Example Trading System

I feel to excuse me for my omnipresence here, and i will be more in background later, but now i feel the importance to do a statement about a psychological issue i meant to read between the lines...

Random reinforcement is the word...Most to everybody get trapped by this...

If a strategy does not work for some days to weeks and you observe that another behavior would have give you the money and you switch to this, you are in big trouble...

It`s very probable to get paid with big bucks for making big mistakes for some period of time, so you think, Yeah, that`s the way it works now.

That`s random reinforcement, you get paid to react to random or timely limited marketbehavior, but that means nothing in the big picture, if you switch to the new working rules/mistakes, you lose money. Market is a lot of random movement, sometimes you get paid for doing the right things, sometimes you get whipped...

So what about trading trendfollowing, you have to stand 66% losers and up to 10 -14 loser in serie, but trading a basket of commodities did work fine for decades, but is extremly hard to stand psychologicaly, 90% are not able to trade such a system.

That`s the same with this ORBO stuff, if you get trapped with sharp reversals in a serie, that means nothing, may be just a random joke of Mr. Market.

Michael
 
@bounty
Are you european? Cause the timestamp on your chart...
What i see is that all fakes occured in the lunchhour(or?), i don`t take Breakouts in the Lunchhour, there is to low Volume to be meaningful, in more volatil times this worked better, but in the nowadays calm market lunchhourbreaks are very often fakes.

My datafeed and broker is Interactive Brokers, but i have to collect by myself and there is no historical backfill if i`m to late with login. So my data are just 99.9% correct, i have some rare holes, so i would not talk about some $ tolerance...

An now i lost your question....moment
 
Ok, If i didn`t make any mistake and the unfiltered Lunchbreaker results are as i mentioned, this is not a question about netprofit, but about if this strat. is realistic to trade at all. I did include 10,-$ Comm, but you can easily experience more, the last trade i got 2 points slippage at the entry, if you are not tricky trying to exit at some limitorders from 1555 or so, and you exit at market this late in the day you will get additional slippage...

With 8,-$/trade a Strategy is not realistic to trade. To have a realistic chance to win the game, i think 30-35 $/trade are minimum, 18 $ are critical, trading some size is not possibel with such low avgtrade, 2-3 points additional slippage is nothing in this chaotic arena.

I have to leave now, but i will check the results again and i `ll post it later...

greetings, Michael
 
Quote from mechtrader41:

I feel to excuse me for my omnipresence here, and i will be more in background later, but now i feel the importance to do a statement about a psychological issue i meant to read between the lines...

Random reinforcement is the word...Most to everybody get trapped by this...

If a strategy does not work for some days to weeks and you observe that another behavior would have give you the money and you switch to this, you are in big trouble...

It`s very probable to get paid with big bucks for making big mistakes for some period of time, so you think, Yeah, that`s the way it works now.

That`s random reinforcement, you get paid to react to random or timely limited marketbehavior, but that means nothing in the big picture, if you switch to the new working rules/mistakes, you lose money. Market is a lot of random movement, sometimes you get paid for doing the right things, sometimes you get whipped...

So what about trading trendfollowing, you have to stand 66% losers and up to 10 -14 loser in serie, but trading a basket of commodities did work fine for decades, but is extremly hard to stand psychologicaly, 90% are not able to trade such a system.

That`s the same with this ORBO stuff, if you get trapped with sharp reversals in a serie, that means nothing, may be just a random joke of Mr. Market.

Michael
More...

Michael, I'm aware of what you wrote. Since it seems as if you replied to my triple loss experience, I need to make clear that those losses alone didn't trigger my desire to change the strategy.

What made me think of a change was that the historical drawdown recently got exceeded significantly. While my simple old ORB strategy (YM range from 09:30 - 11:45 EST without any filters, exit at 16:00, stop at other side of range, additional static stop of 50 points, static target of 150 point) never had two losing months in a row before, whilst monthly losers where much smaller than winners, this changed since december.

Monthly point results where
-44 (December)
+5 (January, but since I missed the first trade of the year with +80 due to vacation, this month was a loss for me, too)
-135 (February, biggest monthly loss ever so far)
and for March -174 yet (looks like it's going to be an even bigger loss than the fresh negative record from february)

These numbers accumulated throughout several months, compared to years of previous data, make me feel a change is necessary.

P.S.: My time zone is CET (EST+6), I guess we live in the same country :-)
 
Hi bounty,
I´m a kraut and live in Germany, Bayern...
Ok, I`m a somewhat dillettant in easylanguage, and I like simple solutions, by the way, what version of TS do you use?

And again i would not debate over cents or even 1-2 $ cause some rare dataholes, but that does not change the big picture.

