Here’s the webinar I’ve held about 6 months ago at Asia’s First Virtual Trading Summit, where I talked about my step-by-step process to becoming a proficient Systematic Trader.
Philip: Welcome back to Online Trading Summit. Our next speaker is Andrea Unger who is renowned for being the only trader to ever win The Futures Trading World Cup Championship three years in a row. A few years later he went onto win the fourth championship to become the tournament’s only four-time winner.
In this presentation, Andrea shared how he got started in systematic trading and illustrated his comprehensive six-steps process to building a systematic trading strategy. There is tons of systematic trading wisdom from him so make sure you hang on to hear what he has got to share.
Philip: Hi Andrea!
Andrea: Hi Philip! Thanks for letting me be here.
Philip: Yeah! Welcome to the Online Trading Summit and we are really very honored to have you coming on board to share with the Summit participants about your experience and your trading knowledge and I am very excited and I think that this is going to be a really great session, but before we go into the part where you’re going to share with us how to become a trading champion, so maybe just very quickly would you be able to share with me and the Summit participants how in the first place did you actually get started in trading? When was that? How long ago was that and how did it evolve in terms of your trading experience along the way? Maybe you can just quickly share with us something about that?
Andrea: Yeah actually that’s a long time ago. It was in the past century! I was still working. I am a mechanical engineer. I have a Master of Science and I was working in some multinational company’s technical marketing, and obviously with our colleagues we were interested in making some extra money. So we started investing in some — in the stock market, mainly based on rumors. We even purchased the shares of the company where I was working for. I started getting interested in that. That was maybe 1997, something like that, and we did make money at the very beginning so I felt very smart! I lost all that money and I lost more of that, but I didn’t want to quit because I didn’t like to lose money. It was more sort of a losing thing that I didn’t accept. I felt to be smart enough to figure out how markets work. So to see that I losing the money was something that was upsetting me, not only for the money but also for the fact that I got a failure, you know?
So I started studying. I started studying and trying to understand. That’s why I started studying. I discovered technical analysis and I started more or less on all of those topics most of you probably studied – [inaudible – 03:11], moving averages, indicators, oscillators, all things like that, and I kept on losing money but slowly I started improving. Slowly, slowly… then I discovered some inefficiencies on the Italian market with some instruments. They were called [inaudible – 03:38] and they were options, options with market makers, and these market makers had some slow software so that their codes were coming out a bit late, maybe five to thirty seconds later that they were supposed to be based on the underlying moves. Being able in maths and making some quick calculations, I was able to figure if there was room for an arbitrage – and there was.
So I started making money. In the same time, I realized that the banks, the market maker, would not keep on forever in going out with such poor software and I was trying to prepare myself for the future. What shall I do when this thing is over? — because in the meantime I also quit my job. So I became a full time trader. I was making money, so I became a full time trader. Like I said, being a mechanical engineer I felt that even if after one or two years I realized that it was not for me I could still be able to find a job. So I was relaxed enough to start this adventure. I started studying. I understood that the best thing to do for me, for my mindset, was to try to find a way in automated trading, developing trading system based on backtesting and on numbers because having a mathematical mind, I was looking for numbers and fixed plans. So I started developing and I tried to put the rules I knew from common sense into the machine. It was a disaster.
Andrea: Then I started trying to mix all sorts of indicators together to find the best ever setup and that was another disaster. So I started thinking it was impossible. I may never find my way! But I didn’t want to quit, and I had a friend, I still have a friend, and he said “Look, I know you are a smart guy. I want you in my team to help me in developing strategies” and he passed a very simple piece of code over to me. This is the starting point. We developed on this. What I thought about this was “Can it be that simple?” — and it was!
It started from that tiny piece of code. I built all of the environment you probably have heard about today because starting from there I understood that it was possible to develop strategies based on a simple process where you put all things together. It’s like when you build a cathedral, you put all the bricks together. You have a plan, but you put all the bricks one after the other and at the end you have your building, your cathedral or whatever you prefer to imagine. I started with that and then having the first profitable strategies I was able to work starting further with a more relaxed attitude because I already had something that worked and I had just to think about other ways of making profitable trading systems.
So that was one way, year after year, adventure after adventure I found other models to build systems out and today it’s what I use. I have some basic models and on these basic models I put strategies, with rules one after the other. I will show you later on an example of my approach. How I start from something and then I build brick after brick the final product which is then what you put live. It’s not that difficult once you get the right ways to go. I decided to go with systematic trading because I believe that it helps psychologically. Obviously once everything is in the machine and running and you made all the decisions in advance then it’s easier to stay in the markets, because when you are a discretionary trader you have to make decisions during the activity and during the activity you are under stress. If you are losing or even if you are making a lot of money, you feel stress and the stress doesn’t help obviously in making the right decisions.
While you are in systematic trading, you measure your idea with backtesting and once you have measured your idea and you believe it’s a good idea that works, you just put it live and the machine does the job for you. In addition, the computer can handle a number of tasks whereas human beings can trade in any one to three markets, certainly not thirty markets or forty or fifty! The machine can. So once you have developed all the systems, they can work. They don’t need our direct mouse-clicking to go ahead.
And also, I remember once I was purchasing some shares and I entered the market with a supposed size of 10,000 shares and it’s so that as the very moment I clicked on the mouse, the stock jumped up the prices. “Oh! I am so good. Look! I am entering and it immediately boosts!”. When I saw that it was something weird and then I discovered it was me moving the share —
Andrea: — because I made a mistake. I entered 100,000 instead of 10,000. So those 100,000 size was actually moving the prices up, so I was pushing prices to the sky! But I had a problem because then I had to get rid of the extra shares, trying to lose as small as possible. So I was really struggling with that situation. I was able to manage it somehow with some patience, but based on that I also understood that a computer would not have made at this stage, because once you program the size into the machine it doesn’t add a zero or things like that.
Philip: I see.
Andrea: So I mean just one more reason to prefer systematic automated trading to discretionary. There are many other reasons, but this is just some to explain to you and to all the listeners why I actually prefer a systematic approach to the classic or discretionary approach.
