As part of a trading plan, a trader should come up with a trading methodology that can be incorporated into a few simple rules for buying and selling. This is the start of trading with a system. Having some sort of system is crucial to a trader; without one, no matter how simple, trading can be random, haphazard, and dangerous. A system doesn't have to be elaborate, etched in stone, or purely computerized; it can be discretionary as well as purely mechanical, but it must give you a guideline to trade with. Systems can be made or bought, but either way they must fit one's trading style. There is no single system that will fulfill every trader's needs. What works for one trader may not work for another, as each trader has a different approach he feels comfortable with. Overall, the point of having a system is to help a trader repeatedly make trades that have been backtested to have a high probability of success.
Simply put, a system is a set of rules a trader uses to make buy and sell decisions. It can be as plain as buy when one moving average crosses the other and sell when they cross the other way, or it can be more complicated and have 10 conditions that must be met before a trade is made. A good system not only will include entry signals but also will account for exits and stops. Just knowing when to get in is only half a system, not a full one. Though some people think a system has to be programmed into trading software such as TradeStation, which then spits out when to buy and sell, it doesn't have to be. A system can be any set of rules, patterns, or conditions a trader uses repeatedly when he sees the market setting up one way or another. I've used systems in which the signals were completely generated by looking at printed daily charts. I would write down where my entry and exit points would be and then wait for the market to approach those levels. Systems can be seen visually on a chart, as in the case of buying a stock that stops going down as the rest of the market keeps dropping. This is hard to program, but it is an entry signal in a system, as it is a clear condition that one can always follow. Systems can also be based on fundamental analysis, such as buying crude when stockpiles are lower than they were the week before and selling when they are higher. Many times I don't use a formal system to trade with: I know what patterns to look for, and if I see them, I'll make a trade. Since I look at different time frames all the time, it is hard to program one system that will read all the time frames. Instead, I use a series of systems and have to look visually for confirmation. I use some discretion, especially when it comes to getting out of a trade. If I see something I like or don't like in one of the different time frames, I'll get out. This is still considered trading with a system because I always use the same guidelines for making my trading decisions.
Traders who use purely mechanical signals, whether they are generated by a computer or not, are known as systematic traders; they get a signal and take it without any thought process, never deviating from the system. Other traders who use system-generated signals may do so on more selective processes, taking some trades and not others, depending on market conditions or as they wait for confirmation from another indicator so that they can time their trades better. These traders are known as discretionary traders. Both discretionary and systematic styles have their pros and cons, as I'll discuss later in this chapter.
I would venture to say that most professional traders use systems for a good part of their decision-making process. Whether it is computer-generated or just a set of rules and conditions they follow, having a system keeps them on track. Some use a fully systematic approach, taking every trade that the computer spits out, while others are more discretionary, using the systems as a guide but using discretion on the final decisions, especially when it comes to deciding how many contracts to trade. The important thing, though, is that they all have a set of rules that they know will produce high probability trades. Even top traders who never write down their rules still repeatedly look for certain setups in the market before they make or exit a trade.
What these professionals know is that a system is there to help them find trades that have a high probability of being successful. They know that by trading with a proven set of rules that have had a positive expectancy of winning, they can increase their chances of making money. Sure, there will be many times when the system is dead wrong (that's okay; if you are right 50 percent of the time, you're doing great), but in the long run if you repeatedly make the same high-percentage trades over and over again, the winning trades should outperform the losing ones. When one trades without a system, then luck becomes a big part of the equation; a proven system reduces the effect luck has on a trader's P&L. The more you can reduce the effect of luck, the better a trader you will be.
Some traders get into trouble because they make trades without any type of system or plan. They have no rhyme or reason to their trades. Each trade they make can have an independent thought process compared with the previous ones: One day they buy a breakout, but the next day, with the same scenario, they may short as the market breaks highs because they don't think it will follow through. To be an effective trader one should have a consistent plan that is followed day in and day out. By having a system, one will have precise rules to follow and will know at all times on what side of the market to be on, if any. There is a much smaller chance of making careless mistakes when one is following a system. There is little guessing as to what to do when you follow a system; the only question is how strictly will you follow the rules of the system. When you do not have a system, you never quite know what to do. Many times I look at the market and think, "We've been in a bull market, and I should be long. No, wait, it's been coming off a little and actually looks better short. Oh, but I don't know, it's already gone down too much, so I should buy." This type of trading can be reckless. With a system you won't have to make decisions like these; the system will tell you what to do.
