Automated trading:does it work?

[Guest zugli]

Hi dear friends!

I'm new here and I like this place.

I just open a new account with TS brokerage and I would like to know about your experience with automated trading. It works? Do you use it?

I optimized some trading systems that I would like to run on 240, 480 and 720 min.

Another question: do you know if it's possible to backtest a ts only in a particular session? and then run it in automated mode in the same session?

Thank you in advance for your help

[mvhanson 10]

Hi Zugli,

 

This only answers one part of your question, but I don't use TS so can't answer the rest.

 

I can tell you that automated (system) trading does work; though I think you need to set some goals about what it is you want to accomplish with your systems before you dive in.

 

For example, I used to work for a hedge fund and the guys that ran it had a good system, but it worked so well for so long they got greedy. They started to leverage up.

 

Then one day the systems blew up. Boom. They couldn't believe it. They kept trading and trading and trading the systems, thinking the systems would recover. They never did. They lost all of their money and their business.

 

My view would be to parcel out the portfolio you have allocated for system trading into several competing systems, across different markets -- one in commodities, one in stocks, etc. Use some of your own, and maybe buy one or two (Aberration, etc.) that you know have good track records.

 

If you get egotistical about something you have created yourself and put all of your cake into it, somewhere down the line, the market is going to eat your cake for lunch and leave none of it for you.

 

Hope that's helpful.

 

 

 

 

Hi dear friends!

I'm new here and I like this place.

I just open a new account with TS brokerage and I would like to know about your experience with automated trading. It works? Do you use it?

I optimized some trading systems that I would like to run on 240, 480 and 720 min.

Another question: do you know if it's possible to backtest a ts only in a particular session? and then run it in automated mode in the same session?

Thank you in advance for your help

[Guest zugli]

Thank you for your ideas, I will take care of your opinions and suggestion.

[GCB 10]

First off I think TS is an awesome piece of software. I use it for both charts and orders and have no feeling that I need anything else. As a former professional programmer I especially like its robust programming language for writing custom indicators. I would be lost without that.

 

That said, I think the automated trading feature on TS is more of a selling gimmick than anything else. I've written several workable though unsophisticated trading strategies and they all had roughly the same results: barely profitable with about 2-1 losers to winners and hundreds and hundreds of trades.

 

The problem with any automated system is translating a trading system into a programming language. It is very hard, without miles and miles of programming code, to produce anything that even comes remotely close to duplicating what a good trader sees and thinks when sizing up even simple entries and exits.

 

Automated systems within the reach of the average trader and programmer are simply not going to perform that well. It would take months of programming to produce anything worthwhile, during which time you could have been becoming a better trader, rather than a better programmer.

[walterw 10]

now GCB thats a good answer you see... the time spent developing a trading skill vs the time spent on programing... very good argument, and the fact that there are many factors that systems cant be trained to see for sofisticated the algorithm could be ... good post. cheers Walter.

[Knyyt 10]

Does Automated trading work? Of course. It is just a function of what you would do in a real trade, if you can find a way to mathmatically describe your method, and if you have a positive expectancy method you could automate it. There are several retail level automated programs currently available,TradeBolt,Ninjatrader, and several others that are able to work with tradestation.

 

Regards, Knyyt

[lrushing 10]

Automated Trading With a Purpose

 

Having a clear set of rules and criteria that are executed on a consistent basis is a good tool to have as a compliment to discretionary trading.

 

Quite often, traders develop automated strategies with the expectation that the strategy alone is all that is needed to accommodate and take advantage of emotional and psychological behaviors of everyday traders. This leads to false hope and disappointment of many traders.

 

Experienced traders rely more on learning to master fewer techniques and focus more on efficiency. This is where automated trading is at its best. Automated strategies are more effective at evaluated 100’s or 1,000 of charts and price action on stocks, futures, etc…for set-ups, market conditions, and pattern failures. These are the opportunities traders rely on for their next trade. So to say that automated strategies don’t work may be an overstatement of their intentions and how the developer has set expectations of them.

 

The simplest strategies usually are the better ones long term. No strategy or technique works well all the time, hence it is important to understand how to filter out those occurrences when the strategy weakens. Trading strategies that are risk tolerance based with reasonable expectations usually do well and have a positive expectancy. That’s all a good trader could expect from a consistent set of rules and criteria, right!

[torero 10]

I have several simple automated strategies that have been doing well so far. One over a year and the other 2 over 6 months. I know these can and will fail at any time so I have cut off points where my Uncle Point is.

 

This is after 2 years of being on TS, learning the code and nuances of the code in describing the strategy into code and the detailed intricacies, factor in as many real life simulations as possible (slippage commission, limit order fill problems etc) but they are simple in the end, just like a discretionary trader should be with this strategy. If the system is too rigid (codes on end with too many filters), it won't survive different market conditions. So the less rigid it is the better chances with surviving through time, but it may take stupid trades but hey, have to take the good with the bad.

 

This is one thing i've learned: the simpler the strategy, the longer the shelf life but the less profits they are. The other is high profits but it can fail at anytime. So some system traders choose one while others choose the other, personal preferences and also understanding the rules of system trading. It's a trade-off.

 

The biggest problem is knowing when the system has failed and your cut off point is. This is the last anyone learns but it has to be taught first I believe. This is the first plan when trading it for real.