Philip: I see. So was all of this systematic trading implemented even before you joined all the trading competitions and during your competition were you actually using systematic ways to trade?
Andrea: Part of it, yes. Part of it because I am still learning and developing today. So it’s an ongoing process, but as you mentioned the competitions you know I won the World Cup Trading Championships in 2008, ’09, ’10 and 2012 and I did that using trading systems. The first additions were with tradings that are running on my computer but the orders were still placed manually, because technology was not to the level it is today at that time ten years ago. But then I implemented in 2010 the whole totally automated structure and I had a trading assistant. Those competitions helped a lot in developing my skills as well because in the competitions I learned how to react to markets which were changing.
Let’s say the behavior of the markets in 2008 was completely different from the behavior in 2009, and again even behavior in 2009 and so on in 2010 and in 2012. So understanding that the markets were continuously changing and having to react to find the right solution quickly, because we are a one-year contest. You have to compete over 12 months because you have to make the best performance in 12 months. So I was able to increase my adaptability with the markets.
My skills to react on time as quickly as possible to figure out what’s going on if something is changing and what to do – which strategies should be in place? On the contest you have a small account. It’s all real money but still small money. It was about $15,000, US dollars, in the very start and then they went down to 10,000. So you cannot try fifty strategies. You have to choose the strategies you believe should be those that will do best. So that is very useful to be able to stay in touch with the markets. So I suggest you to do that. Not to win, I mean obviously he who participates wants to win but it’s also something that can increase your skills in terms of making the right choices.
Philip: I see, I see. Great! So I think right now we are all looking forward to have you share with us how you actually went about achieving that kind of really good results in your competition. So I’m going to hand the stage over to you right now Andrea and I look forward to you sharing with us your experiences.
Andrea: Okay. I will show you just some tips to becoming an automated trader and I will show you a step-by-step process on how I face the market. It’s just an example. It’s one out of tens of examples, but I hope it will be useful so let me show you this.
Okay, I hope you see it. That’s me, okay? So I called this Trading System Attempt as this is an attempt from us; but again, just some quick words about me. Earlier on I told you I’d won four times.
I’m the only trader in the world who managed to win four times. I also said that I have a Master of Science in Mechanical Engineering. I don’t know if this helped or not because I’ve always been an engineer when I was a trader. So maybe being an engineer helps, maybe it’s useless. I don’t know.
I retired from my job in 2001. Since then, I am a full-time professional trader. I also wrote a book about money management – Position Sizing. This book is available in Italian and in Chinese. It has been translated into Chinese! And I am a honorary member of SIAT, the Italian Society of Technical Analysis, which is part of IFTA, the International Federation of Technical Analysis; and been invited as a speaker all around the world usually as today, but also physically. Sometimes I have been to most parts of the world to discuss about trading.
So Trading System Attempt. What do I mean with “attempt”? Well, it might be dangerous as what you see here but don’t be afraid. Attempt keeps in its word a number of characteristics you have to go through to get something in trading, in automated trading.
First of all – Analysis. When testing, you have to test, this is the key.
I Test what I do because I want to know if at least in the past it showed to be profitable.
Tuning – what was something that works when you start doing something to improve it even a little bit more?
Evaluation – This is important because when you have something that works and is good, you have to evaluate if it’s really good and also if it is good for you, because if I give you my best strategy for example and it makes money – it doesn’t mean you will make money with it because if for some reason psychologically that kind of approach doesn’t meet your mindset then you will have problems because you will do something that will harm the results of the strategy, for sure.
Then Money Management as I said, which is very important. The proper size to keep risk under control and at the same time to maximize the potential of the strategy.
Portfolio – I always spoke about strategies, which means one single system can help but will not make you rich. You need to face the markets as a rule. You have to try to diversify as much as possible and the portfolio of strategies is the key, is the target you have to have in your mind.
And then Trading – once you have everything put there you have to put it on your computer to let it go and trade; and also in this case you have to make the right choices in terms of software and generally infrastructure.
Let’s see. Let’s start from the first point which is the Analysis. This is very important. Each market needs investigation. You have to investigate on every single market that you want to trade. In terms of liquidity, is this market liquid enough? Are there enough volumes to trade with no problems? In its own characteristics, every single market has a different way to move. You have to understand what main characteristics are there for every single market you are interested in. This is extremely, extremely important. Believe me.
The lecturers who are out there are telling you that a robust strategy has to work on every market and every time frame. I don’t know if I am allowed to say that is bullshit but it is bullshit; because such a strategy does not really exist. This to find out what every market shows as characteristics is extremely, extremely important and you have to be very careful about this. The tide of moves a market shows – what kind of moves are more recurrent in a market? You have to try to find out. In your analysis you have to find out if there is something specific in this area.
Obviously now you know my argument, “Okay. I like this. How do I analyze? What shall I do?” and we know we read sometimes about econometry, econometric tools, GARCH and things like that. Everything is very useful, but when you use those [inaudible – 20:22] let’s say, one size is even very difficult in my opinion. You need to have a good knowledge to use it. But okay, let’s imagine you have this knowledge which is fine, but the point is once that you use these tools to investigate, at the end of your analysis you have information about the markets but not about the trading style. You might know something about the markets but you have to translate this something into an approach.
So actually I don’t suggest to use these econometric tools to do, but I suggest to simply simulate trading strategies. You simulate trading and you see what happens.
One step back. I spoke about it because we see this simulation on strategies. Here I will show you some markets and I will tell you what I look at when I analyze the markets, which is the Italian Future. I share with you here just because I need it here, but it doesn’t really matter. This is a daily chart and I have two indicators here at the bottom of a chart. One indicator is the volume. The one in the middle. Everybody knows about what volumes are.
The second indicator is a custom indicator which shows the dollar move of the market. What does it mean? You could measure volatility in indicating the range of the daily move or an average range or whatever you prefer, but I prefer to translate this into the dollar move. So if my market moves from here to here in one day, from 100 to 110, how much did this actually mean in dollars? How much having one contract of this market would I have seen moving my account due to that move? And I do it here. I multiply the range – or the daily range – by the big point value which is the dollar value of one point move in a market; and I know how much in money this market is moving.
