Optimising mechanical trading strategies

[JTrader]

pkfryer -
I think we would be better off concentrating on systems that do one thing and do that well. And stick to using them in particular environments including logic in them that closes them down when they detect that the market is incompatible with their rules.

At best I'd have a system that doesn't lose money when the market changes and makes efficient trades when it goes in its favour (for automatic systems) for manually traded systems... Use your judgement and use the right tool for the job.

Thanks pkfryer

this quote more closely resembles the sentiments that I wish to incorporate into my intraday (<1 hour time frame) systems.

Cheers

jtrader.

 

[mr_cassandra]

re only 2 years data

Its my opinion that you may need much more than 2 years backtest on any major index to gain the 'all markets' experience any system must be tested against. Here in the US there was a multi year bear market from 2000 down to 2003. Ditto there was a multi year bubble runup when we had dot com mania. The data base I have goes back to July 1995 and as such has been tested theu both the bubble run, bear and flat markets.

I've seen some folks argue for 12 and even 20 years data but imo, at some point you've got to take your work from the theory table to the reality table.

Regards, Steve

Speaking only for myself, I've found what happened to my program in the real world with real money to be 3-4 times more valuable in improving my system.

QUOTE=jtrader]This quote addresses the principles that I want to stick to. Use a large amount of historical intraday data - say 2 years per instrument - so that a curve fitted system becomes a less likely outcome. This way the system will have experienced bull, bear, ranging and flat markets and will hopefully have been profitable or had minimal drawdown within each market type - and profitable overall.


Thanks again

jtrader.[/QUOTE]

 

[mr_cassandra]

Actually mine does work in all types of markets.

It goes without saying that it does not call every turn perfectly, nor can anything else. However, that being said, its calls enough of them enough of the time, to dramatically outpace buy and hold or even the best mutual funds.

Beyond the backtest to 1995, it has been performing just the same in realtime trading this year.

Ragards, Steve

ps: this is a multi-week-month system so my experience may not apply to short term trading time frames.

I think its asking a lot for a system to be profitable in all market states using fixed system components/variables. Ranging markets dont suit trend based systems and vice a versa. Also, the rally correction intervals shift.

I found that the systems I worked on worked better for rallies, resulting in a set of systems that actually only took buy signals. Buy only systems seemed to work best. As an experiment I worked on a sell only system and that profited very well but try reversing the rules to produce buy signals as well and it only broke even. Trying to make a system do well in all market environments is like trying to produce a vehicle that can fly, drive on the roads and sail on the ocean. Possible in James Bond movies but will produce poor results in markets.

I think we would be better off concentrating on systems that do one thing and do that well. And stick to using them in particular environments including logic in them that closes them down when they detect that the market is incompatible with their rules.

At best I'd have a system that doesn't lose money when the market changes and makes efficient trades when it goes in its favour (for automatic systems) for manually traded systems... Use your judgement and use the right tool for the job.

 

[Trader333]

PKF,

In my view it is not surprising that using the reverse rules of a successful system for going Long does not work for going Short and vice versa. I think that there are a couple of reasons for this:

1) Even accounting for bear times, markets ultimately have always eventually gone up

2) Markets generally drop much faster than they go up so the rules that work for one condition are highly unlikely to work in reverse.


jtrader,

Trading intra-day is an entirely different proposition to that of position trading. I have previously tried adapting successful position trading strategies to intra-day scenarios without success. If you succeed where I was unable to then I will be interested to know how you have achieved it.


Paul

 

[pkfryer]

Trader333 thats the conclusions I came to as well. But point 1 can't account for it as the trades I was using were short-term: days to weeks basis, not long enough to suffer through any lengthy draw down before a bull market re-initiates itself. Also, the rules had reasonably close stop losses in place.

Its also interesting that the market in general would change so much that a system fails that was once very successful. Its also a warning to all system traders that you should never 100% trust any back testing no matter how realistic you made it. A system that works wonders in testing can fail just at the point that you start to use it. If I had hopped on board a few of the systems I tested 15 years ago I would have been very wealthy right now. If I had hopped on 5 years ago I would have had half my account lost. The ironic thing about it, it wasn't very impressive between 20-15 years ago so wouldn't have been chosen to trade live. The markets change all the time!

