Computerized Trading

[Joe Ross]

Computerized Trading

Will trading eventually be done by programmed computers and not by people? Are we really headed that way?

The computer age is bringing about unprecedented change in the markets. Even now it is altering the manner in which we conduct business, interpret events, gather information, and keep ourselves entertained.

While computers can expand our intellectual horizons, they can also limit creative interpretation. There is a tendency these days to let computers do the work of designing and discovering rather than relying upon intuition and imagination. All too often this is taking place even when it flies in the face of reality. In a business context, computers reduce problems to statistical probabilities without necessarily considering the broad effects of events and relationships. No computer can keep you safe from those events which come unexpectedly, and which cause markets to go berserk. Wars, sudden shifts in political power and alliances, and natural disasters, can cause markets to become suddenly and extremely volatile. Even when statistics take such extremes into account, how do you defend yourself if you are long and a market crashes?

I’m not saying that computers shouldn’t be used to prove or disprove theories. But keep in mind that the intuition of the human mind has not yet been duplicated by electronic circuitry. Our educated insights are the critical tools with which we learn and comprehend how markets work.

All the best in your trading,

JR

 

[cj12]

The best computers in the world are your own brains

 

[SOCRATES]

Computerized Trading

Will trading eventually be done by programmed computers and not by people? Are we really headed that way?

The computer age is bringing about unprecedented change in the markets. Even now it is altering the manner in which we conduct business, interpret events, gather information, and keep ourselves entertained.

While computers can expand our intellectual horizons, they can also limit creative interpretation. There is a tendency these days to let computers do the work of designing and discovering rather than relying upon intuition and imagination. All too often this is taking place even when it flies in the face of reality. In a business context, computers reduce problems to statistical probabilities without necessarily considering the broad effects of events and relationships. No computer can keep you safe from those events which come unexpectedly, and which cause markets to go berserk. Wars, sudden shifts in political power and alliances, and natural disasters, can cause markets to become suddenly and extremely volatile. Even when statistics take such extremes into account, how do you defend yourself if you are long and a market crashes?

I’m not saying that computers shouldn’t be used to prove or disprove theories. But keep in mind that the intuition of the human mind has not yet been duplicated by electronic circuitry. Our educated insights are the critical tools with which we learn and comprehend how markets work.

All the best in your trading,

JR

Have you any information on PodG Logic ?

 

[jmreeve]

Will trading eventually be done by programmed computers and not by people? Are we really headed that way?

London equities are already almost entirely computer traded.
However, this has only significantly changed the short term market structure.

 

[clylbw]

I believe one of the advantages of computerised trading is that it takes out human emotions completely, so no possibility of errors in this part. However, it also takes out a powerful tool, the intuitive human analyses, at the same time.

Actually, being a discretionary trader, I do not mind computerised trading at all. It has created lots of opportunities in the short-term markets, certainly in e-minis :cheesy: .

Sometimes it is really interesting to see how the machines generate trades and how such behaviour generates some pretty signals.

 

[Tuffty]

"I believe one of the advantages of computerised trading is that it takes out human emotions completely"

Only if a human has not been involved in the setting up of the computer program, in the running the computer or the computer is using trading money that does not belong to a human. I've not yet met a computer like that, have you?

 

[clylbw]

Hi Tuffty,

Maybe I have been less than accurate. What I meant is, I believe computerised trading diminishes the possibility of some emotion-induced mistakes a discretionary trader like me may make, e.g. hesitating in entries or exits due to fear or greed.

But my apology if I have got it wrong. What do I know? I am always a discretionary trader as mentioned and never a computer guru.

 

[Quenkish]

Have you any information on PodG Logic ?

Please do explain my friend, what is PodG Logic ?

 

[DaveJB]

One often overlooked bonus, in my view, is that each computer program (or set of routines/programs) will trade at least slightly differently... just as humans do now. This offers the discretionary trader with their superior reasoning ability (reason will generally beat an expert system over a long haul) an 'edge' for the forseeable future.

Should computers ever reach the point where they can trade consistently to better effect then humans withdraw, the market then rapidly resolves itself into a stable state as programs cycle against each other in ever faster loops..... or to put it simply, once a computer program can consistently beat a human, and is put into widespread use, the market rapidly moves to stagnation .... and at that point we'll probably invent a new way of providing what the market was actually designed to produce originally - capital for people with good ideas.

I think PC's/programs are badly overestimated in their impact - 'the program approach' is spoken of as if there's just one program producing a standard response to inputs.... well, firstly there are a wide range of programs out there (candle, bar, P&F, Elliot Wave, Gann swing etc), they all have varying degrees of input (fundamentals, price data, different time periods).... the possible combinations are so huge that there simply isn't a 'computer approach'. Essentially I think that despite the advent of computers, there are so many variations in computer implementation that there's little difference between matching yourself against 200,000 people or 100,000 people and 100,000 PCs.... and long may that continue <g>

As long as one trader thinks he/she knows more than the rest of the market there'll be something to trade....