Here I get that information about also how heavy, risky and expensive a market is. If I don’t have a huge capital, I would not be able to trade markets that show excessive dollar roles because those would be too dangerous for me. Here you see we have roughly $2,000 okay? So this is a benchmark for me but you can choose any other market. If this market moves $2,000 I can tell you from my experience that if you try to build an Intraday trend following strategy on a market, your stop loss should be as a starting point to develop at around let’s say 70% of the daily range, okay? In this case, we would be around $1,500 stop loss. So now it’s up to you to say if this $1,500 is too much for your capital or if it is something affordable. If you trade with ten million dollars it’s not a problem, but if you trade with $10,000 that would mean 15% of your account and that’s absolutely too much, of course.
So let’s see. This is the mini-future of the Italian Future and here you will see that the volumes are still good and the dollar size is much, much smaller because it’s a mini contract. That’s Future. You know this probably. Don’t care about the yellow box. It’s something from another analysis which is not important here. The volumes of a DAX Future are about 100,000 contracts a day. So this a very liquid and a good market to trade. Why are the volumes important? — because if there are a lot of volumes, you are relaxed and trading better because you know you will find somebody to sell your contract to or to buy from if you are short. If there are not volumes you might be in trouble if something happens because you will not find any good counterpart to sell or buy to.
And then the dollar move — Look, this is a euro move but it doesn’t really matter. 3,500 here and even more sometimes. This is a very expensive market. So the DAX itself is only 40 for whoever puts capital to trade. On the other side, there is the MINIDAX. The chart is shorter because the contract was born in 2015, but this is one fifth of the big DAX. So this is something that in some cases could be affordable to many traders. So if you like the way DAX moves but you don’t enough capital to trade the DAX, you might consider the MINIDAX, or as an alternative, the CFDs that some other brokers offer on the index.
Mini crude oil – I’m talking about Futures here. Crude oil Future is expensive. Mini crude oil is something that can be considered because it has enough volumes and you see it here and it is 50% of the big crude oil quota. So again if you are looking for some cheap markets, mini crude oil could be a good choice in case you like the energy sector and especially crude oil market.
MINIGOLD – MINIGOLD is not good. Volumes are very small – 366,000. It’s nothing. So MINIGOLD, if you like gold and you want to trade, don’t choose MINIGOLD. MINIGOLD is something you should not imagine to trade because you will have trouble. On the contrary, you could try to trade MICROGOLD because MICROGOLD has good volumes and a very affordable size because it’s one tenth – 10% – of a big gold contract.
So this is something that might be useful for you who want to trade not needing too much money. Now these are just some tips for those who are not heavy capitalized, but I think it’s important to know, because apart from the CFDs or the Forex, who wants to trade the Futures market just because he wants to choose a regulated market, these were some tips to choose from the Futures which are more affordable with small amounts of money.
Now let’s go another step further. We saw something about volumes. I also mentioned about the characteristics of the markets. I can tell you, basically, markets divide in two categories, the category of trending markets and mean reverting markets. What does it mean? Some markets have a behavior where normally once you have a breakout of some level, up or down, then that market tends to continue moving in that direction. It has a tendency to escape in the direction of the breakout. On the contrary, some other markets have a tendency to fade the breakouts. You have a breakout and then a pullback, or to down outside, you have a recovery.
Over long term, the market’s obviously trending; because some markets go up, some markets go down. We have changes which are linked to the economy, to a financial situation; but on a shorter term, I can tell you some markets tend to follow the move generated from a breakout. Some others tend to go back and probably hit your stop loss. How can we decide the characteristics of a market? Obviously, every market is different and there is no unique way to understand what to do, but once thing that could be done that I’ll show you here is to take a daily chart and to use a very simple strategy where you simply buy the breakout of last day’s high or you sell the breakout of last day’s low. No stop loss, no day karate – a simple to buy and reverse strategy.
Here you see some examples. Here we go short at the low of yesterday then we turn long, then short again, long and so on. This is not a strategy that only will make money for you but if you study this, without commissions. Why without commissions? Not because I am tricky but just because I want to understand the behavior of a market. So if a market is trending, has trending characteristics, this strategy should show some profits because if I trade a breakout that trade should try to move in the right direction; the right direction meaning the direction of the breakout in this case. If on the contrary a market is typically mean reverting, this approach, this basic approach, should most probably show losses because every breakout will be a failure. Not every single one, but most of them. So let’s see some results of this.
I don’t know if you can read the numbers but if you can’t okay, buy some new glasses! But okay, here are some Future markets and I have a ranking from the best to the worst, okay? The best market is gasoline, gasoline Future, which shows very nice profits – $377,000 from the year 2000 to some time ago, I don’t know, and the worst market is this one ES0611 is the mini S&P 500. This market loses $100,000.
So obviously you understand that the approach itself does not show common behavior. Some markets on the top of the list make money, interesting money. Some markets at the bottom of the list are a disaster. From this analysis, I understand that gasoline, natural gas, gold, platinum, DAX, crude oil – are all markets with some typically trending characteristics. So if I want to trade trend following strategies, I would probably choose one of these to develop on because I know their identity is stronger in trend following. On the contrary, at the bottom here, copper – ten years low – and especially mini S&P 500 are a disaster in trend following short term. So I would never put my efforts in developing a trend following strategy on a short term base on mini S&P 500 because it would require a lot of work and probably poor results. Why should I put my efforts there when I can easily find some better outcome on those other markets?
So this is to say. This obviously does not tell you exactly, but it gives you a good idea of the behavior of the markets and you see the enormous diversity of some markets. One is making a lot of money and one is heavily losing all the time. So here I get an idea of in which direction I should develop, and by the way, mini S&P 500 is a very efficient market because it has a lot of players and the efficiency of the market is what keeping that from following the breakouts because the level of price is normally the right one. Let’s put it like that, okay?
