#007: BUILDING A HEDGE FUND #3: IMM (INTELLIGENT MONEY MANAGER)
At the end of the year 2013, long before I decided to create a hedge fund, I made a list of tasks for the year 2014. I’ve decided to improve and make my trading more effective during next year. You shouldn’t forget I’m just a trader like most of you - a decade ago I was just an inexperienced trader with a 2.000USD trading account - and the only reason why I’m still trading is because I constantly innovate and improve my trading. Similar lists of tasks are therefore an essential part of my routine.
It isn’t so relevant what else was on the list (some of the tasks were solved, others weren’t). The important thing is that the following point was included:
Find a method to recognize a system which is starting to fail and automatically decrease its contracts and importance in the portfolio. Eventually, switch it off completely.
This looks like an easy point to sort out, but it turned into one of the most difficult challenges you can possibly imagine.
I began to do online research to see if someone had already solved this issue before me (and also for me). Pretty soon I found out that if they had, they wouldn’t have been willing to share that information. And I absolutely understand that (I have to point out that I wouldn’t share this information with you either - if you expect me to share with you all my secrets and know-how in the next parts of this series, keep on dreaming). I hadn’t even found any useful ideas or clues online. All I could do is to start exploring that area by myself.
I started with my favourite method of getting new ideas - meditation. I’m a big promoter of meditation: I train the creative part of the brain and solve abstract problems. Many many years ago, I participated in a Transcendental meditation training and since then, I’ve been meditating regularly. Every time I need to solve a problem, I start to meditate, create abstract images and make something useful out of it (new solutions always come from the abstract, as that’s the beginning of the new concreteness). After a couple of days, I spent meditating, a few ideas popped into my head, but their implementation wasn’t easy at all. Almost immediately, I started to face technology problems. The classic trading software, which you probably know and use, was inadequate for my needs as it couldn’t create the algorithms I needed the software to create and test.
Hence, I asked my friend, elite programer for help. I’ve worked with him many times and he’s always been discreet and loyal to me. I asked him to join this hedge fund project as one of the key persons for a profit share. I described him my idea about a special testing platform designed for searching and tuning optimal algorithms for automatic portfolio management based on the likelihood that the system is about to collapse. My friend accepted my offer without hesitation - after all, we have worked together on a couple of projects before and we both had an inner urge to do something big together.
To get the basic algorithms wasn’t so difficult. At first, I created a couple of sketches in Excel tables for him to understand what my idea was about. But the problem of Excel was the speed - even when we created a couple of specialized testing macros, one iteration of a test lasted minutes and minutes, and we still needed to test thousands of these iterations. This is why we very quickly decided to build our own software solution. We created a testing Framework in C# called IMM (Intelligent Money Manager).
It was a pretty simple application in which we encoded all my ideas for better and more intelligent system management (same as we originally created in Excel ) and afterwards, we inserted the equity curve of my systems - especially those ones that had failed in the past, to see how the algorithms could handle similar situations, and if they could predict them.
We had a great feeling about doing this - it was like being Indiana Jones exploring and searching for something new, something essential. However, we also had a bad feeling because most of the ideas didn’t work or at best were half-working. Some of them turned out to be interesting, but they were too general and had too many variations, different settings and switches; and at that time, I wasn’t able to tell which was the right one. I had a feeling that there was a potential, that all this chaos would bring something interesting and useful, but I had no idea what it would be, how to identify it or even how to extract it and implement it on my portfolio (it happens quite often to me). It felt like being on a date with a gorgeous woman, having a great time, but not being able to make a move.
At that stage, we decided to stop the project and postpone it (it was in September 2013, I think). I didn’t know how to carry on; I had more questions than answers. I needed to stop thinking about it for a while and have some time for myself. So, we stopped it, and after some time, I completely forgot about it. Based on my meditations, I understand that everything takes time and our anxiety was the reason why we got stuck and didn’t continue with the work. It would have been a waste of time anyway.
Luckily, that time off, which I used for solving another million of trading related issues and ideas, lasted only a couple of months.
At the beginning of 2014, on a roof of Marina Bay Sands hotel, I got the missing piece. Out of the blue, I saw the piece which we hadn’t been able to figure out to successfully finish the IMM project, and what was even more important - I suddenly saw the big picture: the mistake I did in 2013. I was too focused on one problem and didn’t take into consideration other parts, which also have an enormous effect on the strategies’ results. Markets are like living organisms and they need to be viewed as complex, from the top and from many different angles.
However, everything was different now. When you look at the whole Singapore city as a unit, from the amazing pool, you have a brainwave and you can see in your mind the whole image of your trading idea for intelligent portfolio management. I was very lucky to get that epiphany. Suddenly, I figured out which of the tested IMM approaches was the right one and what was missing. The only thing I had to do was to take a look at that issue from above. When I think about other links that we need to connect to our algorithm (data exchange within the portfolio, with other algorithms, connection to different market phases), most of the options get eliminated by not supporting all these connections, except for one - “The one”. I have to say that at that time of epiphany, I didn’t fully perceive that kind of things, and the idea was on an intuitive level rather than on a rational one.
We went back to the cruise ship to continue on our round the world cruise and in the same evening, I started to make notes for creating our brand new application for portfolio management (we still called it IMM for some time, but we changed the name to TRADING DIRECTOR). My big idea of our hedge fund started to get a clear and realistic image. And this is the essence:
Create a brand new way of portfolio management, position sizing and risk management using new, highly innovative auto adaptive algorithms, which will be able to communicate with each other and see the big picture at every moment observing all systems and all markets.
back to previous part continue to next part
| part 1 | part 2 | part 3 | part 4 |