Dave

 

[SOCRATES]

PodG Logic

Please do explain my friend, what is PodG Logic ?

Hello Quenkish, I am more or less fiinished for the morning and I regret delay in replying to your question, this is because I have been making background searches. After much detailed research and with the discreet cooperation of intermediaries available to me, I have been able to garner a few details that may serve to cast some light on this mysterious project / topic.

The acronym PodG Logic according to consensus from a handful of very well informed sources appears to stand for Precision Ortho-Dynamic Gradate (System) Logic. It is smugly intimated success has finally been achieved in creating the truly first fully automatic black box system.

Apparently this breakthrough bolt on is capable of selecting instruments in order driven markets with significant liquidity, selecting the entry point, placing the stop, trailing it, selecting the exit, removing the stop and thus executing the trade automatically, putting itself on alert / lookout by sweeping for new signals in a variety of chosen instruments, selection, definition of direction and potential of move, and then replicating the cycle, with failsafe parameters. This would usher in the era of fully automated trading and is a revolutionary concept.

This is the only information available to me from my current sources which I deem to be reliable. When discreetly attempting to widen my circle of enquiry I have met with repeated stonewalling. One in particular was very aggressive and threatening. The others targeted for questioning flatly refuse to discuss it. Nearly all of them have been quite angry with me and demanded in a threatening manner to know the circumstances of how I came across the term, and the information gleaned so far

One source I spoke to mentioned a target date of 10th July, but did not specify any further details. The words Oh Linda or Olinda or Orlinder appear to have some significance in this as they were overheard during to the discussion , I quote "nothing about Olinda". This is all I have been able to find out so far. Very intriguing indeed.

 

[Quenkish]

Hello Quenkish, I am more or less fiinished for the morning and I regret delay in replying to your question, this is because I have been making background searches. After much detailed research and with the discreet cooperation of intermediaries available to me, I have been able to garner a few details that may serve to cast some light on this mysterious project / topic.

The acronym PodG Logic according to consensus from a handful of very well informed sources appears to stand for Precision Ortho-Dynamic Gradate (System) Logic. It is smugly intimated success has finally been achieved in creating the truly first fully automatic black box system.

Apparently this breakthrough bolt on is capable of selecting instruments in order driven markets with significant liquidity, selecting the entry point, placing the stop, trailing it, selecting the exit, removing the stop and thus executing the trade automatically, putting itself on alert / lookout by sweeping for new signals in a variety of chosen instruments, selection, definition of direction and potential of move, and then replicating the cycle, with failsafe parameters. This would usher in the era of fully automated trading and is a revolutionary concept.

This is the only information available to me from my current sources which I deem to be reliable. When discreetly attempting to widen my circle of enquiry I have met with repeated stonewalling. One in particular was very aggressive and threatening. The others targeted for questioning flatly refuse to discuss it. Nearly all of them have been quite angry with me and demanded in a threatening manner to know the circumstances of how I came across the term, and the information gleaned so far

One source I spoke to mentioned a target date of 10th July, but did not specify any further details. The words Oh Linda or Olinda or Orlinder appear to have some significance in this as they were overheard during to the discussion , I quote "nothing about Olinda". This is all I have been able to find out so far. Very intriguing indeed.

very interesting indeed socrates let us await developements.

 

[RUDEBOY]

Tell you what, lets all buy a computer, put in some fancy software, and hey presto, were all on the make, enough said about that subject? First to make a million doesn't stink of poo!

 

[Tuffty]

"This would usher in the era of fully automated trading and is a revolutionary concept."

Why would this would be revolutionary?

Anyway there is no such thing as 'fully automated trading' in my world as computers can't program themselves.

 

[Swang]

Hello Quenkish, I am more or less fiinished for the morning and I regret delay in replying to your question, this is because I have been making background searches. After much detailed research and with the discreet cooperation of intermediaries available to me, I have been able to garner a few details that may serve to cast some light on this mysterious project / topic.

The acronym PodG Logic according to consensus from a handful of very well informed sources appears to stand for Precision Ortho-Dynamic Gradate (System) Logic. It is smugly intimated success has finally been achieved in creating the truly first fully automatic black box system.

Apparently this breakthrough bolt on is capable of selecting instruments in order driven markets with significant liquidity, selecting the entry point, placing the stop, trailing it, selecting the exit, removing the stop and thus executing the trade automatically, putting itself on alert / lookout by sweeping for new signals in a variety of chosen instruments, selection, definition of direction and potential of move, and then replicating the cycle, with failsafe parameters. This would usher in the era of fully automated trading and is a revolutionary concept.

