|
PureBytes Links
Trading Reference Links
|
<FONT face=Arial
size=2>there are softwares that can do multi-precision calculations, They all do
this at the price of speed.
<FONT face=Arial
size=2>
<FONT face=Arial
size=2>Why does Eugene think that he needs to handle these large sized
numbers?
<FONT face=Arial
size=2>
<FONT face=Arial
size=2>I'm reminded of an anecdote about Teller and the early stages of
designing a n H- bomb. The story goes something like this: the AEC had
built a super dooper whooper computer (that is for its time). Then as now they
had to develop an operating system and a high level language for it, and then
run trial calculations to see if it would work.
<FONT face=Arial
size=2>
<FONT face=Arial
size=2>While this labor, called spinning the wheels was going on, one of the
physicists ran an analysis of his own. If he used any mechanical aids they might
have been a slide rule or a desk calculator. He made some assumptions and
approximations. When he had completed his work he brought it in to Teller
and told him that the H-bomb, as they were planning to build it, wouldn't work.
Teller called the carrier of bad news a traitor, incompetent, and a communist
agent. Six months later when they were able to run the calculations through the
computer, Teller had to eat very humble pie. The computer analysis confirmed
that the projected design wouldn't work. The lesson is that a computer can't
overcome the problem of bad judgment.
<FONT face=Arial
size=2>
<FONT face=Arial
size=2>To me we shouldn't use intricate methods in our trading, there is
too much random noise in the data.
<FONT face=Arial
size=2>Look at Darvas' record. In a very short time in the 1950's he
cleaned up $2 million using a fairly simple method (it wont work
nowadays).
<FONT face=Arial
size=2>
<FONT face=Arial
size=2>When optimizing a process, we need to know something about the process to
be optimized. Its a waste of time to run what appear to be random numbers in the
hopes that one of them will come out great. When optimizing a trading system,
one set of parameters may come out really great. Always this means that there is
peak or valley in the chart and one set of parameters found
it.
<FONT face=Arial
size=2>
<FONT face=Arial
size=2>Keep it simple but not simple minded.
<FONT face=Arial
size=2>
<FONT face=Arial
size=2>Lionel
From: Jose [mailto:josesilva22@xxxxxxxxx]
Sent: Tuesday, April 27, 2004 6:46 PMTo:
Metastockusers@xxxxxxxxxxxxxxxSubject: [Metastockusers] Re:
TrendMedium RT core hot update.
Eugene, my calculator can handle values up to 10^100 - it doesn't
mean anything though.Anyway, as soon as I can get my hands on a PC
that can process 160^20 values (or similar) within a year's processing time,
I'll try out your Official MetaStock Add-on. ;)jose '-)<A
href="">http://users.bigpond.com/prominex/pegasus.htmUnofficial
MetaStock add-ons--- In Metastockusers@xxxxxxxxxxxxxxx, "Eugene
Labunsky" <elabunsky@xx..> wrote:> Dear Jose,> >
I'm using genetic for looking optimal params.> More than - core in the
work have 160^20 values or similar... ;-) the main> goal of my work
is development way for find optimal trading system (I mean> formula
kind).> > Best Regards,> Eugene Labunsky.> <A
href="">http://www.TrendMedium.com - Official
MetaStock Add-on
Yahoo! Groups Sponsor
ADVERTISEMENT
Yahoo! Groups Links
To visit your group on the web, go to:http://groups.yahoo.com/group/Metastockusers/
To unsubscribe from this group, send an email to:Metastockusers-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
|