The other important thing a system does is tell you when to get out of a trade. Some people have great entry skills, but once in a trade, they have no idea when to get out. They let losses grow too much, take profits prematurely, and give back a large portion of the profits, or, worse, let winners turn into losers. They never think about the exit when they put on a trade, and so they don't know when it is time to get out. A good system will take care of that for a trader; if he can follow his system, he will know when to get out.
Writing a system is easy and can be done in a few minutes; writing a good winning system, however, can take weeks or months to develop, adjust, rewrite, backtest, and repeat the whole process until it's just right. If you are using trading software, just learning how to program a system is hard enough; when you add all the time and hard work that go into developing a system, you can see why some people abandon formal system writing. They may start doing it but then give up quickly and either trade with a half-baked system or go back to not using any system. But if you want to get ahead, you should have a solid trading system or strategy to follow.
If writing a system is too much work or you don't know how to get started, the easiest way to get a system is to use someone else's. You can buy an advertised system that you see in the back of a trading magazine or on a Web site, but I'd be skeptical of any advertised system. I would never sell a great system; instead, I would keep quiet about it, trade with it, and reap the rewards. If I have something good, I don't want other people competing with me in entering orders. When you buy a system, it is unlikely that seller is using it as well; usually it is an older system that he doesn't use anymore.
Another problem I have with buying advertised systems is that many times the hypothetical results they tout are misleading. They say the system would have turned $10,000 into $132,000 in 3 years. Then in tiny print you see that slippage and commissions have not been included and that all profits have been reinvested. However, nobody in reality compounds his winners on every trade; it is bad money management, as one or two losses will wipe out all the profits. Ignoring commissions and slippage will give you results that vary immensely from real results. When everything is considered properly, that $132,000 may only be a $7000 profit in 3 years. Also, keep in mind that they will report results that have been optimized and tested over the data on which the system works the best. You probably could test the system on another set of data and have losing results. The last problem with buying a system is that a trader may not use it correctly if it differs from his trading ideas. Everybody has a unique trading style, and to feel comfortable one should trade with the rules that suit him best.
Don't let this discourage you. I know people who have bought systems that they were able to follow and make money with. There are some good ones out there that have proved to be consistent, and if a trader is disciplined enough to follow a proven system, he should be able to make money with it. One type of system one could buy is a black box system; this generates signals without the trader knowing how the system is made or what generates the signals. I personally can't trade like this; I always want to analyze anything I use first to make sure it fits my trading methodology. But it does work for some people.
If a trader wants to start using mechanical systems and does not know where to start, buying one may be a good option. This is also a good way to get started writing a system: Take someone else's and go through it, analyzing it step by step to see what makes it tick. Look for ways to improve on it or make it fit your style better or just steal some ideas from it. I've included some simple Easy Language code throughout this chapter from systems I've written in TradeStation, as well as a full working system. You can use some of those ideas as a basis to start making your own system. You also can find free systems on the Web and in trading magazines. A system you make yourself will be easier to trade with than will using someone else's. Even if you start out with someone else's system, play around with it until you make it your own.
For the most part, this chapter and the next one will discuss how to make and what to look for in computerized trading systems, but this material applies to manual systems as well. If you don't have software that lets you write and backtest systems, you have to do it by hand, but it should still be done. None of my early systems used a computer; everything was done by hand, and it worked just as well. A computer lets you do more and do it more quickly. I used to spend months testing systems by hand to make sure they worked even though I was told beforehand that they would; you should never take anything for granted. Ever since I started using TradeStation to write and test systems, I have been able to do things in a fraction of the time, especially backtesting. What used to take weeks now takes minutes, freeing up tons of time. When I didn't know what to do with all that free time, I decided to write this book.