This is the only information available to me from my current sources which I deem to be reliable. When discreetly attempting to widen my circle of enquiry I have met with repeated stonewalling. One in particular was very aggressive and threatening. The others targeted for questioning flatly refuse to discuss it. Nearly all of them have been quite angry with me and demanded in a threatening manner to know the circumstances of how I came across the term, and the information gleaned so far

One source I spoke to mentioned a target date of 10th July, but did not specify any further details. The words Oh Linda or Olinda or Orlinder appear to have some significance in this as they were overheard during to the discussion , I quote "nothing about Olinda". This is all I have been able to find out so far. Very intriguing indeed.

If you dont know just say so

 

[TheBramble]

... computers can't program themselves.

They can and they do.

One big question is (in any context) how to specify the criteria and basis for self-initiated re-programming. Tougher than you might imagine.

But even more important (and a killer) is the method by which the significance of trigger value or higher events will 'organise' the necessary change in perspective of the external environment while being sufficiently robust and self-aware (I kid you not...) to initiate the self re-programme, preserving the internal integrity of the 'ability' of the core code to guard against contamination of the very basis for automated re-initiation.

The technical bit (the code to enable automated re-programming) is the easy-peasy bit.

Stopping the silly sods from altering their own code to the extent they can no longer alter their own code is the hardest part!

 

[Quenkish]

Tell you what, lets all buy a computer, put in some fancy software, and hey presto, were all on the make, enough said about that subject? First to make a million doesn't stink of poo!

great idea rudeboy.

on a more serious note some people do trade this way mainly the quants look at de shaw, renaissance technologies & to some extent jwh.

it is rumoured that renaissance technologies are now the most successful trading company in the world at the present time.

 

[JohnDoe]

"I believe one of the advantages of computerised trading is that it takes out human emotions completely"

Only if a human has not been involved in the setting up of the computer program, in the running the computer or the computer is using trading money that does not belong to a human. I've not yet met a computer like that, have you?

Excellent points.

As for the rest......

Computers are most widely used to efficiently execute large commercial orders. About 60-70% of computer volume is executed for this. The remaining part is executed for speculative reasons. Most scalp for cents and ticks due to the superior speed offered, rather than any other reason. A couple of servers whirling in the store room are a lot more cost effective than a desk full of traders. They may not make quite as much, but they sure do save a lot more though, so the bottom line goes up. Is a business decision more than a trading decision per se.

As for 'swing' trading or other tom-foolery black box systems, well they are all heading for the inevitable!

D.E. Shaw are very proactive in this area - but they tend to be at the cutting edge and early adopters of computer trading, and dont make as much money as perhaps they think they should. Others tend to develop their methods further and profit more - mostly as ex-employees leave and go work for other funds.

It is estimated that there are only a handful of truly original algorithms in the world. All others are simply modified and re-hashed versions, again, adopted and adapted by employees as they move around Wall Street from job to job, learning from colleagues (these IT folk like to massage and show off their egos to their new boss far too much to be in this business) etc.

Some of the biggest names in quantative models include Taleb, Niederhoffer, and of course everybody's favourite...LTCM! Need I say more??

 

[SOCRATES]

F.A.T.C. (Fully Automated Trading Capability) and PodG Logic (Precision Orto Dynamic Gradate Logic) I can confirm as the result of news just come in, are not connected to Victor Neiderhoffer at all.

It appears as if the bolt on fully automated trading project has split into two divisions, apparently one for Forex and the other for everything else except traded options.

My source promises to have more information later in the week. Apparently trials are under way but everyone is very cagey and secretive.

It is intimated there is a difficulty, but despite repeated gentle prodding have not been able to discover anything further. Any body else know or heard anything ?

 

[donaldduke]

The human mind has spent millions of years evolving into something that is extremely bad at trading,
probably even worse than a random coin flip.

Why rely on something that is naturally so flawed at trading.

Computers do not suffer from fear, anger, hope, laziness, revenge,
impatience,stress,tiredness,greed,boredom,excitement, pride,ego .. the list is endless.

Computers dont gamble with losses and take profits early, they dont double up after losing, they
dont try to get back at the market after losing.

Computers do not suffer from these things, sure humans still have to develop and run the programs
but you want to reduce the human element as much as possible.

Some people have reprogrammed their brains to do the unnatural things that work in trading or were
lucky enough to be born that way, but for the rest of us not blessed with such luck the computer offers
a way to reduce the exposure to such tendencies.

 

[donaldduke]

Tell you what, lets all buy a computer, put in some fancy software, and hey presto, were all on the make, enough said about that subject? First to make a million doesn't stink of poo!

There are many fully automated hedge funds that pull in millions out of the market each year.

Computer are only a means of executing a system with consistency and discipiline,
there is very little difference between a mechanical system and a discreationary system.

Traders have made/lost millions with mechanical systems and without using mechanical
systems. There isnt much difference. For some reason people think if you have a profitable
mechanical system you should be able to take over the world, this doesnt happen for
the same reasons as someone with a profitable discreationary system doesnt take over
the world either.