But now let’s look at this matrix where we have the order in the middle or the center and then we have some points to discuss. Why, when, which and etc. Why means why am I entering the market? This is one of the questions to answer. Is there a reason why I decide to enter the market? And the answer to this question might come from a setup. You might have a specific setup. And bias, what is a bias? A bias is sort of a seasonality of a market. If you discover that a market has a specific tendency to move bullish or bearish or in some periods of the day for example, you might decide to add long or short depending on your formation at the beginning of that period of the day and close at the end. They’re random. You might even enter randomly into the markets and then try to manage your position with some position and risk management. I don’t recommend this because it’s too much theory, but it’s something that you could consider.
The second point is when. When I want to enter, when shall I enter? There might be a best time of the day. If I want to trade a breakout, is it better to do it immediately after the market opens or maybe it’s better to wait for two or three hours to let the moves develop a bit? I would prefer the second. Best days – are there some days where it is better to enter or some day when it’s better not to enter? That is depending on, again, bias. Maybe this day is more bullish, maybe. So it’s better to trade long or days depending on some other characteristic. Maybe you don’t want to enter on a Friday because if I do override I don’t want to keep the position open during the weekend (just an example).
And then chart patterns. Chart patterns are a great tool to decide if a trade has to be taken or not.
Chart patterns tell us something about the recent moves and the behavior that might depend on these recent moves and they give us then information if it’s good or not to enter the market after a specific setup. Order – which order? Select end to a stop, a stop limit, limit market touched – this opens the world. Depending on the type of strategy you might decide about this. Normally classical trend following strategies use stop orders or a stop limit, while mean reverting use limit orders.
A profit is a typical limit order example. You stay there and you wait, okay? Markets touched are now kind of all the world. These are a substitution of a limit order normally, especially on markets like the mini S&P 500. Once you put the limit order in the order book, you have a risk that market trades at your level, but doesn’t serve you, because you have too many players there and out of a thousand contracts 500 are eaten, but you are sitting in the remaining 500. So you stay there and you don’t go anywhere. So if you use a market touch, as soon as the level you want to enter is touched, first of all you send the market by. In this day for sure you enter a market, but you will most probably lose one tick because you entered one tick worth. So it’s up to you to decide.
Market orders – what orders? Obviously you know what they are. Many exceptions – we might consider some exceptions and we might decide whether to close them. For example I never enter a market during the main economical data release. First trial every month, unemployment situation in the States… I don’t want to enter because there is too much volatility and all things like that, okay? This is something I can’t code or something I want to use but not code. For example – Brexit. We knew about the Brexit decision. Why code that which is once in the history? We simply switch everything off the day before. So this is a decision, an exception, we decide to put in place but we don’t need to program it because we exactly know what to do and it’s not repetitive, so we just do it once.
Now after we did this, the analysis, we go into the testing and now we show a real example of what we can do. So would you remember the ranking of the breakout and you remember the mini S&P 500 was the worst of all the markets to trade on the breakout? Perfect! So I can take this to trade and to develop a mean reverting strategy, a counter trend strategy, and I do so. I do so. I imagine to enter the market. Here are some mistakes. Sorry about it. There is any counter trade and the “development” is missing a “T”. I assure you will understand it anyway. It’s still easier than my Italian accent on English!
So let’s say there is a breakout of the highest high or the lowest low of the last five days. Five days is a very important window of time because five days is a week of activity in the markets. So let’s say after five days breakout, I decide to enter, but not in the direction of a breakout because I know that mini S&P is not good for this kind of approach. I enter on the opposite way. So if there is a breakout of the lowest low of the last five days, I’ll enter long. On the contrary, breakout of the highest high – I enter short. I do the opposite of a breakout because I just discovered that mini S&P is not good for breakouts.
These are the rules. The software here is Trade Navigator by Genesis company, but any testing is okay. I use this because it’s very fast to a wide sum and types of rules, but any software is fine of course. This is the result from ’97 to 2012, okay? We already get good results. 64,000 is positive. So we know this is winning, okay? This is winning and it confirms that counter trend is better on mini S&P rather than trend following. Perfect. Okay, but the equity line of a strategy is not really exciting. I mean maybe you like it, I personally would not try this because the drawdown is a bit heavy and it’s not very stable. So okay, never mind. It’s positive, so I can work on it. It’s positive.
So the first thing I say. Okay, let’s take a condition where my breakout is a bit stronger. So I say that the close after the breakout must be at least 10% of the range far away from the lowest, from a level. It’s not just one tick below or above the level of breakout, but at least 10% of a range, okay? So I want a stronger breakout. The results are pretty similar. I got 64,000, before now 63, so it’s the same but the drawdown increased and we see it from the equity line. Here we see it much better. This curve is still not the best in the world, but it is better than what we have seen before. I mean, at least it’s there.
Perfect; no problem. So now I am going to add a stop loss because I always trade with a stop loss; and here I have an optimization of the different levels of stops and the optimization is not run to find the optimum, it is run to understand how things change, changing the values of my stop size. Here we see that no stop – zero means no stop – shows good results. Very tights stops on the very left corner of the chart show losses. So again, the legend of tight stops to win are again bullshit. You can put a beep on this if you edit the video and I say this, but tight stops don’t work.
Then we see some profits and we have some stability zones here and here, zones where changing a little bit of a stop does not lead to excessive changes in the results of a strategy. I will choose the first one with a smaller stop level, but this is absolutely flexible. You can choose a value here on the left or some here. I choose $600, just for the sake of example and there we have here a strategy.
With the stop loss we get some improvement in the results but nothing dramatic, but the equity line is again a little bit better, and this is a safe equity line because it is again protected by stop loss. So here is something we can at least know how much risk we put on the table when we have a little with our protected stop loss. Now our strategy will be ready. We have a setup, an entry and a stop loss. The strategy is ready but I said there is a tuning phase, a fine tuning phase where we want to improve things, to clean up what we have and to see if we can get something which is even better.
First thing is the day of the week. Because I believe in bias, I have attended some Level Winners’ courses. He works a lot on this, and I myself there is something. If we look at the results of a strategy, on single days of the week, we see that at the end of the week actually we don’t get very much. Apparently it’s much better to work on Monday, Tuesday and Wednesday rather than on Thursday and Friday. Here I do it now in the opposite way because I believe there is long and short difference. So I test the long and short entries in the opposite way. I don’t say now this test is not how much do I make on the single days. Here the results are how much do I make avoiding entries on a single day.
