washington, dc
Posts: 3 since Jan 2013
Thanks Given: 4
Thanks Received: 1
|
Have not really found a good answer to this problem:
Diversification is fine and helps risk-adjusted returns. However, DIVERSIFICATION FAILS WHEN YOU NEED IT MOST. Correlations can run 0.5 to 0.7 in normal conditions, but as soon as markets start running, those same correlations approach 1.0 and even the most diversified portfolios, if on the wrong side of markets, can quickly fail, exceeding the expected losses of even well-backtested systems.
The best theoretical approach I've seen is Ralph Vince's work, which basically tries to backtest based on correlated maximum drawdown to arrive at optimal trade size. Theoretically, this limited trade size will not suddenly transform into a de facto overleveraged position when the time comes.
But this is more theory than practical.
How many markets should be traded, and, maybe give some examples of combinations.
Thanks in advance!!
|