Here the code i used, please check the rules.

Commissions = 10$/Roundturn

{DATA1 = 5 Min YM}
Input:StartL(1900), Fakt(0.3), Trigger(1);

Vars:Lunchrange(0), Lunchhigh(0), Lunchlow(0);

if time = 1900 then
Lunchhigh = highest(high,15);
Lunchlow = lowest(low,15);
Lunchrange = Lunchhigh - Lunchlow;

if EntriesToday(date[0]) < 1 then begin
if time > startL
then buy at Lunchhigh + Trigger stop;

if time > startL
then sell at Lunchlow - Trigger stop;
end;

exitlong at Lunchlow - Lunchrange * Fakt stop;
exitshort at Lunchhigh + Lunchrange * Fakt stop;

if time > 2200 then exitlong {at close limit};
if time > 2200 then exitshort {at close limit};
setexitonclose;

The results are in the pic, remarkable that nearly the whole profit was made with shorttrades !

Michael
 

Attachments

Well, next the afternoonbreaker...

Commissions = 10 $/roundturn

Openrange = 210 Min, first 3 1/2 hours

Stop = 0.6 * OR

Exit after 2200 (1600)

results are in the pic, the shorttrades did perform ~50% better as the longs with lower drawdown as the longtrades
 

Attachments

To repeat some words about the market, the reason why some strategies don`t work anymore is the very low Vola, the S&P trades with 0.7-0.8 % daily average tradingrange at the moment, in Mid of 2004 the averagerange was ~1.4%. In 2003 we had times with 3-4% average daily range. 2002 with 1.5-2.5% or so, same in 2001...

We are at low Volatility not seen since mid 1990, so Breakoutsystems profiting from large trenddays have hard times, i further suspect less trenddays at all and more z-days or calm creeper days. Unfiltered Breakouts on the longside are not profitabel at the time, better are shorts, cause the law of the faster falling prices, on the shortside is mostly more volatility.

Cause Volatility is cyclic this will change one day and we see again high vola, but no one can tell when this will happen, in the mid 1990 there was a very calm time for 3 years, so a change can happen in some weeks/months or in 1-2 years.

Cause the low vola, we have to be more selectiv in taking trades, it`s a good advice to look for the days/times with higher probability of success.

pic with daily average range in %

Michael
 

Attachments

@bounty
did look on your set of rules, fixed $ stops and targets make no sense in trading, i would always prefer some ATR based stops and targets adapting to marketvola.

The stop exactly at the other side of OR is fools play, everybody`s stops are there, the results are better with widther or tighter stops. Fixed $ stop i found 40 pts best and as target a 100 pt target.
Yes, 3 negative months, but i would say that`s not unusual for such a very simple system, i would start to worry after 6 months...

I definitely would not trade on mondays and would cancel all orders in Lunchtime>>>you get more or less an afternoonbreaker, the ATR(5)filter is very logical und useful, after a widerange bar you have a very high probability of a range/Z-day>>> losing money...

Pimp my system in the pic

Michael
 

Attachments

Ever thought about the positionsizing stuff, means how much risk in % from your account per trade? Uuuh not...

Bad...

The size is the key and makes the difference between losing with a winning system or getting millionaire soon...

With the same system....

Here the basic and the "pimped" system i checked for bounty with some Moneymanagement rules..
 

Attachments

@bounty
from own interest i did some tests about dayofweek with the Lunchbreaker, on the left side you see the basic version, on the right side the one with the enhancements i described, also with a Lunchhourbreaker mondays are the worst, the most stable conditions are for any reasons on thursdays, thursdays are wellknown for afternoon reversals, may be that`s the point.


It`s of importance better not get caught in Range/Z-days, for the Lunchbreaker or any other nonscalping Strategy looking for some trend, the results for the LB are even better if you not only don`t trade if yesterday was a Widerange day, but the day before to, none of the last 2 days should have been a widerange day. The resulting equitycurve is at the bottom of the pic.

Meanwhile this Lunchhour thing looks promising for me :-)

And more interesting, it seems much more important to look for conditions excluding rangedays and look for higher probability for a Higher Vola or Trendday than the method of entry itself. I was quite surprised that i could slide/move this timewindow of 75 Min across the time of day and it doesn`t matter to much if i took the 75 min one hour earlier or later, a break of such a timewindow was profitabel shifted up even to 0145 -0300 pm, but the best results did the Lunchhour. So the point seems more to look when NOT to trade, than the exact method of trading, here we meet the subject "The ZEN of trading" :-)

Michael
 

Attachments

You must log in or register to reply here.
Back
Top