Zero means I trade all the days, I don’t avoid any single day. One means I avoid entering on Monday. Two – I avoid entering in this case long on Tuesday. Three – I avoid entering long on Wednesday and so on, and this side, the right side, is the short entries. I skip days. I skip Mondays, Tuesdays and so on. Immediately on the long side we see that skipping Fridays is good. We see the best results, skipping Fridays. On the short side, I would never skip Fridays because you see that it’s hard on my results, but I could consider skipping Thursdays. We remember Thursdays and Fridays were weak in the outcome, so I can decide to skip long entries on Fridays and skip short entries on Thursday.
This is what I can decide. I do it now to see what happens and we get this improvement – 86,000 – it was 68 I think, and this equity which is an even better one we saw before. So now we get a very important improvement in our strategy. Is there something else we can do? Yes. Be careful. I show you all these steps to show you the way I think. I mean, you see what’s leading my development is mostly curiosity. I am curious about the markets.
So at the beginning I was talking about mini S&P and we were talking about breakouts okay, but now the question is, when is a breakout more important?
When there are more players or when there is nobody? Obviously when there are no players. Mini S&P is open 23 hours a day, but there is a day session when the day of world trade opens, okay? So we have a night session and a day session. If we consider only the day session of a mini S&P, so that we cover the session when the stock market is open, I believe that the behavior of that portion of data gives me a much more realistic picture of what the market is doing, rather than taking the whole set of 23 hours; because during the day session I believe that the moves are much more significant, because they reflect what the big players are doing all over the market because also the stock data is trading and everything is going on.
So instead of considering the breakout of a lowest low or a highest high of the last five days of the mini S&P, I take only the day’s session of the mini S&P. I’m still trading 23 hours, but as a setup I take only the day session of the mini S&P. I call it my way for reference. What happens if I do so? I get an improvement – 10% improvement – and again, if you remember previous pictures, a much cleaner equity line. So taking the day session only of the mini S&P to measure the breakout of the last five days gives me an improvement because my setup is more stable, more precise and more affordable to get some good percentage of winners.
We are still in the tuning phase. This should be my last tip. I look at the chart with the trades. This is a chart. Here the losing trades are in red. I know that Chinese people prefer to see red as something positive, but this is a typical Western software so the settings are different, but okay, never mind. What you can see here is that we have some periods where they have more losing trades in a row, you see. So it is sort of against the market but it looks like when there is a momentum or an acceleration in the market I risk to place more losing trades in a row.
So actually I could add the condition that it’s written here but if I am stopped out, if I got a loser, I wait at least three days before entering again into the market, to avoid the array phase. It’s not pleasant. So I wait three days expecting that in those three days the markets will slow down and go back out of that momentum mood which is harming my counter trend strategy. With this addition, I get the same level of profit but I get a much better average trade because I re-use the number of trades but increase the quality of the trades and get a better line. It’s difficult to see from here but if you look carefully at the numbers you will see that there is an improvement.
I was tricky! There is one more piece of fine tuning stuff here. It’s not the last bit. One is – I am in the market now. Shall I stay there forever? No. We see that after a certain number of days that we sit in the market there is no benefit to you waiting forever, and after all we can say ten, twelve bars there is, it’s time to close anyway because we don’t get benefits later on. So we add this condition and we close after twelve days I think and this is our final result.
Now we have to evaluate the strategy. Does the report show good values? Yes, it does. Is there consistency over the years? Yes because the equity line was going up more or less throughout the period we analyzed. Are results replicable in normal conditions? Yes, because we have setup, we know it the day before, the night before, and then we have all the time to place other trades so there is no risk to not be able to do it and also condition. In real-time you can obviously trade it. That is no problem. It’s not something developed out of thirty seconds where you have to place orders every thirty seconds and you could not do it manually or even not with a computer.
And then are these results real or there might be there some software bug. Well in this case there is no problem in considering them real, but believe me, there are many strategies where you get extremely good results which are not real because they are a bug in the software; and a rule of thumb, the better the results look, the more danger of a bug there is. If you have something which looks too good to be true, it probably is not true. And then, last but not least, is this approach in line with my personality? This is very important, because if you don’t like to go against the trend, you will not be able to trade this strategy. You will feel bad. So you have to really sincerely consider if what you have here is something that you can feel is good for you.
Psychological strength is one. So now we are thinking about financial strength. Do I have the financial and psychological strength to trade with? Do I have enough money to go through the drawdown periods? Do I have enough money to trade this with no excessive risk? Do I have the psychological strength to go through a $10,000 drawdown? All these things you have to put on the list and to answer with sincerity. You don’t have to cheat yourself. Now I have also [inaudible – 54:30]. The strategy was developed in 2012 which was here more or less… here somewhere. It kept on making money also afterwards. So it shows stability.
We get now to money management. All we are seeing has been done with one single control just to see how the market behaves. Now we have to put in place. There is no risk on how many contracts you can use for each trade. There are many. I am just am not explaining them here obviously, but you need to consider it and you have to be aware that this can dramatically change the result of your strategy. Just to give you an example – fixed ratio model by Ryan Johnson with a Delta of $500. Most who know the fixed ratio model know what a Delta is. Those who don’t, please go and study.
Here you see the red lines. They are normal lines with one contract. The blue line is the result having applied this model, and you see we would have made 1 million dollars in the same period. Similar to this – fixed ratio with 2.5% risk which is moderately aggressive but with one single strategy it’s feasible. Again, you see a difference. 1 million dollar profit instead of something more than 100,000. Position sizing can dramatically change the results of your strategy.
Porfolio – this is an example. One approach on specific model. You should try to mix sectors. Index, which was index, energy, currency, metals, stocks, bonds. Approach is what’s counter trend. You should put together trend following and counter trend, but with following on another market obviously because this one is not good. Trade duration – this is daily, maximum 12 days. You can put Intraday, short term, whatever.
