Discretionary Trading vs Automated Trading

[januson 10]

TheBramble-> No I don't think you misunderstood something, it' was perhaps me who answered a question without any thoughts.

 

What do we need to know about a stock's price and volume profile to assess whether it is simply being ignored or is being quiety accumulated.

 

So if I understand you correct, then we need to know what the throughput normally would be to keep the price unchanged. What marketmakers buys and whom sells and at what rate.

 

Could a sort of a advanced broker statistic tell us what we need to know?

- For the last couple of months I've been thinking about a system which collects sell and buy for a specific stock, it should be self-learning in that manner that every time a broker buy stocks it would accumulate his total holding and vice versa with the selling.

If the broker sells more stocks than he owns could we then say it's shorting or do we need to adjust his total holding?

In the beginning such a system would be very very unprecise, but as time goes it would reflect broker holdings and maybe also behavioral patterns.

[notouch 10]

There are a number of products on the market now that enable trading decisions to be made instantaneously on the basis of news feeds. For example this from Dow Jones:

 

"Now, Dow Jones, the leader

in real-time financial news,

introduces a faster, structured

news data feed for automated

trading and analysis: the

Dow Jones Elementized

News Feed. This advanced

solution delivers headlines

in precisely tagged elements

on a low-latency feed, all

designed for direct integration

into quantitative analysis

models and computer trading

programs."

 

"The innovative Dow Jones feed

revolutionizes how newsflow

can be interpreted and acted

upon for algorithmic trading,

quantitative analysis and

execution management.

The Dow Jones Elementized

News Feed gives hedge funds,

banks and securities firms

an enhanced news source

for analyzing and identifying

trading, investing and hedging

opportunities – and moving

on information in milliseconds."

 

These sorts of products are the reason the market moves instantaneously to, for example, Bernanke testimony and FOMC statements as well as terrorist attacks.

[Thoth 10]

Back to Oddball, wouldn't this system become obsolete because everyone is trading at the same time and eventually cancel themselves out? Eventually the market will absorb and change to make the system becoming useless. I'm speaking of too many using the same system, not systems in general as useless.

 

The core idea of oddball was really sound thinking. In many variations I think that lots of institutions are still using it. Look at the volume of sp trades on the top of the hour and it may tell. It really was an amazing system.

 

Maxux

[Guest fx-programmer]

Guys, trading is a mathematical game, so in essence you cannot compete with a calculator.

I hold a Doctorate degree in computer sciences,& I worked for intel for a couple of years until I left to the trading world.

The first year into trading I programmed every possible indicator but I was loosing all the time, until I finally realised that there is something I was missing.....Thus, I joined with two past NYSE pit traders who actually gave me the perspective I was missing, but I still lost since I tried to capture pure mathematical nuances which are essentially Math patterns that work for a certain market, ie trending ....But I still lost since the market is trending only 40%+- of the time....

 

Anyway, after two years I finally made an algo so complex with artificial intelligence analysis etc and now it finally works. To make a long story short, I now execute for a little fund, and we monitor about 50 dif vehicles with pin point precision 24/7 - Forex and Futures - and to the point : there is no way a human can immitate and execute with such accuracy........I used C#, C++ and direct interbank FIX protocols to beat spikes etc...

 

PS - If anyone is interested in automating a strategy I can help - so PM me on that.

Anything you can describe in words I can program !!!

[brownsfan019 17]

... there is no way a human can immitate and execute with such accuracy...

 

A human trader doesn't have to match the program's speed unless you trading vehicles that are so volatile that you can't blink w/o missing something. Quite a few of us here trade the e-mini futures and it's not exactly hard to get your trades executed... even in quick markets.

 

I'm sure this really, really complex computer program is making money now, but over time, just like any other program out there, it will die. We have yet to achieve the point where a computer program that is made by humans can adapt to changing market conditions. So the question then becomes what happens when (not if) this 2 yr project stops working? Back to square one? No thanks... I'll just do the adjusting myself and stay in tune with the markets.

[Guest fx-programmer]

The core of AI is the learning capabilities of the computer program to adapt.

Essentially your brain is a biological computer program which is prone to tiredness and emotional psychology.

The program does loose, but it learns.

 

Elliott wave analysis is based on human psychology but lacking in momentum, yet big funds still use it.

It takes only a few mil to move the futures market, but the Forex can only be moved by a state.........

 

My program uses fib retracements,fib confluence,market profile,trend confluence points,pivots,gaps,momentum,external news params,Elliott waves,convergence,divergence, etc - now can you track so many params on 50 vehicles at the same time??? I think not !

 

Sorry mate, it is not my place to say that humans lack the capability to profit consistantly, but if you traded long enough you now know that the best way to hedge your bets is to trade many markets at the same time - and this my friend can be done better by a network of CPU's like we do.

 

 

 

 

A human trader doesn't have to match the program's speed unless you trading vehicles that are so volatile that you can't blink w/o missing something. Quite a few of us here trade the e-mini futures and it's not exactly hard to get your trades executed... even in quick markets.

