Ten important things I learned in ten years of trading

Although I started trading more than 10 years ago – almost 12 to be exact – my blog and public efforts in trading started in the fall of 2007, almost ten years ago. I was 21 at the time and I viewed trading in a completely different way although my core believes – which is what has kept me trading all this time – have remained basically the same. Today I want to share with you a post about some of the most important things I have learned through the past 10 years of trading, small pieces of wisdom that I hope will help you reach your trading goals faster and make smaller mistakes in the process. Although I still have a lot to learn and a lot to understand I believe these ten years have taught me some key lessons worth sharing.

It’s all about statistics. When I started trading I was in the middle of my major in chemistry. Although I knew how important statistics and math were in science I did not fully appreciate their importance in trading. It has become more and more apparent through time that if you need to pick one area to learn about to become a good trader, it should be statistics. Someone without a strong core knowledge of statistics is most probably going to fail in trading, basically because they can be easily fooled by randomness – have a good month by chance and believe they have it all figured out – and are prone to pick strategies that are mathematically proven to fail (think grids, martingales, etc), just because of the way in which they skew short term rewards in exchange for future catastrophic failure. Statistics are the sharpest tool in the shed, use it.

Automated trading – any trading – is emotional. One of the reasons why I decided to focus so much on automated trading was because I knew that it was very hard for me to trade in a rational manner when I had to make every trading decision by hand. However it quickly became obvious that automated trading is not emotion-free. Algorithms need to be traded and decisions need to be made about whether to run them or not. You need to decide if you run system A or system B and if either of them performs poorly you need to decide whether that’s normal and the system needs to be allowed to recover or whether it needs to be removed. There is no such thing as emotionless trading, because there is no such thing as decision-less trading.

Everything eventually fails, everything. When you begin trading and find something that works pretty consistently you feel like you have the entire problem solved. Through the first half of the last 10 years I developed around 4 systems that performed with amazing consistency for 3-5 years and then, they all failed, almost at the same time. Having systems that have been performing so well suddenly start to get into drawdowns far more prominent than anything within your historical tests – even your Monte Carlo simulations – is a pretty discouraging experience. You feel like the roof is crumbling over your head, with no way to hide. Since everything fails, you need to make sure no piece of your trading arsenal is vital to your trading. Everything fails, so everything needs to have a replacement.

Trading should be boring, it’s not a casino. If you feel a rush of excitement when the market moves in your favor or sick to your stomach when you lose, then you’re definitely risking too much. In the beginning I used to trade with risk levels I would never touch now – haven’t we all? – but I realized that this meant that my decisions about what to trade – which systems to use, etc – became clouded by the emotions that were attached to watching them trade. Watching an algorithm trade shouldn’t generate any sort of strong emotion, if this happens, lower your risk, you’re gambling. 

The past is the past. A key insight about trading is that historical profitability is necessary but not sufficient. Even if you manage to eliminate all cognitive and statistical biases when designing a trading system – which you cannot do – a system can still fail under future market conditions. This is basically because there is no rule saying that the future will be equal to the past in all specific ways in which it can. If there are say 100 million potential algorithms that worked in the past that would not be expected to have worked just due to chance or mining bias the market might evolve in a way where 70 million of those are either unprofitable or behave much worse. Trading profitably requires a measurement of these future expectations, there must be some knowledge of how likely things are to develop in a certain way. The past is limited in possibilities while the future is limitless.

Failure to innovate is failure. We’re no longer in the 1970’s trading Donchian channel breakouts, this is a reality. Algorithmic trading systems that are simple tend to have deteriorating long term edges compared to more complicated strategies that work, simply because they are more accessible to people who are searching for inefficiencies. This is why the drawdowns for things like long term trend following systems tend to become worse with time, because they are traded to the point where trading them becomes painful for those willing to operate them. More complexity requires more knowledge for the evaluation of things like statistical biases so there is a much higher entry barrier for more complex trading systems. Complexity causes higher failure when done incorrectly, due to increases in bias sources and things like chances of programming bugs, however well done higher complexity  – algorithmic complexity, execution complexity, etc – will tend to have deeper edges, just because of the accessibility problem (or do you think Renaissance trades the turtle system?).

Lacking insights is having a disadvantage. Think about the trading system or systems you are using right now. Could you answer any question about them? Can you answer what the probability of winning or losing is? What is the probability of winning or losing given this week’s market conditions? Last week’s? Can you pin-point what a market condition is and how to define that mathematically? A lack of ability to answer a question means that you’re at a disadvantage against someone who has that answer. Trading – particularly in negative sum game markets like Forex – is a cut-throat, shark tank, last-man-standing competition. Answers are the gold currency here, each answer you lack means a bit less of an edge for you, a bit more for your smarter competitor.

Success is not a website, video, system or book away. It is really tempting to believe that success in trading is easy, but there’s a ton of Dunning-Kruger effect here. Traders who are new and have read a few books or gone through a few seminars/webinars or youtube videos – especially those with a few months of profits under their belts – tend to underestimate the difficulty in trading. I remember when I was 21, I thought that I was a trading wizard after only a few months within the game. It’s perfectly normal to believe that something that you’re doing apparently successfully is “easy” but time eventually shows – as the market cycles and traders are crushed – that trading is a very difficult game indeed. No website – including my blog and trading community – video or book, will ever be enough to make you profitable on their own.

It’s something you don’t do fundamentally for the money. Through the years the people that I have met who are successful traders – either discretionary or algorithmic – sincerely don’t do this fundamentally for the money. Trading just for the money is a huge trap because you won’t have the necessary focus and driving force necessary to develop the understanding needed to succeed but you will be obsessed with things like avoiding all losing trades, over-trading, reaching a certain % target per day/week/month etc. Certainly the money is important but it isn’t the underlying reason why I pursue trading nor why all good traders I know do so. If you don’t love math, difficult challenges and innovation – by the way, it’s not a crime if you don’t – then quantitative trading is probably not a good choice for you.

Programming is vital, learn to do it. Besides building a strong base in statistics, the next corner-stone for any trader should be coding. Programming is not only good to learn if you want to automate algorithms but it is also vital for running experiments to answer questions and apply all your knowledge in statistics. As I mentioned before, the ability to answer questions marks a fundamental difference in someone’s ability to find and profitably trade edges compared to someone who can’t. Programming is a tool to answer these questions and lacking programming knowledge puts a person at an enormous disadvantage against someone with that knowledge. While it took a discretionary trader 2 years to develop a trading system, it might have taken a quant 5 seconds to develop 1000 of those algorithms using data-mining tools. From systematically evaluating sources of bias to properly quantifying and describing market conditions, coding is key. Learning to program will only be a positive thing for you – even if trading doesn’t work for you as it’s a very generic skill – do a few online courses and start coding.

I hope you enjoyed this list. Although I realize some of the above might be controversial for some people, especially new traders, these are some things I have learned through this journey that I thought many of you would like to know about. If you would like to learn more about automated trading and how we approach some of the problems mentioned above please consider joining Asirikuy.com, a website – not a holy grail – filled with educational videos, trading systems, development and a sound, honest and transparent approach towards automated trading.strategies

Print Friendly, PDF & Email
You can leave a response, or trackback from your own site.

Leave a Reply


Fatal error: Call to undefined function mcsp_html() in /home1/danielfp/public_html/mechanicalforex/wp-content/themes/modern-style/comments.php on line 71