And then you have to trade. Now the question is okay, but how should I trade? What should I do? These are some platforms, some of the most well-known platforms out there, just as an example. I am not promoting any of these, but just to let you know, Trade Station has its own data with a good history, so you have a good storage on your data set; a great community, and it’s easy to program because the language is an easy language, okay? As Cohen said not all markets are covered. We know American people think that America is the world, so they don’t consider very much about the rest of the world. Some markets are American but cannot do trade and the technical support of Trade Station is very poor. I have to say this. It’s true.
MultiCharts – you can use different data feeds, different brokers as well and it’s very flexible on position management because you can really put many services together. It’s easy to run. The language is mostly the same as Trade Station, but it might be considered expensive if you had to purchase the license, the data feed and maybe something else; then it starts being expensive; and then some banks are not always considered by the support because they are not very flexible in treating the customers. This is what I can tell you.
Trade Navigator is what you have seen in the development. It has a great coverage of US markets. It’s easy to program from a daily time frame up. Inter-market programming – so building conditions to trade the market based on a behavior of another market is very easy to program and it’s nice graphically to see. But it is low on Intraday data, not even very flexible in programming and if you want to program using variables it’s not the most intuitive tool to do so.
Ninja Trader – It is very flexible if you are a good programmer. There’s a very technical support and you can use different data feeds and different brokers, depending on the version, because you know Ninja became a brokerage so they are more restrictive of this now. As a disadvantage, if you are not a programmer it’s not easy to program because this product is in C#. I am not a programmer myself so I have some trouble with this. If you are using Interactive Brokers there is no gateway interface, so you have the problem of daily disconnection. There are work-arounds but there is no official solution, and there is no portfolio tester which is, in my opinion, a con.
AmiBroker – It’s very fast in testing. It’s very versatile. You can really program everything and it’s very good, really good, for portfolio testing; but it requires a hard learning curve. It is not easy to learn it properly and it’s not very flexible in real time. So to trade off a broker you need to adjust things a little bit.
MetaTrader4 or 5, okay – The pros are the data comes from the broker. You don’t need to purchase data. It’s a great community. It’s free, free of charge, which is not bad; and it’s very good and stable, stable software in real-time. It’s a good software, but it’s not easy to program. You need programming skills and it has an approach to constructing which is different from any other platform. So it’s a completely different mindset. So when you switch from this to an another one, you have some trouble.
Cons – Data from broker, but it was a pros, you say you get it here, you get it here. Yes, but what is this? It is an advantage because you don’t have to pay for the data, but it is a disadvantage because normally the data set is limited and it is brokering, because it depends on the broker. Sorry about my coughs. The quality of the data depends on the quality of the broker itself.
How much capital to trade? This depends on the investment you have to make and sometimes it’s not just the money you put in trading but also the money you invest. For example you should probably consider having a cloud or a DPS. This is my newest friend, Matthew. It really depends on your risk profile. Forex is good, but there’s more amount of money lost because of the leverage but also because of it’s scale-ability. Just to get your idea with this, this is a very technical setup of my cloud server. Just to know how this is setup.
Well, this was my insight into the automated trading world. Here you can get at least some links with some more information. If you want there is a PDF where you get some more insights about automated trading. This is free of charge, so you can get it just getting the email out of this website. I hope you will download it because I hope that you’ve got some curiosity in what I do from what I presented now. Obviously, now you will probably ask me some questions, but I hope so. If you have other questions, you can send them all to Philip and he will then revert them to me at his first occasion. Now I leave this and I am back here with my face. Sorry for that!
Philip: Great, great! Well that was really a very fantastic session, especially for me. I have not really been into systematic or algorithmic trading before, but what you have shared actually offers some insights to what could potentially be possible, especially from someone like me who has always been more of a visual trader. So if I could just put in some context; if we just put it your systematic approach as to how you actually trade in the trading competitions that you won. What were the similarities and what were the key factor that allows you or your strategy or your way of trading help you win the trading competition?
Andrea: First of all, in automated trading you can build a plan because you have your strategies. You know how they are supposed to perform and you can base some plans on this because you can plan your path from January to December during the competition. You can plan the mix. In my first competition I mixed four strategies together. I used it on the DAX Future and on the Italian Future to Intraday trend following strategies. I had a counter trend strategy Intraday on the mini S&P 500, one good four hundred and another, trend following on Euro FX Future.
So I had already sort of a diversification and a small portfolio. Small because with $15,000 you cannot diversify enormously, but I was able to put these things together and to have a rough idea of what could happen, so that I could also measure the results during my path and understand if things were going properly or if there was something to change.
Philip: I see. So during the competition itself it was based on real money right… that is being traded?
Philip: And could all those strategies and all those methods that you used be totally replicable to real life in other asset classes, other instruments as well?
Andrea: No. The main approach, yes, but then every single model was tweaked to the market where it was trading. So I had two trend following strategies on DAX Future and the Italian future; very similar, but each had some slight differences to adapt to the market it was trading.
Philip: I see. So correct me if I am wrong. So based on the approach that you actually advocate, we should always start with analyzing the characteristic of a specific instrument?
Philip: And then we come up with a strategy specially for that particular instrument, rather than having —
Andrea: This is what I do.
Philip: And rather than compared to having a certain strategy into algorithm and applying it across different instruments? No, not that way, right?
Andrea: My approach is to understand every single market as much as possible and then to tailor suit for that market.
Philip: I see, and is there any possibility because there is always this saying that when traders use systematic and algorithm trading, there is always this tendency that they were too much curve-fitting, such that it appears so beautiful, but over the long run it is not a robust strategy. Have you encountered that and what was your take on that?
Andrea: Tuning is okay, to clean up from something. Too much is wrong because you do it only to get the best, appealing figure but then in real-time it will get broken because of other fitting. You have to find a good compromise, which is not easy, but there’s no rule to say no more than ten conditions. There is no rule. It’s based on common sense and experience. So obviously, the better the equity line, the more dangerous it is because you would have put probably too many conditions and constraints into it. You can have something that looks good, not extremely good but you know as Perry Capone says, that loose pants fit everybody.
