As a trader, having a good portfolio is very important especially when you want to make trading your career. Therefore, the goal of a portfolio model cannot be understated. This is because the model acts as the arbitrator among three components: alpha model (the optimist), the pessimist (the risk model), and the cost-conscious (transactional cost model) before making the decision about how to proceed. The model is important because if you focus on one of the above, chances are that you will make a tremendous mistake. For instance, if you focus on the alpha, you will tend to be greedy and ultimately lose all your holdings. On the other hand, if you are pessimistic, you will underperform and if you are a cost-conscious person, you will tend to hold positions indefinitely.
These models are particularly important to people who focus on quantitative trading. To build the models, you can either chose to use the rule-based model or the optimized model. In the former model, the rules are based on heuristics defined by the quant trader. Since these depend on the trader, they can be simple models. On the other hand, the optimized model uses optimizers which use algorithms to get to a goal defined by the trader.
In rule based portfolio construction model, there are four main ways to do this. These are: equal position weighting, equal risk weighting, alpha-driven weighting, and decision-tree weighting. In the equal positioning weighting strategy, the users believe that if a position looks good to buy, then there is no other external information is necessary to determine the size. In equal risk weighting, the model will adjust the position size inversely to their volatilities. As a result, more volatile positions are given smaller allocations while less volatile positions are given bigger positions. For instance, in a portfolio that has Facebook and Exxon Mobil, Facebook which is more volatile will be given a bigger position than Exxon which is less volatile. The alpha-driven weighting model on the other hand determines the positions by the alpha model where the model dictates how attractive a position is likely to be.
For anyone thinking of becoming a quant trader, the portfolio optimizer topic is a very important one. The optimizers are based on the principles of Modern Portfolio Theory. In short, this theory tends to explain that investors are inherently risk averse. This means that if two assets offer the same return, but at different levels of risk, they will buy the less risky asset. This led to the concepts of risk-adjusted return and mean variance optimization. In the optimizer, mean and averages will be the key inputs. Other inputs will be: the size of the portfolio in currency terms, the desired risk, expected return, expected volatility, expected correlation, and any other constraint. By combining all these single inputs, one will be at a good position to optimize his trading.
For new traders, these concepts might seem complicated. I know. In fact, these models are recommended to the sophisticated trader, particularly one who wants to try building his own algorithms to execute trades on his behalf. As I have written before, none of these models is perfect. A good example of this is a hedge fund which nearly led to the collapse of the financial markets in the United States. Long Term Capital Management was a highly sophisticated hedge fund with partners who included 3 Nobel winning economists. The fund also had top-notch PhD physicists to develop the systems. For three years, the fund returned more than 30% of net returns. In the fourth year, the fund collapsed. Cases of failed quantitative funds are rife to this date. Therefore, if all you need to do is to make money, there are other strategies which I have explained in the past including: scalping, hedging, and swing trading among others. You can use these instead.