|
PureBytes Links
Trading Reference Links
|
Unfortunately, I am not as smart as I sound in this matter. What I suggested in my earlier post is much easier said than done. The markets by nature are not a deterministic system and do not really exhibit stable distributions. My only defense has been simplicity and diversification. And I've tested my stuff on lots of history. I don't have any easy quick answers. What I can say with certainty it is critical to at least recognize the link between system stability and the appropriateness of optimal f type strategies. Vince calls this f shift.
Truth is, the Holy Grail would be *knowing* that your system edge would be stable in the future, regardless of how big the edge was. Making money would then be reduced to a casino operation. It really would be a money machine.
Optimal f gets a bad rap undeservedly IMO. All of the arguements that I've heard against it violate base assumptions that Vince himself would be the first to emphasize.
1) Huge drawdowns. A base assumption of optimal f is that the investor's utility is for maximum equity growth with no concern for drawdowns. Simply put, if your utility is something other than for max equity growth, optimal f is not the right tool for you. We would be foolish to ridicule the large backhoe as useless simply because we are just trying to dig a small hole and a shovel is more appropriate. It may be useless for our job at hand. But try excavating 1000 cubic yards of earth with a shovel.
2) My system blows up. This has to do with my assertion about system stability. If your system blows up using optimal f, it was not optimal f that did the damage, rather the distribution of returns used to calculate optimal f was not in fact the distribution of returns for your system going forward in real time. IOW, the system was not stable.
Scott Hoffman
Issaquah, WA
-----Original Message-----
From: Lincoln Fiske [SMTP:lincolnf@xxxxxxxxxxxxx]
Sent: Wednesday, September 09, 1998 3:52 PM
To: Scott Hoffman
Cc: 'omega-list@xxxxxxxxxx'
Subject: Re: optimal f and futures lots traded
Scott--
What's the best way to measure "the stability of the distribution of returns from the system"? Is there a minimum threshhold of this value over which optimal f works? Can optimal f be implemented in Easy Language? I've heard so many mixed reviews of optimal f that I'd written off looking into it, but guess I should attempt to understand it.
Thanks,
Lincoln
|