Philip: I see and I would suppose that things like winning streaks and losing streaks will have a huge impact together with risk management and position sizing in terms of how the outcome on your equity curve is going to be, right?
So what’s your take on that and based on the systems that you have developed over the years, are there occasions where you see losing streaks happening for a long time and there’s a risk of ruin if you size too much? Like let’s say 5% or more than 5% every trade?
Andrea: Yes. I have got many strategies which went broken. So the proper position sizing where I limit the risk of each trade is the way to save my account, because even when things go wrong, I am still losing no more than what I can afford. That’s the key. You have to risk and trade small, much smaller than you can imagine, because the losing strike is always behind the corner, you know? Risk or ruin is something that is very far from my imagination because I risk very, very small. I never took a lot of leverage. I always took enough risk.
Philip: Mm. So it’s based on the –
Andrea: And in my account my margins are normally blocked by one third, 30%, when I am heavily invested. So you see I would not even need all the money I have on the account to trade what I trade just because I am very, very conservative. You have to stay humble. When you think you’ve got it to make you rich, it’s that moment when you will start losing.
Philip: I see. So I suppose leverage for you only comes in when you know that you still have room for risk exposure but based on the original amount of capital it’s not enough to open new positions, right?
Andrea: Leverage is not something that comes into my mind. It never comes into my mind.
Philip: I see.
Andrea: I never start from how much do I want to lose maximum if things go wrong, and from that I size my positions, being with how much smaller, it will never occupy a sensible part of my account.
Philip: I see, mmm.
Andrea: I will never to start from how much from the account. You have to say “how much can I afford to lose?” and then based on that, you calculate your position. One, two, three, ten contracts; just because in case of the worst case scenario, I will be losing so, so much. This “so, so much”, is it reasonable for me or not? Yes, it is. Okay, go. Whenever it’s too much, scale down.
Philip: I see and I suppose for most strategies and algorithms, what would be your recommendation in terms of — I’ll not say recommendation, but your preference in terms of the maximum percentage to risk in any trade? Is it 2.5% or it could be —
Andrea: I cannot recommend it because it is something that is depending on the risk profile of each of you. Personally, I measure not the single trade but historically the worst day on a single contract phase and I measure my position limiting that worst day to 1.25% of my account.
Andrea: But this is my approach. It’s my idea. Somebody could refer the worst week or the worst trade and change that 1.25% to 0.5 or to 3%. Obviously, there’s more to the matter.
Philip: I see. So you said earlier on that most of the strategies before you actually implement you do backtesting to at least see how much better it works well or what over historical time.
Philip: So when would be the time that you actually start to question whether your strategy is still working and whether you should do a review, fine tuning or overhaul of the strategy?
Andrea: Wait, wait, wait, wait. There is no single way to do that. You can use some indicators of performance, for example, the drawdown. My biggest times when I had a few strategies, each of them I was running a Monte Carlo simulation to see what the average drawdown could be and the standard deviations. So I decided to stop the strategy if it was exceeding two standardizations of the average historical drawdown. But this is not possible when I have fifty, sixty, strategies. So now today I simply build a ranking of performance of the strategies and from this ranking I get a number of strategies, a load to trade for the coming period; in my case, the coming month and all the others are resting and stopping. So I have an automated let’s say, selection, of profitable strategies based on a number of conditions.
When the strategy shows clearly to have become weak or not working anymore, it’s better in my opinion to throw it away rather than trying to tweak it, because when you take a strategy and you start tweaking the parameters, maybe you get a positive result but that’s normally of a fitting; and then most of the time you get after a while better results on the old version rather than the tweaked one, but I don’t suggest to try to re-animate a strategy that shows fear and weakness.
Philip: I see and how — I mean I’ve heard traders, systematic traders, saying that you have to keep thinking about new ways, new ideas or coming up with new strategies as market changes, as environments change, as technology changes.
Philip: So is that something you do on a regular basis? How do you actually come up with new things that pop into your mind? “Hey, maybe I should try this strategy?”
Andrea: Oh. It’s not really about new ideas. You might try new setups or more likely new approaches. Looking at the markets every day, you might notice something. “Oh, I noticed this”, you go and test “Can I take advantage of this?”, but you have to explore new markets and try to understand more about what’s going on rather than on finding rocket science things because there’s no rocket science out there; but you can say “Okay, I noticed that on this market there might be an opportunity to enter between noon and 1pm”, so test it. Look at these kind of things because you should find an edge somewhere. You can test it and use that if it goes, because it’s not always that things last. The higher the number of traders who discover some inefficiency, the more of an inefficiency it will become.
New ideas mean that you notice something and you test something around what you notice if that thing can make money. Then, you have to verify if your new idea is maybe too much correlated with some other of that which you are trading. In that case, use less to keep it because if you double your exposure on something it’s just doubling the risk.
Philip: And what, in your opinion, would be some of the reasons why some systematic strategies might lose their edge over time?
Andrea: Because of a loss of the edge itself. Strategies base their profits on inefficiencies. The higher the number of traders trading an inefficiency, there’s more of an inefficiency because there is always somebody who tries to anticipate, to go out, and then you lose this potential move. It causes traffic because you have too many people trying to do the same thing. It’s like when you get a discount at the mall. The more people go, the less items you will find left and at the end you don’t find any item, and all because everything was sold out.
Philip: I see. So it is in a way to find a strategy that has got an edge. It has to be contrary from a certain perspective because if everyone is pursuing the same strategy, the same rules, then the edge actually will start to lose itself, right?
Andrea: Lose strength! So it fails to have an opportunity to make a hundred pips. Then players come in. Maybe it’s ninety, eighty, seventy, sixty… and at the end there is no more enough to cover your costs or to cover generally your trading efforts let’s say, and it’s gone. So you have to move somewhere else because there is no more real advantage in doing something where too many people are trying to do that thing.
