Thursday, January 28, 2010
When Risks Exceed Rewards: The Costs of Overtrading
In my recent post, I took a look at some of the dynamics of one's trading edge. That post assumed a relatively modest degree of risk taking to achieve a 10% annual return over the course of 100 trades.
In this post, we'll once again turn to Henry Carstens' P&L Forecaster and see what happens when we increase the amount of risk needed to achieve the same level of return. Instead of a standard deviation of daily returns of 1%, we'll now assume a standard deviation of 3% to earn that 10% annual return.
That means that our $100,000 account holder is swinging plus or minus $3000 on returns for approximately 2/3 of all trading days. While that might not seem like a lot of money, it's a large swing for the sought return of $10,000 over the course of the trading year. Personally, I don't know of any reputable hedge funds that seek such levels of volatility among their portfolio managers.
I ran Henry's Forecaster 20 times and here were the projected P/L results after 100 trades:
$6676, $15,226, $18,378, $8932, $6633, $15,503, $14,586, $15,733, $12,259, $6174, $19,608, $11,057, $5794, $18,365, $9778, $12,992, $13,378, $7473, $10,976, 7850.
Now compare this with the annual returns when risk was at a 1% standard deviation:
$9379, $12,097, $12,861, $9210, $10,934, $11,529, $9779, $7992, $10,694, $11,827, $10,839, $11,535, $10,300, $11,738, $9324, $9052, $13,197, $11,712, $10,736, $9683.
What you can see is that, even with the higher volatility, the presence of the trading edge is apparent. The "Rule of 100" still holds: After 100 trades, with any reasonable degree of risk-taking, you know if a trading edge is present.
Notice, however, how the size of the edge relative to the random swings in P/L is diminished when risk is increased while returns remain the same. Whether the trader makes more than 15% on the year or less than 8% is completely due to chance. Risk levels magnify the chance component of returns. This happens most often when traders "overtrade": they take more risk (by putting on more trades, trading larger size, or both) without achieving a positive edge for the added risk exposure.
I ran the simulation one more time (see P/L chart above) to show what that means in practice. Note how a trader with a positive edge but high level of risk (i.e., someone who is overtrading their edge) goes through massive P/L swings during the trading year. In the above scenario, the trader is losing money for half the year, with double digit negative percentage returns, only to sharply swing higher later in the year, steeply drop, and then rebound.
Realistically, could a trader stick to his or her edge while drawing down in this fashion?
More than likely, such a trader would abandon a winning strategy during an extended, deep drawdown and/or double down on the strategy during a big winning period. Both would create fresh problems.
The moral of the story is that, as Henry says, the path of one's P/L contains as much information as the endpoint. It is not absolute returns alone, but also the risk taken to achieve those returns, that matter in the long run. If your daily swings in P/L are much larger than your average daily profits, it will be difficult to stick with your strategy and edge.
The wise trader seeks positive risk-adjusted returns, not just large returns on capital.