 

I'm sure this really, really complex computer program is making money now, but over time, just like any other program out there, it will die. We have yet to achieve the point where a computer program that is made by humans can adapt to changing market conditions. So the question then becomes what happens when (not if) this 2 yr project stops working? Back to square one? No thanks... I'll just do the adjusting myself and stay in tune with the markets.

[Thoth 10]

I found this page on this method, is this as true to the original concept you mentioned?

 

Trader.Online.pl® - MetaStock™ Zone - Oddball S&P System by Mark Brown

 

Pretty Much I think that is Mark Brown's OddBall System, he proved you can look at one market and trade another. Also note the web sites bottom note:

 

This avoids the pitfall of basing real trading on unrealistic system tests that generate signals on the close at the end of the session

 

 

Note: If a trade is signaled at 4 p.m. EST, you have 15 minutes until the close of the market to place the trade in the S&P futures (it is not possible to do this when trading the SPY). This avoids the pitfall of basing real trading on unrealistic system tests that generate signals on the close at the end of the session, when the trade can in practice only be initiated the next session. In such cases the price may have moved farther away from where the test indicates the system was filled, giving a false reflection of the system's performance.

 

OddBall was a ground breaking system and it still works pretty good - it was free keep in mind. Noted by profs at MIT as one of if not the best trading system ever developed.

[torero 10]

Is anyone using this system or a modified version of this system? Any success?

[Jakew 10]

My take on the subject is that both automated and discretional trading can be valid.

 

However very short term trading is probably better served by discretional decision making, especially as to entry and the type of setup to use (because of the "market climate" prevailing), though not the exit decision. Exit must be automated even in discretional trading just because human nature will statistically do more harm than good in the realm of exits. I other words I belive in intuition at entries but not exits. It is so because when thinking/feeling (whole brain) about entry you are not in the trade yet, and the condition of being in the trade changes your intuitive capabilities for the worse. For entries at least I can use emotional triggers with considerable benefit, it doesn't work with exits in my case. Emotional input cannot be really automated, or in a way it can be (say 1 for I feel this looking at the chart/tape, 0 for I don't feel this) but would it be still... well... automated trading, with emotion of an individual coded in?. Perhaps, but not a black box for sale, just your own one of its kind wholistic system?.

 

For relatively longer term trades I suppose automation can work better than for very short term trading because of the logic of the R-multiples of these trades: it is not as important if the entry suceeds (and it is very important in very short term trading) becuase the opportunity cost of not taking a systemic entry is way too high to even think of not taking an entry - just because you risk very liitle with your stop loss compared to the much more formidable risk of missing a high R-multiple win.

[waveslider 10]

Good point Jake about how your mindset changes once in a trade. If you are trading discretion it is very hard to see the market move your way without moving your stop up too aggressively. Even if you have a sound methodology, you trick yourself into getting out too early in order to hold onto profits.

 

A successful fund manager once mentioned to me his idea of the best mechanical system:

 

Take a room full of newbie traders and tell them they are trading real money, let them trade it however they want. Then fade their emotional decisions.

[Reaver 10]

I firmly believe that if a system were able to be automated to the point it could make money...eventually it would end up with all the money in the markets, given enough time of course.

 

That is why I do not believe 100% fully automated systems are feasible as far as profiability goes.....just doesn't make sense to me. Although of course, the paradox lies in the fact that many system traders are in fact profitable...so to each his own...Yes I am aware that I contradicted myself...life is full of mysteries, huh?

 

 

But I have my own beliefs, and they preclude me from automated trading.

 

I prefer using my brain and my understanding of the market-however small it may be- to make decisions...maybe it is the challenge of figuring things out that makes we lean this direction..but I find it rewarding to make money based on my decisions, and not the decisions of a formula, system or computer....

[Dogpile 10]

thought this was interesting rumor/fact on something that came out end of this week regarding 'program trading':

 

"one of the largest firms specializing in algorithmic "black box" auto systems (market making), had taken a rather large hit the prior few days, and had turned off all their auto-trading systems."

 

http://futurepathtrading.com/content/view/362/69/

 

(good blog by the way)

[Reaver 10]

thought this was interesting rumor/fact on something that came out end of this week regarding 'program trading':

 

"one of the largest firms specializing in algorithmic "black box" auto systems (market making), had taken a rather large hit the prior few days, and had turned off all their auto-trading systems."

 

http://futurepathtrading.com/content/view/362/69/

 

(good blog by the way)

 

 

Thanks for the link.

 

Not knocking those who choose automated systems...but if I am going to lose my money, I want it to at least have been my call to do so.

[Nvesta81 10]

This link is recent and somewhat pertinent to the discussion.

 

http://money.cnn.com/2007/08/09/markets/goldman_fund/index.htm?postversion=2007080915

[brownsfan019 17]

Interesting articles guys, thanks for sharing. The thing is, we'll never really know what has caused this tremendous volatility. My guess it's a combination of a whole bunch of things occurring at the same time.

 

The only question I really have is - when will this volatility dry up and why?