Philip: I see; and in recent years it seems like things like AI and machine learning has become a very hot topic in terms of using it to actually analyze the market and deriving some kind of strategy on its own, or if let’s say based on a certain strategy you have, machine learning over time might be able to improve the strategy on its own. What’s your take on that? Do you actually try —
Andrea: I’m not a fan of machine learning and not even AI. I don’t like that very much because I think they are mostly over-fitting tools. I strongly believe in the human action behind the tools we are using. I analyze with software, I put numbers together, but again it’s still only me making the final decisions on what to put together. I don’t take things that I don’t understand. If I don’t understand the reason why something could work, I don’t trust trading it, because if I don’t understand the reason I fear it is just a coincidence. Things put together, you get something that works. I mean I don’t trust that. So all the AI tools, all machine learning tools, nobody do this. They find something because they dig very much, but is that something real or is just a coincidence? In my opinion, it’s more in the coincidence sphere and I prefer to keep an eye and power action on top of the development as a human being. In my opinion, we are not yet so far to really delegate the machines’ every kind of choice in trading. That’s my personal opinion of course, but I have not seen really successful AI guys out there so far. I mean, maybe I ignore it. Maybe I am too stupid, but to my knowledge I have not seen any revolution there as of today.
Philip: I see. So right now in your usual daily trading, do you still encounter anxiety despite the fact that you are using pure systematic and algorithm? Do you have any kinds of occasions maybe because Trump did something and you felt like “Oh, maybe my system is going to be at risk of something” and you decide to stop and pause and do something about? Has that occurred to you?
Andrea: No. No… not a lot. I have a temptation sometimes of course if it loses, but I try to stick to a routine. I do that only in case I see that it’s clear behaving wrongly, that it is doing something wrong technically. So then I say “There’s something in the market” and in that case I stop it, but if it just losing because there is a losing strike, no. I keep it alive because I stick to the routine. In the past I noticed that making decisions of this kind normally were producing poor results overall than letting the machine do what it was supposed to do.
Philip: I see. So events that happen overnight and Black Swan events and financial bubbles and crashes and stuff like that, you would just still continue to trade through your system as though it is supposed to run as per usual?
Andrea: Um, a little bit different. If there is a major event where markets are disturbed by the event, I might decide to stop everything; but it’s different because my systems are based on the nature of markets, because I start with a single market. So if the markets are disturbed, so that they behave different, like if you go out and you drink too much, you are drunk. It’s not Philip anymore. It’s a different version of yourself, and maybe discuss it with you, in that occasion it’s not very productive, you know?
So the same case with the markets, with a difference that you would also be losing money maybe. So if I expect that the markets will have a different behavior, then usually because of this major event I might decide to stop everything; as said to Brexit, American relations and other things like that, they might change everything for a while, two or three days, one week. So I think in that case you might decide on certain markets just to put it right. Now this is a specific exception. If there is something like Fukushima or that kind of disaster, you can keep everything running because normally the effect is exhausting quickly.
Andrea: Normally… today.
Philip: I see. I see.
Andrea: Which is not nice I mean because it shows that there is no real sorrow about what’s going on from people, but the reality of the fact is that also when we have terrorists that attack today there is no longer an effect on the markets.
Philip: Yeah, because it’s become so common.
Andrea: We are becoming cynical! So it’s normality, which is very bad from one point of view; but if we look only at the markets, that’s what’s going on.
Philip: I see. So in most circumstances there’s no reason to stop the system from running.
Andrea: Yes, in my opinion.
Philip: Except if you anticipate that. I see. I see. Great. So it was really wonderful for you to share with us so much useful information. So before we end off today’s session, would you have any kind of advice? Maybe three tips you have for any trader who has been trading for a while and is really keen to go into systematic trading? What top three advice and tips would you have for someone of that nature?
Andrea: My general advice for any trader, not only a systematic trader, is to stay humble. Stay humble because you have to be careful when you lose, obviously, but also when you win. When you win very much, stay humble. be careful. You are doing something wrong. If you are winning too much money, you are doing something wrong. You are probably over-exposed. Once in your life, in the right direction, but be careful because it will reverse, sooner or later.
Then obviously if you really want to learn about automated trading, there are resources out there and myself I have Unger Academy. So I’ll promote it just a little bit. You’ll find a lot of resources, but I have also my YouTube channel and Facebook page where there are free resources. You have free videos, just to start getting tips about the day-by-day happenings of a day trader.
Don’t under-evaluate the power of automated trading, but don’t even underestimate it because obviously, depending on the conditions, it does make any bigger problem. Don’t think that just because you are working through an Algo, you become profitable. The machine is smart, but not so smart to put money in your pockets. It’s not a slot machine, or maybe it’s a slot machine because if you play you know that you never win; but it’s not the ultimate miracle and solution to trading.
It’s a way to overcome some psychological issues, to get a proper way to plan your trading style, but it’s not something that makes you your sorrows of tomorrow. It’s something that helps. I chose that because I have a mathematical mind. So if you like it, you can approach it. You don’t need to be absolutely very skilled programmer to start automated trading. There are just some basic knowledge things that you need to know, but you don’t need to be a wizard of computers to do it.
You need first of all passion. Please, don’t trade for money. Trade for passion, because if you trade for passion, you will do it happily and you will overcome the bad periods. If you do it for money, there may be other ways to make money out there which will get you less stress. Money is not easy to make, but trading is one of the most difficult ways to make it, in my opinion.
Philip: Haha, wonderful! That was really good advice and I think it really brings a lot of traders back down to earth in terms of the kind of expectations.
Andrea: Yes, expectations are always too high.
Philip: Yes, yes and I think that is exactly what we are trying to achieve in this Online Trading Summit. Not to let traders feel that it’s an easy thing to just participate in some strategies, holy grail and you will make a lot of money from that but really teach them what is the difficult thing about it but it can be done if you really set your mind to it, actually. Thank you again so much Andrea for your time today. I am very sure your teaching is going to be very helpful to the Summit participants and I wish you all the best in whatever that you do and hopefully I will have a chance to interview and to chat with you and do some kind of presentation with you again in the future.
Andrea: Thank you. Thank you for inviting and thanks to everybody who has been listening to this. I hope you enjoyed it and stay humble. Be careful.