 

[JTrader]

Why may adapting good position trading strategies to intra-day strategies not work?

mr_cassandra
re only 2 years data.............Its my opinion that you may need much more than 2 years backtest on any major index to gain the 'all markets' experience any system must be tested against.

Hi mr_cassandra

I am talking about intraday data backtested between time intervals of 5-60 minutes. I can see that more than 2 years would be necessary for an EOD system, but 2 years sounds to be plenty for these smaller timeframes. Do you agree?

Trader333 -

jtrader,

Trading intra-day is an entirely different proposition to that of position trading. I have previously tried adapting successful position trading strategies to intra-day scenarios without success. If you succeed where I was unable to then I will be interested to know how you have achieved it.


Paul

Hi Paul

I plan to operate my systems within the style of intraday momentum trading and not scalping. Also, I do not intend to fully automate my systems (initially) - but just have them create mechanical rule based entry and exit signals - that I will execute manually.

I do not really understand why such a strategy may not work as well on intraday data and timeframes, as it would do on an EOD basis............My expectation (as a novice to back-testing &MTS') would be that the principles of trading and technical analysis - whether looking at EOD or intraday timeframes - remain the same. The differences in timeframe and other parameters must surely be superficial.

With regard to intraday trading, on another thread that I started, I seem to remember that you highlighted the difference between the chart alert price and a difficulty in getting filled at the same price or better with your broker - as a problem that you faced when you initially started intraday mechanical trading.................I would be very greatful if you could highlight any other obstacles/problems that are likely to arise.

Thanks again everybody

jtrader.

 

[Trader333]

jtrader,

What do you mean, or understand, by momentum trading and how do you intend to trade using this approach ?


Paul

 

[mr_cassandra]

AGree on intraday

I forget sometimes that I'm looking at a very different time frame than traders per se.

Hi mr_cassandra

I am talking about intraday data backtested between time intervals of 5-60 minutes. I can see that more than 2 years would be necessary for an EOD system, but 2 years sounds to be plenty for these smaller timeframes. Do you agree?



Hi Paul

I plan to operate my systems within the style of intraday momentum trading and not scalping. Also, I do not intend to fully automate my systems (initially) - but just have them create mechanical rule based entry and exit signals - that I will execute manually.

I do not really understand why such a strategy may not work as well on intraday data and timeframes, as it would do on an EOD basis............My expectation (as a novice to back-testing &MTS') would be that the principles of trading and technical analysis - whether looking at EOD or intraday timeframes - remain the same. The differences in timeframe and other parameters must surely be superficial.

With regard to intraday trading, on another thread that I started, I seem to remember that you highlighted the difference between the chart alert price and a difficulty in getting filled at the same price or better with your broker - as a problem that you faced when you initially started intraday mechanical trading.................I would be very greatful if you could highlight any other obstacles/problems that are likely to arise.

Thanks again everybody

jtrader.

 

[JTrader]

Why may adapting good position trading strategies to intra-day strategies not work?

jtrader,

What do you mean, or understand, by momentum trading and how do you intend to trade using this approach ?


Paul

Hi Paul

by momentum trading I basically mean identifying, trading and hopefully capturing profits, from when the instrument is/is likely to be moving in a pattern/trend during the course of the day. On a real-time time frame of probably 5-30 minute time interval charts. Not holding open positions overnight.

My understanding of this is different to scalping, whereby trading involves the rapid and repeated buying and selling of a large volume of stocks within seconds or minutes. The objective is to earn a small per instrument profit on each transaction while minimizing the risk. I imagine that scalping is done on tick charts, or 1-minute to a maximum of 5-minute time interval charts.

Therefore with intraday momentum trading, the stops, attained profits and length of time within the trade are likely to be longer than with scalping.

Thanks again

jtrader.