[Soultrader 12]

I have come to a conclusion... discretionary trading wins hands down. But how many traders will ever go on to master the art of discretionary trading? Just a few....

 

I say this because discretionary traders can constantly see warning signals and changes in the market with their own mind and vision. Therefore every setup is a different situation. Systems will never know what hit them until its too late.

[waveslider 10]

I agree with you, James. Discretionary trading does win hands down - in the right hands! There are so many pitfalls out there though that the successful discretionary trader is less than 5% of the trading population. We all know the drop out stats.

 

All mechanical trading does is try to emulate what a discretionary trader would do, but on a larger more diversified scale (one trader wouldn't be able to execute or watch the same number of signals as a computer).

 

This advantage I see in mechanical trading -diversification. Humans simply cannot watch many markets at once with any hope of taking into account all of the data that a computer can.

 

Another advantage is consistency. Humans are emotional and tend to be subjective. On the positive side they are also adaptive. The problem is, is the ability to adapt really helping or hurting?

 

If a computer can identify a consistent edge, it will attack that edge consistently, every day. The markets may change, and in those cases the system may lose. But if the edge is proven over time, the edge will bring profitability.

 

A good discretionary trader will be able to chose between a variety of methods (systems) that are appropriate for that particular time. Mechanical traders often take a shot-gun approach and apply all systems, or just train the system to stay out of the market when it is inappropriate.

 

What you said in your second sentence is all important : But how many traders will ever go on to master the art of discretionary trading? Very few will ever trade consistently. I consider consistent profitability in trading as mastery. The reason the majority fail is not because of the lack of a good method, it is the fact that the method is not applied consistently. Human emotions just screw it up for most of us.

 

So bottom line, if you are one of the few (don't kid yourself) who is able to see 100% objectively and know when emotion is affecting your judgement, then you are one of the few successful discretionary traders who will succeed. But think about it, if you are 100% objective and acting without emotion, you are pretty close to a machine yourself.

 

These comments may disturb some of the discretionary traders on this board. If what you are doing works for you, great! I just think that those that may be interested in the mechanical route should not be dissuaded. Being able to act with the confidence of a statistical edge behind you is an amazing thing, for sure!

 

ws

[enlightedtrader 10]

The recent volatility spike of the markets is a good testimony of discretionary trading vs automated trading. We heard numerous sad stories about hegde funds which trade primarily on computer-driven trading strategies. To name a few, Gold Sachs Global Alpha Fund is down 26% this year. On the other hand, we also heard a lot of discretionary day traders made a killing in the recent market turmoil.

 

For me, this huge contrast shows clearly the edge of discretionary trading over automated trading, i.e., when the market experiences high volatility and turmoil. The more chaotic the market, the more edge will discretionary trading shows over automated trading. Of course, this comparison is based on the assumption of a good discretionary trader.

 

I used to think that discretionary trading would be in danger with the proliferation of automated trading. Now I'm more convinced that discretionary trading can significantly outperform automated trading in the long run. As a result, I'm more determined to become a better discretionary trader.

[Thoth 10]

if taking the time "to do it right" and quality of programming, a human mind "can always be beat, by a computer".

 

humans = 0

computers = 100

[PYenner 10]

The computer can be good for processing data to give signals.

But the decision to act needs to come from an experienced human brain, far far more complex than software still.

One guy said it well, a human knows when it is time to head for safe harbour.

The one thing big trading programs don't like to do is stop trading, that would be like admitting defeat. But knowing when and how to get out or stay out, thats priceless.

[Reaver 10]

if taking the time "to do it right" and quality of programming, a human mind "can always be beat, by a computer".

 

humans = 0

computers = 100

 

Negative ghostrider.

 

Haven't you ever seen Terminator?

[PYenner 10]

Last I heard, IBM had only just got to a supercomputer that could match a chess grandmaster. So I aint sweating about trading computers for a decade or so.

[syswizard 10]

I've been following a small hedge fund manager whose been posting his "systematic" trades on Elite Trader.com. Omigosh, he is just getting HAMMERED and whipsawed to death....he was up HUGE in short YM futures 5 weeks ago, and he held and got stopped out last week for a loss because that's what his system told him to do.

[brownsfan019 17]

We can beat this horse to death.

 

The programs are built by humans. Computers just do what they are told and cannot react to changing conditions unless told to do so.

 

Now, that will probably change years from now, but I hope by then I've got my bank roll together and can watch from the Bahamas. Or maybe Aruba.

 

[feb2865 10]

This is one of the most famous and well-known soap operas in the whole trading world

 

Stops vs. no stops

 

Scaling out vs. all in

 

Indicators vs. naked trading

 

the list goes on.

 

I have seen a lot of traders trying a lot of mechanical stuff, bot's, black boxes, ect

 

none has win consistenly on the markets.

 

But, I am willing to learn, you know, it's wise to have an open mind about everything. If anybody has some mechanical stuff that win consistenly, with average losses, 5 years in a row (a whole cycle according to experts) I am willing to take a look at it.