MY FIRST SYSTEMS I was lucky enough to use systems from the beginning when I started to trade. Though they were quite simple, I did backtest them by hand, and they worked. My first system was using a point and figure chart I kept while in the pit. In the ring you don't have the luxury of having a computer, and so floor traders keep track of the market by hand with a point and figure chart, filling in the chart as the market moves a set amount. I had learned to make one as a clerk, and then an experienced trader showed me what patterns to look for and how to trade using the chart. Basically it was a simple breakout system that involved taking trades in the direction of the trend. I would measure exits by the width of the previous congestion or exit on a signal in the other direction. Later, when I started looking at more markets, I would keep my commodity perspective (daily paper charts) with me all the time, updating the markets I traded throughout the day. I kept track of about 10 markets and used a reversal day system that I still incorporate into my trading. That system was simple: If today's low is lower than yesterday's low, then buy on a stop if the market goes positive by a small filter that varies by market. For an exit I used the low of the day as a stop or a signal in the opposite direction. |
In making or choosing a system, there are a few things to look for. You want to make sure it is suitable for your style of trading, and you want something that is simple rather than complicated, that is easy to understand, and that works. The more complicated a system is, the more likely it is that it has been customized to the data. You also want a system that will work over different time frames and markets as opposed to one that works only on a specific market and time frame. A good trading strategy should work no matter what; if it doesn't work across the board, something may be wrong with it. As far as results go, I'll cover that in the next chapter, but you want a system that has a positive expectancy, is consistent, and has relatively small drawdowns compared to profits.
Systems should be kept as simple as possible. Overdoing things doesn't make a system better; on the contrary, it can take away from a good system. Trying to make a system too complicated with too many indicators and variables is a common mistake with some traders; some of the best systems are the simplest. As a rule of thumb, a system should fit on the back of an envelope and be easily explained so that someone can understand what every indicator does and every rule does. Otherwise it's too complicated. Always remember the old adage "Keep it simple, stupid" and you'll be okay.
There are hundreds of combinations of indicators and ideas a trader can look at and use, but for the most part traders rely on just a handful of their favorite ones. I never overdo it with indicators and ideas. I look for indicators that work in most markets and market environments and try not to get too fancy. Money management is more important to a trader's success than are all the indicators in the world; find a few that you like best and keep it simple. If you have a system with 42 variables in it, it gets a little ridiculous. The more indicators and parameters you have, the more likely it is that something will go wrong, and when it does, you won't be able to figure out what it is. As people put more and more rules into a system, it becomes curve-fitted to the data and too difficult to analyze when one is trying to improve on it, as there are too many parameters to look at. People sometimes get carried away writing filters to make their systems catch more trades or avoid others. This is something that should be avoided. Filters should be kept simple, as the more you have, the less likely it is that the system will work as well in the future. If you have a simple system that worked in the past, it is likely that it will work in the future and in different markets better than a complicated curve-fitted system would.
One of the most important things about a system is that it has to fit a trader's style and beliefs about the market. One person may have a great system that works like a charm for him, but if another trader uses it, he may lose money with it. This is the case because the other trader may not believe in the signals it gives if they differ from his trading style. Some people, for example, will buy as a market breaks out of a range, but a trader who likes to look at stochastics may second-guess this and not buy because the market looks overbought. Some may like to hold trades for only a few minutes, while others hold for a few hours; this is a personal choice, and it is hard for people to switch holding times once they have set beliefs. I know I am not a quick scalper no matter how hard I tried to be one in the past; if I have a trade that's working, I like to hold on to it. I allow for this in my systems, and make systems that hold on to winning trades longer. A trader with no tolerance for loss may be more comfortable using stochastics in his systems, while a trader who is looking to hold for the long term may have a moving average-based system that keeps him in a trade longer. Some traders just can't short; they feel comfortable only buying. If that is the case, they can make a system that will only go long, ignoring all sell signals.
Rule number one of any system is to make sure you are comfortable using it and believe in the signals it generates. To do this you need to know what kind of trader you are and what makes you tick; you need to analyze your personality and figure out how you want to trade. Are you a person who is in it for the excitement or to make money? Would you be happy trading once a week if you knew you'd make money or do you need to trade 50 times a day? Do you like to trade reversals, trends, or breakouts, or do you have another preference? Do you have a full-time job so that you can't watch the market all day? If so, you may need a system where orders can be placed before the open or at night instead of one that makes trades in the middle of the day, which a full-time trader may use. No matter what type of trader you are, you will work better when you are comfortable with a system that complements your beliefs.
As you try to figure out what kind of trader you are and what types of systems you will need, ask yourself the following questions:
How often do you need to trade?
What time frame do you like best?
Are you more comfortable trading with the trend?
Do you like trading breakouts?
Are you a contrarian trader, always looking for reversals? What indicators or pattern do you like to use?
