|
PureBytes Links
Trading Reference Links
|
"If my new system can't beat the performance of the "Buy and Hold" system
over a long period of time, why bother?"
A number of reasons. The myth of buy and hold in testing is that it starts
off in a buy position from first date in data. Your indicators have to come
up to speed. Lets say you dropped in a 200 MA. you don't get first reading
on it for nearly a year. Buy and hold either profits or loses during that
time frame. In this Giant bull, that year can make a difference.
Drawdowns are ignored when you look at bottom line. B/H might be ahead now
but you could have set through a 50% drop.
Time is money. While a 5 year buy and hold may show a larger profit, your
system could have had you in the stock a total of two years in that time.
Your money could have been active elsewhere.
Then think of the unfair aspects of saying to B/H, you have to buy on first
day of my data, even "investors" have some reason for entry. The TA system
enters and exits on an objective measure but not B/H in the test.
Yes, I love to say "It beat B/H", but in reality The amount of risk and the
time/money aspects allow you to beat it without the bottom line. And since
we aren't back at the first day of data trying make a decision. I fall back
to the objective measurements our indicators provide. I am not ready to pick
a stock and say come hell or high water, I will buy and hold.
If someone had bought IBM at its 1987 high, it would have taken 10 years to
get even. It would take someone that doesn't care about their money to have
bought and held for that long.
One aspect Equis might change is not start count till there is a system buy.
Another more realistic change would be to stop compound trade amounts and
let us say buy/sell 1000 or 100 shares on each trade, instead of putting all
profits back in each trade.
Richard Estes
-----Original Message-----
From: Chip Anderson <chipamy@xxxxxxxxx>
To: Alberto Torchio <atorchio@xxxxxxxxx>; metastock-list
<metastock-list@xxxxxxxxxxxxx>
Date: Friday, February 06, 1998 1:12 PM
Subject: Re: Buy and Hold
>Hi Alberto,
>
>Simply use the System Tester dialog to create a New system and enter
>an expression that is always true in the "Enter Long" box. I use "C >
>0" which is usually true ;-)
>
>I always make sure to compare the results of this test with any other
>system I am developing. If my new system can't beat the performance
>of the "Buy and Hold" system over a long period of time, why bother?
>
>Hope this helps,
>Chip
>http://coolhistory.com/ChipsCharts
>
>---Alberto Torchio <atorchio@xxxxxxxxx> wrote:
>>
>> Dear Listmembers,
>>
>> I am sure it has already been on the mailing list, so please allow
>me to
>> apologize.
>> The question: how does one write a "trading system" to obtain the
>same
>> performance as buy and hold strategy?
>> Thank you all.
>>
>>
>> Alberto Torchio
>> Torino, Italy
>>
>>
>>
>
>_________________________________________________________
>DO YOU YAHOO!?
>Get your free @yahoo.com address at http://mail.yahoo.com
>
|