|
PureBytes Links
Trading Reference Links
|
Scott Hoffman wrote:
> 1. Too many degrees of freedom consumed in the system. IOW, to many parameters.
This could be narrowed down to not knowing or understanding what it takes to build a robust system.
I would propose that a system have three components in order to hold up over time. One: would be a
short term counter trending method. Two: would be a long term trend following method. Three: would
be a combination adaptive stop/profit taking method.
> 2. Favoring edge over robustness in system design.
Yes, you put it kindly, I will call it greed! Only losers will fail to understand it.
> 3. Not enough data to thoroughly backtest a method.
Precisely! I marvel (and snicker) at systems results I see posted on the web showing two weeks of
results. It is a pity that they sell Super Bikes to first time motorcyclist!
> 4. Not deducting enough for realistic slippage.
This is valid but the least of our concerns for now. This is also a VERY important issue. This one
single factor should tell you if you have built a system like 99.99 percent of the so called
technicians out there or not. If your system suffers from slippage and it becomes a factor then you
have built the system wrong. You must make slippage work for you not against you. This could be
narrowed down to the simple title of inverse thinking.
> 5. Not being able to distinquish between true system failure versus temporary drawdown.
This could be thrown in with number 3.
> The list could go on...so many ways to get it wrong, so few ways to get it right.
You have certainly given me enough obstacles to explain already. Thanks
> Scott Hoffman
--
TC
|