Are you an aggressive trader, or a risk-averse trader?
Can you trade a boring market or stock or do you need fast movers?
Can you hold positions overnight?
Do you need to get out by the close in order to be able to sleep at night?
Are you too high-strung, or are you laid back?
Do you worry about every tick?
Can you let a trade breathe to maturity?
Do you want lots of small wins or are you looking for big moves?
How big of a position can you handle?
Are you looking to make a living or to have some fun and make a few bucks on the side?
How much are you comfortable losing per trade?
Do you know how to lose?
How often do you need to be right?
What percent of your equity can you handle as a drawdown?
How much can you afford to lose per trade?
Once you can answer these questions honestly, you can start to come up with trading ideas and systems that will suit you and that you'll most likely be able to follow.
There are many different trading approaches that work. Finding the one that a trader feels the most comfortable with is what's important. The next few sections give examples of some different types of systems you could use and can get you started on the path to writing a system or two. I've included the Easy Language code for TradeStation users.
The oldest, simplest, and most effective systems are breakout systems. The reason these systems work well is that they will get you into a trade at the start or during the continuation of a major trend. Every trend or major move starts with a breakout of the previous high or low, and if you want to get in on it, a breakout system is for you. A trader using these systems must be willing to be wrong quite often, as they will produce many false signals that can cause one to buy many highs and sell many lows. The key to making money with these systems is that one or two breakouts will be quite substantial and can more than make up for the false signals. Breakout systems are best suited for a patient trader who can wait for a retracement and then hold on to winning trades as long as possible. If you get out too soon, you could end up missing great trades and come out a loser in the long run. Many people like to use stop orders when getting into the market on a breakout. I don't think it is always a good idea to do this because sometimes it's best to wait for the pullback if the market is already overbought. What I do is have a system that alerts me when the market has met the breakout criteria and then I use a different system on a smaller time frame that waits for a pullback so that I can get a higher probability trading opportunity.
The easiest type of breakout system to program is the kind that gives signals when the price exceeds the highest high or lowest low of the last x periods. I will cover exits and stops later in the chapter; for now I'll just worry about the entry signals.
TradeStation users would write the long entry signal as follows:
Input: Length(10); If Close > Highest(High,Length)[1] Then Buy On Close;
This buys when the close is higher than the highest high of the last 10 (Length) bars beginning one bar ago ([1]).
The Input: Length(10); allows you to easily change the lookback period of the highest high when you are using the charting software without having to go back and change the code. All the inputs for a system go at the top of the system. The [1] ensures that you start counting from one bar ago so that you don't include the current bar in your signal. Since you don't know what the high of the current bar is until the bar is closed, you shouldn't include it in your signal.
If you wanted to get the buy signal before the close of the bar, you could write the signal as follows:
If High > Highest(High,Length)[1] Then Buy;
One way to avoid receiving false signals and getting whip-sawed is to add a filter to the system that gives you a buffer zone before getting a signal. You can do this in several ways. One way is by adding a few points to the breakout as in the following example, which buys when the close is higher than the highest high of the last 10 bars plus 5 points. This prevents you from getting in if the breakout was just a quick peek through the previous high:
If Close > Highest(High,10)[1] + 5 points Then Buy On Close;
You can also add a buffer that's based on the volatility of the market. Volatility changes with the market, and so as the market gets more volatile, a breakout has to be more significant to be taken. This is done by adding a standard deviation measure to the system, which ensures the market has moved enough outside the breakout level and that this has not happened by chance. If you want to consider the volatility of the market as a buffer, you can add a standard deviation move from the highest high. You can either write it directly into the line of code or make it a separate variable, as I did in the following case:
Buffer = StdDev(Close,10)[1] If Close > Highest(High,10)[1] + Buffer Then Buy On Close;
Another buffer in a breakout system may be to buy when the market closes above a 35-period moving average for 2 days in a row. This ensures that the break of the moving average wasn't a 1-day fluke. You can add as many periods as you like to this kind of buffer:
If Close > Average(Close,35) And Close[1] > Average(Close,35)[1] Then Buy;
You also may want to throw in a condition that takes a trade only if volume is getting strong. When a system has more than one requirement to get into a trade, you will need to set up conditions and then tie them together to get a signal. In this case something like the following would work:
Inputs: Length(10),LengthV(5); Condition1 = High>Highest(High,Length)[1]; Condition2 = Volume>(Average(Volume,LengthV)*1.25); If Condition1 and Condition2 Then Buy On Close;
The first condition is the break of the highest high of the last 10 bars, as in the first example I gave. The second condition requires a filter that the volume of the current bar is 25 percent greater than the average volume over the last 5 days. This will give you signals only on trades that break out of a range on better than average volume, which is something you want to see, as it may indicate the break may follow through.
