces will push it one way or the other. Although they will often have a different perspective than the technical trader on what might move the market, the fundamentalist is similar to the discretionary technical trader in that he uses forecasts of market action as the main road map. Forecasts about the effects of outside economic forces guide this trader into speculative positions that can be long or short, and do not have to be in harmony with the prevailing trend.
Contrarian traders look for markets they believe are overextended and therefore providing a chance to take the opposite side of the trend that has been prevailing. To be fair, many times the discretionary or fundamentalist traders will fit into this group. There are probably not that many pure contrarian traders, or ones that use such a strategy exclusively, but maybe I just don't know of them. Trend-followers never fit into this group.
Hedgers represent the producers or major consumers of commodities. Farmers, airlines, processed food producers, etc.... they take positions not for the purpose of making money on the position, but for the purpose of hedging the price of the product they produce or use in great quantities. They are looking to lock in prices at certain levels. Hedgers are not really speculators, but I included them here because they are part of the market of buyers and sellers. Hedgers need speculators in the market so they will have someone to take the other side of their hedge and so that there is ample liquidity in the market.
As you can see, all of these groups are typically competing with each other and at the same time working together voluntarily in a process of price discovery.
Price discovery is important in free-markets
The Soviet system had no price discovery. The North Korean system does not have proper price discovery. People starved under the Soviet system. People Starve under the North Korean system. Both the Soviet system and the North Korean one were designed as supposed solutions to the presumed problems of a market economy, or one with price discovery.
Prices communicate the message of supply and demand to producers, entrepreneurs, and consumers. Generally speaking, rising prices signal producers and risk taking entrepreneurs that there is possible reward in bringing more product to the market to serve the demand. Suppliers compete to serve those on the demand side, which then brings prices down as supply and demand converge. If prices are lower than the cost of production, then suppliers and entrepreneurs leave or reduce exposure to that market in order to divert resources to operations that can meet other areas of demand. Without price, no one really knows what to produce and what quantities to produce it in.
As was already stated, anti-free-market systems such as that of the former Soviets and that of the current North Koreans have tried planning this from a top down approach, with people starving as a result. Prices tell producers what to do and how much to do it. It is spontaneous anti-authoritarian voluntary order between producers and consumers. In fact, the line of who is a producer and who is a consumer is often blurred. It's really just voluntary spontaneous order. And that is a good thing. Prices convey a vital message for voluntary interactions.
The real culprit is the Fed
I don't want to go on and on making this blog post supper long, but I do want to mention the Federal Reserve Bank. They have been destroying the value of the US Dollar since 1913. Prices, as we have seen, fluctuate up and down due to normal supply and demand and the price discovery process, but sometimes it is the value of the dollar going down rather than prices going up.
Just look at Gold and Silver over recent years as an example. A high quality suit has cost the same as one ounce of gold for generations. But look how much a suit costs in dollars over that same period of time. It has increased exponentially! The Fed and it's expansionary monetary policy distorts markets.
A quick Google search should pull up some charts of phenomenon for you. Below are a couple I got just from doing a simple search. I didn't even have to go to sites I normally visit, so this is not an endorsement of any particular site.
Source: http://mungowitzend.blogspot.com/2011/12/every-picture-tells-story_06.html
Source: http://seekingalpha.com/article/136589-dollar-s-purchasing-power-annihilated-the-chart-they-don-t-want-you-to-see
These last two charts were just the results of a quick Google search that you can do yourself. To be clear, the trading system charts are mine, and the system is a strategy I wrote the code for.
You can get some ultra long-term charts of oil from: http://www.thechartstore.com/. I found some posted around the internet, but noticed a bunch of copyright notices at the bottom of them. With the way IP battles are going these days, I'll just leave it up to you to look it up.
As you can see, speculators, which is just a derogatory way of saying entrepreneurs, are not the cause of oil prices going up. I don't know if the media intentionally wants to take your eye off the ball or not. They are possibly just clueless pretty faces meant to entertain. Who knows? All I know is that I'm sick of hearing about evil speculators, but it did give me something to get fired up about while riding the bike at the gym. I turned the TV at my house off a long time ago.
2/28/12
2/27/12
Over Filtered Trading Strategy
It occurs to me that we might have a tendency to over filter trading systems. At least many of my less successful system codes reflect that. After all, what is the first thing we want to do with a robust system? Make it better of course! The problem is that these efforts can make a system less robust, less profitable, and less reactive to price trend. Recently I took my robust system code and tried to filter trend signals further. The following summarizes my findings.
First, RSI was tried as a trend signal filter. Performance was reduced, not enhanced. Second, after removing the RSI filter code, ADX filter code was inserted. Performance was also reduced with the ADX filter. Increasing filter thresholds only made matters worse. Threshold reduction on either of these calculations increased performance, but never back to up to the performance of the already robust system without either RSI or ADX filtering.
The devil is in the details. Some markets did end up showing better results and better equity curves. Most markets did not improve though. If you were only testing on one market, then you could be under the impression that the system had been improved. That's why you have to test a portfolio of markets.
Could these filters improve other systems? Sure! Most trend following systems are similar enough, but I have definitely seen performance differences in the various codes I have written for catching trends. Like most things in life, there is a lot of art in writing and picking the code / system you want to use. RSI and ADX are definitely worth testing as additional trend filters on other systems.
These results are encouraging. We often learn the most from failures. This time we learned that our simple and robust system does not improve by making it more complex in these specific ways. Simplicity is a good thing; especially when we can objectively discover that it works!
First, RSI was tried as a trend signal filter. Performance was reduced, not enhanced. Second, after removing the RSI filter code, ADX filter code was inserted. Performance was also reduced with the ADX filter. Increasing filter thresholds only made matters worse. Threshold reduction on either of these calculations increased performance, but never back to up to the performance of the already robust system without either RSI or ADX filtering.
