|
PureBytes Links
Trading Reference Links
|
Hello Cameron,
The process is no diferent than eod except more complicated usually. I
can't help with you question bec I don't know your system. I sugest
you get a few intraday data systems--one's that test out or not--and
study the coding. Soon you will understand some of the differences.
Here's one I've never used and don't remember where it came from but
it is somewhat complicated and illustrated many techniques for
intraday systems. Gook Luck.
{System JB 30 NL BO - the Jake Bernstein 30 minute new high / new low
INTRADAY breakout system
as described by Jake Bernstein at the October 21, 1997
This system is a simple new high / new low of the first iTime bar
breakout system as described by Jake Bernstein. Jake expressed that only 3 out of a 1,000 people
would ever trade this system because of three minor problems with this system:
a) Our tendency to want to take profits (even small ones) and run,
b) The system requires the ability to risk a lot of capital to make
big profits,
c) The system does not use a stop loss - hence the only way to exit
the trade is to reverse positions or to exit at the end of the day.
Rules for the system:
a) Calculate for the first 30 minutes (or in our case, iTime minutes)
the new high and new low for the day. No trades are taken the first iTime minutes.
b) At the end of the second (and additional iTime frame bars), check
to see if the close of that time period closed above the 1st 30 minute bar high plus 2
ticks or below the 1st 30 minute low minus 2 ticks. If not, even if the high of that bar
was above the 1st 30 minute bar's high plus 2 ticks / low of the bar was below the 1st 30
minute bar minus 2 ticks -- NO TRADE. This system works only on where the positon of
the close is in relationship to the 1st 30 minute bar's high / low plus / minus 2
ticks.
c) If the close is above the 1st 30 minute high plus two ticks, place
a order to buy at the close.
If the close is below the 1st 30 minute low minus two ticks, place a
order to sell at the close.
d) No new trades in the last hour of trading.
e) Exit the position at the end of the day.
f) Exit the position when a new indication is given to reverse
positions. That is, if you are long one position, on a new short position, exit and go short. Same
for reversing from short to a long position.
h) This system does not use any other exit routines - no trailing
losses, no maximum losses, no money management losses, etc.
(Note: This system was coded for intraday stock trading an instead of
the 2 ticks, Randy used 1/4 of a point as a substitue. You should change this if you want
to use it in the futures markets. Likewise, the system was coded to buy/sell 1000
shares of stock via the parameter sShares. Change this default to the number of contract
you are interested in trading in the futures markets. This system as built will not trade
the first day of test intraday data but uses it to set up correctly. It uses the firstord
varable to by pass this first trading day. The number of bars used can be set very low because there
is no requirement for a large number of lookback bars.)
This is a day trading indicator which when matched up with the JB 30
NLBO indicator would suggest that large profits and a large percentage correct trades can
be accomplished by a mechnical trading system. This system will not trade the first
iTime minutes of the day and should not make any new entries into positions during the last
hour of trading. This last hour is accomplished by the variable e1time which is twice the
amount of the input variable of iTime (which can be modified for testing purposes).
Jake mentioned that this works best for the 30 minute time frame based
upon computer backtesting. This system works best on futures --- the S&P, Bonds,
Currencies but does not work very well on anything else. The system should only have an
average of one trade per day, but could have up to 3 or 4 in a very moving index or
currency. You can contact Jake at http:\\www.trade-futures.com\ for additional
information or training.
This system is built to be placed upon any 1 to 60 minute scale chart
to correctly plot test the system. While this could be used on a tick chart, it
was not built with that in mind.
inputs: iTime(30);
var: cH(close), cL(close), s1time(0), vtime(0), e1time(0),
sShares(1000), Firstord(false);
if s1time = 0 then begin
s1time = MinutesToTime(TimeToMinutes(Sess1StartTime)+iTime);
vtime = MinutesToTime(TimeToMinutes(Sess1StartTime) + (2*iTime));
e1time = MinutesToTime(TimeToMinutes(Sess1EndTime) - (2*iTime));
end;
if date > date[1] then begin
Firstord = true;
cH = high;
cL = low;
vtime = MinutesToTime(TimeToMinutes(Sess1StartTime) + (2*iTime));
end
else
if time <= s1time then begin
if ch < high then
ch = high;
if cl > low then
cl = low;
end;
if time >= vtime and
firstord and
time < e1time then begin
vtime = MinutesToTime(TimeToMinutes(vtime) + iTime);
if close >= cH + .25 and
close > close[1] then begin {enhancement - test to see if trend is
up}
buy ("JBBO Buy") sShares Shares at Market;
end;
if close <= cL - .25 and
close < close[1] then begin {enhancement - test to see if trend is
down}
sell ("JBBO Sell") sShares Shares at Market;
end;
end;
if MarketPosition <> 0 and
time >= 1555 then begin
ExitLong ("Exit 1") at Market;
ExitShort ("Exit 2") at Market;
end;}
Best regards,
Jim Johnson
Tuesday, June 05, 2001, 9:33:16 PM, you wrote:
CJ> Hi All
CJ> So i have this sytem , and it uses a tight stop in the S&P. And it does
CJ> really good with tick bounce at 10% , But it is average at 15% and pathetic
CJ> at 20% bounce. So i now have 5 min S&P data thanks to don the ledgend. So
CJ> now how do i run my EOD system using the five minute data to get an acurate
CJ> backtest?
CJ> Help Please
CJ> Regards
CJ> Cameron
--
mailto:jejohn@xxxxxxxxx
|