Some breakout systems require patterns that are hard to program, such as channels, trendlines, and double bottoms. With those systems, you will have to look visually at a chart to get signals, but if you come up with a set of rules for a trade, it is still a system even if you can't program it.
For those who want to trade with an established trend, moving averages and trendlines will be the core of their systems. Since trendlines are hard to program, one is better off relying on moving averages in programming trend-following systems. Those who use patterns, such as channels and trendlines as a trading guide most likely will have to make visual systems or draw their trendlines on a chart so that the software can compute signals.
As with breakout systems, those who can hold on the longest will fare the best with these types of systems. In a choppy market you should expect to be wrong a lot. One way to avoid getting whipsawed is to use longer period averages; the downside is that when a trend does start to develop, you will get a signal later than you would if you had a shorter average. This trade-off is something a trader must determine. There is a little time lag between the time a moving average system shows you that a trend has started and the time it actually started. Moving averages are lagging indicators, and so any system that uses them will get you into the market only after a move has been established, but if a strong trend develops, these systems can catch a good portion of the move.
The most basic moving average system would be a two-aver-age/crossover one, buying when the shorter average crosses over the longer one:
Input: Length1(10),Length2(35); If Average(Close,Length1) Crosses Over Average(Close,Length2) Then Buy On Close;
This will give a buy signal when a 10-period average crosses above a 35-period average. Though the signal is given at the close of the bar, one would place the order at the beginning of the next bar. This is the case because you never know for sure until the actual close if a signal will be given. Sometimes it looks like it may, but in the last few minutes the market moves drastically the other way. You could rewrite this so that you are buying on the open using the information from one bar ago, which would read as follows:
Input: Length1(10),Length2(35); If Average(Close,Length1)[1] Crosses Over Average(Close,Length2)[1] Then Buy On Open;
One condition in a system could be to take trades only in the direction of the 50- or 200-day moving average; this will always put you on the side of the major trend. You can find the direction of the average by having a condition that let's you see if the moving average is higher now than it was 10 days ago or whatever lookback period you like. If it's higher, it's moving up:
Input: BarsBack(10) Condition1 = Average(Close,50)> Average(Close,50)[BarsBack];
Another condition could be that you want the current bar to be as close as possible to the moving average before buying; otherwise you may buy an overextended market that could eventually pull back in toward the average. I would do this by adding a condition that says that the price must be within one average true range (ATR) of the moving average. You could use more or less than one ATR, or you can use standard deviation, or points. The following condition ensures that the current close is less than one ATR from the 35-period moving average:
Input: Length2(35), ATRlen(10); Condition2 = (Close — (Average(Close,Length2)) < AvgTrueRange(ATRlen(10));
There are many more things one can do with moving averages. These are just a few examples; the rest is up to the trader's imagination.
No matter how many times you tell people that trading with the trend is the way to go, some will insist on looking for bottoms and tops. Systems based on oscillators are good for these counter traders, who are always looking for reversals, bottoms, and tops. When a market is stuck in a range between support and resistance levels, an oscillating system that buys at oversold levels and sells at overbought levels works better than does anything else, especially in the short term. Oscillator-based systems also can satisfy the needs of a person who always needs to be in the market. There are many ways to write signals based on oscillators, but trying to come up with a computerized signal for the best use of an oscillator–divergence between the market and the oscillator–is not easily done. Divergence is something that has to be seen visually on a chart, but there are many ways a trader can use oscillators in computerized systems. Here are a few signals and conditions that could be used in oscillator-based systems.
The most typical signal using the stochastic lines is to buy when the SlowK crosses over the SlowD. This is written as follows:
Input: Length(14); If SlowK(Length) Crosses Above SlowD(Length) Then Buy On Close;
One also could add a condition that the stochastics has to cross above the oversold area (BuyZone) while the SlowK line is above the SlowD line. Since the SlowD line will be below the SlowK line, you only need to ensure that the SlowD indicator has crossed a certain level to get a signal:
Inputs: Length(14), BuyZone(30); If SlowK(Length) > SlowD(Length) and SlowD(Length) Crosses Above BuyZone Then Buy On Close;
Using the RSI, one could use a similar crossover rule:
Input: RSILen(10),BuyZone(30), If RSI(Close,RSILen)Crosses Over BuyZone Then Buy On Close;
You also could change the area where a signal would be given to any number. If you want to buy only if the indicator goes above 50, change the BuyZone parameter from 30 to 50.
If you want to avoid buying when a market already has made a move and is in overbought territory, you can add a condition to your stochastic system that will take trades only if the indicator is below the overbought area. This will keep you from buying potentially overbought markets. This condition could be used in any type of system to potentially keep you from chasing a market:
Input: SellZone(70); Condition1 = SlowD(Length) < SellZone;
The last example I'll give is one that can be used in a strong market. This one is used to buy when the oscillators are in overbought areas, as the market could stay up there for quite a while. This is the opposite of selling when the market is overbought, but in different market conditions one strategy works better than another. You may want to combine one of your entry rules with a rule that takes the average directional index (ADX) into account. The following will buy when the ADX is strong indicating a strong trend and confirmed by a strong reading in the stochastics:
If ADX(10) > 30 And SlowD(14) > 85 Then Buy On Close;
I'll get to exits soon, but with the above example you may want to have a rule that exits the trade if the stochastics drops below 70, which may indicate that the strong trend is dying down. Again, these are just a few ideas one can use with oscillators. Go through Chap. 7 again to get more ideas and then use the ones you like best in writing a system.
Traders need to be flexible to adapt to different market conditions. In a choppy market one set of rules may work better than it will in a strongly trending market. If I see that the market is choppy, I stay out or use a system that is based on stochastics instead of moving averages. I especially avoid anything that has to do with breakouts. I won't buy highs but instead look for the market to reverse at the extremes. In contrast, when the market is strong, I use trending indicators and rely on oscillators to confirm the trend or to wait for a pullback. I avoid trying to pick tops in a strong market but may use oscillators to buy when they indicate that the market is oversold.
One thing that can help in deciding what type of system to use is to employ the ADX to help determine whether the market is trending or trendless. A trader can have one set of rules if the ADX is below 20 and another if it is above 30.
This could be done as follows:
If ADX(Length) > 30 Then Trending Market System Else; If ADX(Length) < 20 Then Choppy Market System Else; Middle Ground System
A system is not complete until it has stop and exit parameters in it. It is important to work as hard on the exit side as you do on the entry side. A system with just entry signals is like skiing downhill for the first time: You can get started and on your way easily enough, but you may have to hit a tree to stop, which is not lots of fun, trust me. Money is made on the exits, not on the entries, so spend as much time working on the getting out part of your trades as on the getting in part. Though I do have rules built into my systems to get out of the market, I very often end up setting my exit parameters visually, using different time frames to do so. Sometimes I get out of positions before I get an exit signal because a position just stopped performing the way it should. There is no need to wait to get stopped out or to give back profits. When a trade stops working, I get out. If it returns to the right path later, I'll get back in, but for the time being there is no need to risk a pullback.
The simplest way to make sure you account for stops and exits is to have a stop and reverse system, which means the system is always in the market. These systems will have you either long or short and use a signal in the opposite direction as an exit/stop. The problem with this is that on short-term systems there will be many times when you are trading against the major trend of the market. Though I used to use stop and reverse systems frequently, now I prefer to sit out when I get a signal that goes against the main trend.
I have a standard stop I use in all my systems. It's my canned stop that I will use if I have not gotten out of the market beforehand. The stop I use gets me out of the market if it has moved 2 standard deviations against the spot where I got in:
Exitlong From Entry("Buy1") at$ Close -
2*Stddev(Close,10)[1] Stop;
at$ Close refers back to the close of the original entry bar. To use this type of stop, you have to have named the entry signal. The following example names the entry signal "Buy1":
If SlowK(Length) Crosses Above SlowD(Length) Then
Buy ("Buy1") On Close;
Other stops could be to get out when the market has dropped below a moving average by some buffer zone:
If Close < (Average(Close,Length1)-Buffer) Then
Exitlong ("Stop1");
or if the market has just broken lower than the 5-day low:
If Close < Lowest(Low,5)[1] Then
Exitlong("Stop2");
Besides a signal in the opposite direction, some popular exits are to get out after x number of bars:
If BarsSinceEntry = 10 Then Exitlong;
or if the stochastics is in overbought territory:
If SlowD(Length) > 85 Then Exitlong At Close;
or if the market has moved too far from its trendline or moving average:
Input: SD(5), Length(35), Period(10); If (High - Average(Close,Length)) > StdDev(Close,Period)*SD Then Exitlong At Close;
This exit will get you out when the price moves more than 5 standard deviations from the moving average. You would want to do this because when a market moves too far from its support level, eventually it will tend to snap back. These are just a few examples. You have to test different ideas to see what works best for you. You also may have more than one set of exit rules and stops; if any one of your conditions is met, you can get out.
You don't have to use only one system. Some people trade many systems on the same stock or commodity at the same time. They may trade five systems and take every signal they get so that at times, if all their systems are indicating that they should get in on the same side, they may have five contracts on. If they are getting mixed signals, they may have no positions on, as the signals will cancel each other out. It's a good idea to have several systems because some will work best in choppy markets while others will work best in trending markets. By having one for each type of market environment you may always be involved no matter what the market is doing. When a bunch of systems start working in sync, the odds of a trade working are probably better. When this happens and you take all the signals, you will be trading with more share size, and so you can capitalize nicely on it.
There is a debate about whether system traders should be purely systematic or have discretion in regard to what they trade. Purely systematic traders rely on backtested systems and rules to generate every trade they make. They like having clear-cut signals and will not second-guess the system. They ignore all emotions and thoughts about the market and are disciplined about making a trade as soon as the system alerts them to. Discretionary traders may take some trades and not others; they may use the signals as an alert and then use discretion to time trades, especially if the market has had a run-up. Some people may use discretion when they are trading patterns that can't be programmed, and some may not use a concrete system at all. Regardless, good traders who are 100 percent discretionary, still have solid buy and sell rules they adhere to every time.
One problem with using a system and not taking every trade is that the one trade you decide to skip could be the one that would have made up for the last five losses. People who do this end up blaming good systems for their losses, even though the system actually ended up performing well. It was the trader, not the system, who was responsible for the loss by deciding not to take every signal. Though it's okay to use discretion in trading, a systematic trader shouldn't get into the habit of overriding the system whenever the mood hits. If you are using a good system that has been backtested and you want to be a system trader, you have to take every signal regardless of what you think about each trade, because you never know which will be the one that works like a charm. However, as I mentioned before, it is impossible to program some proven patterns and methodologies into a system. A good trader can see patterns in the market that a system cannot capture. It's hard to program things such as what wave in an Elliott wave pattern is the market in or if the market has retraced 38.2 percent yet. Head and shoulder, cup and handle, saucers, and flags are some patterns that are close to impossible to program in TradeStation. Some of these patterns make for great low-risk trades, and I am always on the alert for these kinds of trades. Sometimes you'll know about an upcoming report that you'd rather be on the sidelines for instead of taking undue risk. Hunches are also impossible to program. There are many times when I'm in a trade and don't feel right about it, and so I get out even though it hasn't hit my stop or target area. One time you may want to override a signal is when there is news that makes the market jump and leaves a big gap. A system may generate a signal, but do you really want to take it if the gap is large? The difference between where the signal was generated and the current price could be several hundred dollars per contract or share. When this happens, it may be best to do nothing and wait to see if the market gives you a better opportunity to get in. You still want to trade in the direction of the signal, but there is no need to rush in.
Overall, it's hard to say whether one should use a fully systematic approach or not. People can make money both ways and lose both ways. The only thing I know is that if you have something that works, don't second-guess it.
Finding or making the right system is not easy, and people make many mistakes along the way. Aside from using a system that doesn't suit a trader's style, there are several mistakes traders make when they are using a system to trade with. Though I'll discuss them in more detail in the following chapter on backtesting a system, some of these mistakes include not having enough money to trade a system, giving up on a system too soon, trading a system that does not have a positive expectancy, trading a system that has been curve-fitted, and trading a system that has not been backtested properly or with enough data. If a system is not properly backtested to have a positive expectancy, don't use it; it can cause a lot of unnecessary losses. It's amazing how many people trade a system without ever backtesting it first. Even when a new system has been backtested properly, don't risk a lot of capital on it until you have tested it with real money for a while. Start small or in a market that has low volatility until you are sure the system will perform well in real time before risking your normal amount.
Some people keep working on a system forever, but they are never happy with it; they keep changing it or adding to it, trying to make it perfect. They spend too much time writing it and never get around to trading the system. You can avoid doing this by having a goal for the system. Know what you want before you start. If you have well-defined goals for your trading, finding or making a system will be easier. If you are looking for a system with a 55 percent win/loss ratio and with winning trades being twice as profitable as losing trades, with a drawdown of no more the $3000 and 5 percent per month profit, when you get close, be happy with it. Don't try to keep stretching it or you will never get to trade it.
Here is a simple system for TradeStation that works. It is based on the signals I explained earlier. Everything inside brackets {} does not get read by the system; it is there to make it easier to understand. This sample system should give you a good idea of what goes into writing a system and what it should look like.
This system buys when the market breaks above the highest bar of the last 10 periods, with a filter of a 0.5 standard deviation move of the last 10 bars. The sell signal is just the opposite. The entry is simple enough, but I made the exit more interesting by using the ADX to give me different exit conditions. If it is strong and the market is trending, the system will stay in until two moving averages cross; if the ADX is weak, the system takes profits after 10 bars; and if it is in between, it will exit if the stochastics reach overbought territory. Finally, the system exits on a stop if the price is more than 2 standard deviations from the entry price.
Becoming a better trader means using some sort of system in your trading. Whether it's computerized or done visually on a chart, is simplistic or elaborate, a trader should have a solid set of rules that have been proven to work. Unless a trader can start making the same smart trades over and over, he not only will probably lose, he will not know the specific reason he is doing so. With a system, if you are not making money, it is because the system is no good or the trader can't follow it. If the trader can't follow it, he should get one that fits him better. If the system is not good, he should be able to see that and abandon it or fix it. By backtesting a system before using it, a trader can reduce the chances that he is trading with a bad system. The problem when you don't have a system is that you may have no idea why you are losing, as your trades may be made randomly.
Even when you are trading with a solid computerized system, you can still use discretion on some trading ideas, as you won't be able to program every good strategy. There will be times when things don't feel right and you want to get out before a signal is given; as long as you are not cutting winners short on a repetitive basis, this is fine. Overall, though, if your system works, you should try to take all the signals it generates as you never know which will be the good ones.
One thing that is important about systems is that no system is just a signal to get into the market; it should always provide you with exits and stops as well. Even if you use visual stops and not programmed ones and do so all the time, they become part of your system. Either way, it's important to incorporate them into your trading. What's good about having the exits already established is that you don't have to think as hard or worry about getting out too soon or too late; you let the system take care of that.
Making and backtesting systems is not easy work, but those who want to get ahead will take the time to do it, as it is crucial for making consistent high probability trades. System writing can be quite time-consuming, but a successful trader will put the work into doing what it takes to come out ahead.
Why People Lose Money with Systems
1. Abandoning a system too soon
2. Ignoring commissions and slippage
3. Lacking the discipline to follow the rules
4. Second-guessing the signals
5. Relying on misleading hypothetical results
6. Using a system that doesn't fit your style
7. Using a system you can't afford to trade with
8. Using a system that hasn't been backtested properly
9. Adding too many variables and conditions
10. Curve-fitting systems
11. Not keeping things simple
12. Trading with a losing system
High Probability System Trading
1. Use only systems that have a positive expectancy of winning.
2. Learn to backtest a system properly.
3. Use systems that adapt to different markets.
4. Have solid rules to trade by.
5. Work as hard on the exit as on the entry.
6. Include stops in your system.
7. Use a system that has consistency of returns.
8. Keep things simple.
9. Use a system in a higher time frame to alert you of trades and to monitor stops.
10. Use a system in a smaller time frame to time entry and exit into the market.
11. Use systems with a small relative drawdown.
12. Know the most you can lose.
13. Look for signals in different systems to confirm a trade.
14. Use discretion when it is needed.
15. If you are going to be a true systematic trader, take every trade.
Helpful Questions to Ask Yourself
Do I really have a system?
What is my system?
Is it too elaborate?
Does it have stops and exits?
Has it been backtested properly?
Do I believe in it?
Do I second-guess it too often?