 

Finding a profitable system takes time just like learning to discretionary trading and it gets better and quicker in finding new ones, just like discretionary trading.

 

Good luck.

[Maverick 10]

I don't use TS but I do use Amibroker's Auto Trading Interfaced to IB and here is why I like Automated trading.

 

More fun. It is mesmerizing and great fun to see your orders being placed, modified and filled faster than any human trader could ever do and do so without lifting a finger.

 

Less stress. Not having to punch in all those orders makes trading a fun activity.

 

More efficient. Whether you trade Intraday or EOD manually placing complex orders can be time-consuming. Using the IBc you can automate all order placements and perform them perfectly in a fraction of the time it would take you to do it manually.

 

More flexibility. You can make up your own order types that react instantly to price changes in real time.

 

Less emotional. We all know that emotional trading can kill even the best mechanical system. Your Automated system will follow your trading rules flawlessly and never ever second-guess signals.

 

More responsive. If a price changes your prices can be recalculated, orders placed, and perhaps even executed, before you can even move your fingers to the keyboard.

 

More accurate. No entry errors, ever.

 

Trading Niche. While popularity of automated trading is rising fast it is still mostly practiced by a small percentage of elite and knowledgeable traders. I believe that in High-Frequency Automated-Trading there will be less competition than in other trading areas. The price excursions and volumes are just too small for fund traders but are a perfect niche for the small trader.

 

More profitable. If you are trading a profitable mechanical system now adding automation to it will almost certainly increase your profits.

[Kojak 10]

Automated Trading may not be possible for all trading strategies, but trading strategies can be modified to fit into the automatic trading. You can seize the advantages that an automated (or even partially automated system offers ).

 

Like you can

 

a) Operate in smaller time frames (any time frame actually) up to High Frequency

b) Operate on a larger basket of instruments

c) Have more complex strategies when the basic rule of any manual discretionary trading system would be to Keep it Simple ( and "Stupid"? "Successful" the traditional manual traders would insist. The KISS syndrome.)

d) Operate beyond the limits of human endurance

 

Like it or not Algos are taking over the world. Best to transform and become a trading cyborg early, before they run over your trading systems. People who have been trading for a longish period know how trading has changed with the introduction of automated trading systems. Like whiplashes that trips up your stop losses. Reversals at targets etc which happens at a speed that a manual trader cannot respond to. Unless of course your trading system is immune to such movements.

 

The real issue is that with automated trading, there are opportunities that exist now that you could never dream of seizing.

 

On a lighter vein see this Ted video: We are all cyborgs now and on a less lighter vein..this one How algorithms are taking over the world

 

Jose

Chennai

Edited August 3, 2011 by Kojak

[zapisy 10]

Yes, it does work, but like others said; it is long time to create one, then test it.

 

One advice. Do not optimize systems in the hard way, just to find good equity curve - it is called curve fitting, and it never works the next day.

 

The best it to find strategy/create one, that does not need any optimization, and adopts itself to the market. If you have to optimize, then always take at lease few months from the chart (latest few month), then optimize, and then apply the new data to see if it works.

[sadscorpion 10]

Hi

auto trading work in my point of view but if you make some rules and these rules are more important as your strategy rules. If you are using any commercial ea then must know its strategy that is programmed in it.

dont use any ea in all market instruments or market conditions(like 24/5) give them some rest .thanks

[mslk 10]

Of course it works ... check out this cool contest. News - Automated Trading Championship 2008

 

-mslk

[n_bagadia 10]

hi guys ...

i am from mumbai .. .and i am trying to automate my strategies on amibroker to i.b. ... is there a way i can use the scaletrader algorithm in my stratedy and can i send orders in a sliced way ... and their respective profit order also should go in the same slice size ... if any one can share the code of the same .. or share some idea. ... and some references for coding automation on amibroker ...

there is lot of difficulty i am facing to find a reasonable platform ... i.b . is good but the brokerage they r charging is very high in options ... can u guys suggest some good and reasonble platform ... i am also a subbroker ... so direct access to market would also be possible ...

we need to also spread awareness here in india ... as traders r only thinking that automation is arbrtrage systems or hft ... they r not participating in the coming revolution here ...

[ratmpower 10]

Automated trading works very well if you can come up with profitable algorithms. The trick here is to realize that there are others who are doing exactly the same thing as you, and so you want to be one step ahead of them. You should thoroughly test your strategy before executing in real market.

[Ianprotrader 10]

Automated trading works for big players and institutional. I don't believe that a personal trader can come up with a robot that will win the market constantly. Only if you have a manual technique that works, than you can try to implement it on automated trading, though you probably won't succeed.

[Obsidian 10]

majority of personal traders don't have any programing skills

[electroniclocal 10]

Yes, autotrading does "work", "work" meaning that you can create an automated trading system that is profitable over time with an acceptable drawdown.

 

The expectations for a fully automated system should be different to expectations of an expert trader trading on a discretionary basis as, in most instances, a lot of quick dynamic reaction is lost in an algo. That being said, autotrading can take less skilled traders to CP by mitigating many of newbie reactions it may otherwise be difficult to eradicate.

 

I'd start by reading Bob Pardo's book.

 

EL