The devil is in the details. Some markets did end up showing better results and better equity curves. Most markets did not improve though. If you were only testing on one market, then you could be under the impression that the system had been improved. That's why you have to test a portfolio of markets.
Could these filters improve other systems? Sure! Most trend following systems are similar enough, but I have definitely seen performance differences in the various codes I have written for catching trends. Like most things in life, there is a lot of art in writing and picking the code / system you want to use. RSI and ADX are definitely worth testing as additional trend filters on other systems.
These results are encouraging. We often learn the most from failures. This time we learned that our simple and robust system does not improve by making it more complex in these specific ways. Simplicity is a good thing; especially when we can objectively discover that it works!
2/23/12
Do Tight Initial Stops Improve Trading System Profitability?
That's one question I had during the final review of system code I wrote and plan to soon deploy in trading a diversified list of commodities. All of my trading has been fairly discretionary and on a smaller universe up to this point.
There are three exit, or "stop", strategies in the original system. All are trend following in nature. We don't shoot for targets. We ride waves.
One of the three stops is what I simply call the initial stop. Real creative naming skills! The initial stop is usually very tight and related to the price bar where the entry signal was given. The main stop/exit algorithm is volatility based and steps along with the market trend as the trend moves in direction of the trade signal. There is another stop heuristic that is somewhat related to same measures that calculate the entry signals. My code synthesizes all of these exits and uses the tightest one. The initial stop never moves, and the other two are not allowed to retreat even when volatility expands.
The initial stop tends to be much tighter than the other two right at the beginning of the trade. Perhaps I am giving up too much, but I should also mention that the position sizing algorithm is calculated off of the volatility based stop only, which tends to be the widest stop at the initiation point of the trend, or until (if) the trend moves in the direction of the trade.
Why the initial stop? Contrary to common perception, commodity traders tend to be pretty risk averse. At least the responsible ones. We want the most bang for the buck and tend to pinch pennies in this regard. Every trade is a mini business plan. How much will it cost me to see if a particular trend works? If we throw all available capital at every business idea, then we will go out of business as soon as we get a bad idea or trade. Bad trades are pretty frequent. The initial stop was designed to help us discover a bad business idea much earlier than the other two exit heuristics.
Should we be so quick to judge the trade idea? Today's due diligence research included the very simple task of taking this initial stop out of the system and running it against all the markets we have been testing. Although the current system tests profitably across a good number of markets, we want to know if we are initially giving the potential trends room enough to breath? Can profitability be improved by removing the initial stop in lieu of simply using the two trailing stops? That's the question I wanted to answer.
As a result of removing the initial stop, profitability increased in some markets and decreased in other markets. Overall, profitability increased significantly for the portfolio tested. Fewer markets failed to perform with this change, increasing the probability of success across all markets in the portfolio traded. In the original test, we found three markets in our group that were in negative territory at the end of the period tested. By removing the initial tight stop, the number of markets that did not end in positive territory over the extended test period was reduced to only one, which was only barely in negative territory.
The goal is to seek exponential gain and to compound capital by using limited and linear risk in robust and repeatable processes. Removing the initial tight stop from my main intermediate to long term system has increased profitability, simplified process and code, and increased robustness of the system.
There are three exit, or "stop", strategies in the original system. All are trend following in nature. We don't shoot for targets. We ride waves.
One of the three stops is what I simply call the initial stop. Real creative naming skills! The initial stop is usually very tight and related to the price bar where the entry signal was given. The main stop/exit algorithm is volatility based and steps along with the market trend as the trend moves in direction of the trade signal. There is another stop heuristic that is somewhat related to same measures that calculate the entry signals. My code synthesizes all of these exits and uses the tightest one. The initial stop never moves, and the other two are not allowed to retreat even when volatility expands.
The initial stop tends to be much tighter than the other two right at the beginning of the trade. Perhaps I am giving up too much, but I should also mention that the position sizing algorithm is calculated off of the volatility based stop only, which tends to be the widest stop at the initiation point of the trend, or until (if) the trend moves in the direction of the trade.
Why the initial stop? Contrary to common perception, commodity traders tend to be pretty risk averse. At least the responsible ones. We want the most bang for the buck and tend to pinch pennies in this regard. Every trade is a mini business plan. How much will it cost me to see if a particular trend works? If we throw all available capital at every business idea, then we will go out of business as soon as we get a bad idea or trade. Bad trades are pretty frequent. The initial stop was designed to help us discover a bad business idea much earlier than the other two exit heuristics.
Should we be so quick to judge the trade idea? Today's due diligence research included the very simple task of taking this initial stop out of the system and running it against all the markets we have been testing. Although the current system tests profitably across a good number of markets, we want to know if we are initially giving the potential trends room enough to breath? Can profitability be improved by removing the initial stop in lieu of simply using the two trailing stops? That's the question I wanted to answer.
As a result of removing the initial stop, profitability increased in some markets and decreased in other markets. Overall, profitability increased significantly for the portfolio tested. Fewer markets failed to perform with this change, increasing the probability of success across all markets in the portfolio traded. In the original test, we found three markets in our group that were in negative territory at the end of the period tested. By removing the initial tight stop, the number of markets that did not end in positive territory over the extended test period was reduced to only one, which was only barely in negative territory.
The goal is to seek exponential gain and to compound capital by using limited and linear risk in robust and repeatable processes. Removing the initial tight stop from my main intermediate to long term system has increased profitability, simplified process and code, and increased robustness of the system.
Subscribe to:
Posts (Atom)
Disclaimer:
