Manuel Ochoa
Submitted by (TRIAL USER):
How do know whether a system is “breaking down” or just going through a drawdown period?
Manuel Ochoa:
When you are a trend follower and a market has a big move up or down and you either break even or lose money. The system was supposed to catch big moves in the market and yet it didn’t because of either to many entry rules or poor stop placement.
Submitted by Trial User:
Is most of your trading systematic or descretionary?
Manuel Ochoa:
It is about even. The best traders in the world use both. There are good purely mechanical traders but the guys that have been around the longest with the best risk reward ratios are the ones that use both. The systems keep them grounded and the discretionary part tells them when to lighten up on positions. That is usually the way it plays out.
Submitted by trader21:
Can you explain what you mean by “optimization?”
Manuel Ochoa:
In most historical testing software, there is a module where you can see what the results would have been in the past had you used differnt inputs for the SAME indicator. Optimization is abused when the user simply picks the best set of inputs for the indicator thinking that these inputs will perform the best going into the future. I have found that what worked best in the past probably will NOT work the best in the future.
Submitted by dbundy:
I like to do short-fuse option trading (intraday and no hold longer than four days). Have you developed a trading system for this type of strategy?
Manuel Ochoa:
No. I believe that options have to high of transaction cost to be effective. By that I mean the commissions are to big a percentage of the value of the option and so are the spreads. I would look into shorting premium. I know very successful traders that do this. The only downside is that you need a lot of captial but it is worth it.
Submitted by cmattes:
How much price data should you test a system on? Or do you think in terms of generating a certain number of sample trades?
Manuel Ochoa:
Ideally you want both but either one is good if you can get it. You want your data to span a period of years so that it will reflect the different economic cycles of the past. Markets behave very differently depending on what is going on in the economy. Just look back at the ’70’s and the high inflation and see the kinds of moves you got in the markets back then. It was dramatic to sey the least. Today’s environment is much more tamed because of the lower interest rates around the world. Get the most data you can.
Submitted by ASLEEK1:
What is required to make and test systems?
Manuel Ochoa:
Three things: 1. Some type of general knowledge on the ecomomy and the markets. Read the news papers. 2. Historical data to plug into your testing software. There are many vendors. 3. The testing software itself.
Submitted by Trial User:
Are there markets in which you prefer to trade discretionary approaches and other markets in which you prefer to trade systems?
Manuel Ochoa:
Yes. Since I was a business student in college, I understand the financial ecomomy such as Bonds, stocks, and currencies. When reports come out on these markets I can make some sort of opinion out of them. On the other hand, when reports come out on the grain markets I don’t feel as comfortable doing this. That is not to sey that I don’t trade the grain markets with some discretion, I just dont do it as much.
Submitted by Moderator:
Welcome to the tradingmarkets.COM Live Forum, featuring Manuel Ochoa.
Manuel is a hedge fund manager who began researching and developing trading strategies and models while he was still a student at the University of Southern California, from which he received a BS in Finance in 1990. Manuel’s market commentary, “Financial Futures Insight,” appears every Monday, Wednesday, and Friday at 5:30 p.m. ET on tradingmarkets.COM. In it, he talks about what he’s doing in the T-bond, S&P 500, and currency futures markets, and what’s making these markets tick. Manuel has enjoyed great success trading both his hedge fund and his personal account since launching his career immediately after graduating college. He has been profiled as one of the hottest new traders in Futures magazine, and one of his funds was ranked third by Mar Hedge for its performance in 1997. He continues to develop trading systems and portfolio diversification and risk management models. Today, Manuel will begin the forum by discussing how trading system design and testing. To ask a question, simply type it in and hit the “Submit Question” bar–that’s all there is to it. You also can create a short subject heading for your question in the title space (it helps if you do). Past questions appear in the left-hand portion of your screen for easy browsing. This is a moderated forum, so we ask that you respect the other guests and our featured speaker. We try to get as many questions as we can, so please be patient. Shortly after the forum is completed, it will be archived and available for review.
Submitted by Moderator:
Is a system good if it “works” in only one market, or should it show profitability in a wide range of markets?
Manuel Ochoa:
It always better to have one that works in a variety of markets. It means the system has a better chance of making money in the future.
Submitted by mlbv:
Do complex trading systems with many rules tend to perform better than simpler systems?
Manuel Ochoa:
Usually not. Complex systems are usually the result of over optimized system developments. The past is never just like the future. This is very important to realize when you start testing out your ideas
Submitted by epkes:
Please explain shorting premium.
Manuel Ochoa:
With options there are three possible outcomes: 1. the market moves up and you make money on your calls. 2. The market goes down and you lose money. 3. The market goes nowhere and you still lose money. When you short premium you have 2 of 3 scenarios in your favor. The probablities are on your side.
Submitted by kenka:
How do you know whether a system is “good” or not? Are there specific statistical measures or benchmark performance figures you look at?
Manuel Ochoa:
There are many but I usually look at: 1. Max Drawdown, the most amount of money the system lost from the peak dollar amount in the account until the system stopped losing money. 2. The percentage of winning trades: This is key because it lets you know what to expect in the future. Are most of your trades going to be winners or lossers. This is important to know before you start trading a system. Those are the basics.
Submitted by ccolson:
Can a trader be successful using only one system?
Manuel Ochoa:
Yes, but he wont have as good a risk reward ratio as one that uses multiple systems. When you combine multiple systems you have the benefit of adjusting to changing markets on a day to day basis rather than on a less frequent basis.
Submitted by Moderator:
Could you describe the difference between a “robust” system and a “curve-fit” system?
Manuel Ochoa:
A robust system is one that you test out on forty different markets and it makes money on all of them. The over optimized system is the one that makes money in two or three and fails on the other 38 markets.
Submitted by bknapp:
How many parameters would you say the typical “robust,” profitable system would have?
Manuel Ochoa:
Three or four. Any thing more is suspect.
Submitted by valky:
How do you determine realistic commissions and slippage amounts to factor into your system test?
Manuel Ochoa:
It depends on the market. A super liquid market like Bond futures has very little slippage. On the other hand, a market like Lumber futures or a small cap stock can have tremendous slippage. You have to investigate it. Call your broker and get the spread on these markets to get a feel for them.
Submitted by magnum:
You mentioned that option commissions were too high. Would $15 (in and out) be too high in your opinion?
Manuel Ochoa:
Probably, because if you look at the percentage that the 15 dollars is to the price of the option and then you compare that to the percentage that it would be to a futures contract or a share of a fifty dollar stock you will see what I mean. When you trade slippage and commissions are your enemies and they will reduce your profits by significant amounts over periods of time. Just do the math and you will see.
Submitted by Moderator:
This concludes our Live Forum. Thanks for all your questions, and thanks to Manuel Ochoa. Be sure to check out Manuel’s commentary every Monday, Wednesday, and Friday at 5:30 PM ET. Come back next Thursday, May 6, for our next forum.
This forum will be archived shortly for easy review.
Submitted by Moderator:
What about disadvantages to system trading?
Manuel Ochoa:
Trading systems can be harmful if you don’t understand the logic of them. It is good to follow a system but on the other hand there are times when system start to break down. If you don’t understand the logic of them you will probably not be able to identify this situation.
Submitted by philrock:
Is there any way to know whether a trading idea is a sound basis for a mechanical system? Do you have any personal criteria?
Manuel Ochoa:
The main criteria is that whatever idea you have has to have some sort of logic. For example, if you trade Bonds you know that the Fed has little control over long term interest rates. They are more dictated by perceptions of the market. This being taken as a fact, you can then discern that trends will tend to develop in the long end of the Bond market as perceptions change. You can then use some sort of trend following technique to capatalize on this characteristic.
Submitted by johna:
What do you think are the most common mistakes people make when designing and testing trading systems?
Manuel Ochoa:
Not understanding the behavior of the markets. What most rookies do is test systems based on technical studies and then optimize the indicators that made the most money in the past. It turns into a numbers game with hindsight being the main decision maker.
Submitted by dbundy:
What do you feel are the most important ingredients in developing a trading system?
Manuel Ochoa:
If you are a trend follower, it would be a filter such as a moving average to determine whether you should be look to go long or short. It will keep you on the right side of the overall trend. Then you can make your rules for the actual entry.