#010: BUILDING A HEDGE FUND #4: THE BIG PICTURE
If you want to aim for big results, you need to start seeing The Big Picture. Otherwise, you’ll never evolve from 2000USD, which was my initial trading capital, into a big multi-million hedge fund. I understood this in Singapore that afternoon and since then, I’ve followed this rule even more than before.
What does “seeing The Big Picture” mean?
I’ll try to explain it in a simple way: imagine trading as a battle (as a matter of fact, it really is a battle) and the market as a battlefield, where the strategy is your army. The strategy has to enter and exit the positions in the market for profits and to win the battle. Do you agree?
REALLY?
Well, then you have a big problem. This is not The Big Picture I’m talking about. The Big Picture needs a much more complex view.
First of all, the market isn’t a battlefield. It’s a mega-war. It’s a cluttering, nuclear disaster. And there are many other armies that even Darth Vader would be at least concerned about. If you enter the turmoil of this battle, you’ll be crushed into small pieces in a second.
At first, you need to have many armies, not just one; you need a quantum. That means a lot of strategies, a fully diversified portfolio fighting on different battlefronts and surviving at least on a couple of them.
But, for something like that, you need an important part - a general staff. You need a brain that manages all those fighting armies. Someone who decides which armies should retreat and which should keep on fighting. Someone who sees which armies should be reinforced and which can be reduced. And most importantly, there must be someone who, at every moment, re-evaluates the whole situation comprehensively and from all different angles, carries out new orders, adapts to the current situation, changes the tactics and reorganises your armies and their sources in real time. You need to perceive the whole battlefield, be flexible, react immediately and understand that losing one army doesn’t change anything - you need to focus on The Big Picture.
I grasped this concept in Singapore when I started to understand that all the effort I had put into the IMM application was just two dimensional, too binary. It isn’t only about when to switch the strategy off and back on. It’s also about when to decrease or increase the contracts, which systems to link together into one big “mega-strategy”, when to do so, how to promptly change the capitalization based on the battlefield situation, how to find the hidden spots on the battlefield which can be used to attack and be one step ahead of your enemy, but most important - how to monitor all armies altogether from above and redistribute available sources in real time (i.e. having all strategies in mind, not only one). You need to create an intelligent control unit and establish communication channels for all armies to have all the necessary information to make decisions. These decisions should be made based on the current results of every single army, but also on their current morale. In other words, you need to create many more algorithms exceeding the IMM and every algorithm will have a slightly different purpose in the management process. But all of them together will make one unit, the general staff for all armies.
We were still in Singapore and I had already seen lots of areas which could be managed by algorithms, but I had limited them (to begin with) to these:
1. The main algorithm evaluating in real time the risk of a whole portfolio and also the worst possible scenarios (we called it RPMC Algo).
2. The algorithm for evaluation of the current behaviour of all strategies in a portfolio and adequate decrease or increase of contracts bearing a whole portfolio in mind (part of RPMC Algo).
3. The algorithm to switch off systems, that show “unhealthy behaviour” and switch them back on automatically after certain conditions are met (part of RPMC Algo).
4. The algorithm for the adaptive quality evaluation of each buy/sell signal and adequate change of the risk exposition according to success rate probability (we called it “Signal Quality” - SQ but changed it afterwards to BIN Algo)
5. The algorithm for the adaptive evaluation of equity curve changes and adequate changes of resources assignment (we called it STEP Algo)
6. Strategy grouping according to certain criteria (we called it GROUPING) - e.g. according to market quality criteria (we called it Market Quality)
Everything described above might sound too difficult and mysterious, but don’t be afraid. As I wrote before, we decided to explore the unexplored to find out that it was impossible to manage a whole portfolio in a simple way. This is why I started to meditate again and think in abstract, which led to the construction of all I stated above. I’ve attached an overview picture for you, loyal readers of my blog, to get an idea of what I’m talking about:
Let’s start with the fact that we have a PORTFOLIO. Set a couple of different strategies designed for different markets and different timeframes. All these strategies are initiating buy/sell signals in a different order. None of these signals is executed immediately, every signal has to be checked by a series of algorithms within the Trading Director (TD) application to get a number of contracts or sometimes even no contracts at all (signal rejection).
Every enter signal has lots of important information that TD has to process. It processes information about the market, which the signal is meant for (processed by BIN Algo), about the quality and character of the signal, underlying the market and also the character of the current situation of a whole stock exchange. But at the same time, TD receives information about the recent performance of the system that created the signal (sent by STEP ALGO) and most importantly, it gets overall information about all possible risk scenarios of the strategy, and also of the whole portfolio which is evaluated by RPMC Algo.
Trading Director processes all this information really quickly (all is optimised for speed and most of the calculations are done in advance before any strategy initiates an entry signal) and makes a final decision about the system exposition. In other words, it decides how many contracts the strategy should buy/sell (we don’t use any standard position-sizing models; the number of contracts depends on a complex information evaluation of all algorithms). It can also skip the current signal or turn the system off for some time (or turn it back on). Of course, the system is much more complicated than that and the final decision depends on many more factors like GROUP HEAT, PORTFOLIO HEAT, etc. I think that the simplified description above should be more than enough for basic understanding.
As a result of this process, we can earn more money with less risk, extremely manage the risk (just the part that can be managed) and earn money with much more confidence. Don’t get very stressed by the performance of individual strategies because what really matters is The Big Picture. It’s as if you had a big company that works perfectly because of the communication channels and internal rules and processes - you don’t have to deal with one employee underperforming, or even with one who resigns or retires. This won’t put your company at risk, and it can keep going on without any problems.
I have spent the rest of our stay in Singapore, and the following weeks of our cruise, thinking about that and I have created an image to see how our Trading Director application, the most innovating point of our future hedge fund, should look like (this application IS NOT AND WILL NEVER BE for sale - it’s been exclusively developed for our fund). It was great to finally see “The Big Picture”, have the structure of algorithms and their mutual connection solved and finally start working on it. We had no idea that the realization would be incredibly tricky, 1 year behind our schedule and a real challenge for our brains like nothing has ever been before. The process of creation has turned into a true nightmare.
Happy Trading!
Tomas
