Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Trading Myths and Some Forex Math 337 replies

What are your favorite trading myths? 145 replies

Popular Trading Tips: Truths or Myths 36 replies

Ridiculous Trading Myths 323 replies

Dispelling Some Myths on Trading Practices 23 replies

Trading Myths

Quoting merlin
this is a great topic to debate

here is my take...

quant systems... there is NO universal system. however, the more universal the system, the better. its always a good sign to see a system work on the EUR/USD and the USD/JPY, as the pairs have very different characteristics so the system must be robust to be profitable in both. it would be even better to see the system work on a completely different market like crude oil. with this said, most of my quant systems are taking advantage of market inefficiencies that only exist in a very specific place. therefore, i trade very few universal-type systems. in quant trading i feel its more profitable to have a tight focus.

discretionary trading... YES, there are universal strategies that work. as i have pointed out many times, i usually trade the "human element" of the market (ie reading the order flow and guessing what the major players are thinking/doing). i use the same exact philosophy in the stock market, currencies, oil, bonds, you name it. and it works everywhere, in every timeframe from 5mins to 5years, because the human element remains the same everywhere. so while i dont believe in universal systems, i certainly believe there are universal strategies and approaches that are profitable in every market in every timeframe.

SOOOOOO TRUE
Just Trade It
Great post Merlin
Quoting merlin
...the "human element" of the market...
Yep that's why machines will never outdo humans in the long run... besides... machines are only as good as the guy who made them (they're not perfect either). Lucky for us it's a lot easier to adapt to current conditions and future conditions. All markets change every few years in some way or another and we can adapt without having to scrap months or years worth of coding and testing.
Quoting MoForce
besides... machines are only as good as the guy who made them
exactly!!!! im fond of saying that more decisions go into building a system than goes into a year of discretionary trading.

i dont know about you guys, but when i build a system there are thousands of little decisions involved. systems trading IS discretionary!
Relax and be happy.
Quoting MoForce
Yep that's why machines will never outdo humans in the long run... besides... machines are only as good as the guy who made them
There's been genetic software evolution algorithms for years. They learn and adapt. Some of the algorithm have been showcased by putting them into games where you get creatures that actually "learn" the things you show them / teach them. This includes language e.g., albeit this is all in it's infancy. It's no small wonder they also tried to see whether they could teach those algorithms how to fly a combat jet on their own...(I believe it turned out they could, just the landing bit was a bit complicated for em Then again I don't think that's gonna bother many..) Btw, speaking of machines, FF is completely broken for me. 404s, totally broken formatting, etc...
Trust price. Know yourself.
Quoting merlin
this is a great topic to debate

here is my take...

quant systems... there is NO universal system. however, the more universal the system, the better. its always a good sign to see a system work on the EUR/USD and the USD/JPY, as the pairs have very different characteristics so the system must be robust to be profitable in both. it would be even better to see the system work on a completely different market like crude oil. with this said, most of my quant systems are taking advantage of market inefficiencies that only exist in a very specific place. therefore, i trade very few universal-type systems. in quant trading i feel its more profitable to have a tight focus.

discretionary trading... YES, there are universal strategies that work. as i have pointed out many times, i usually trade the "human element" of the market (ie reading the order flow and guessing what the major players are thinking/doing). i use the same exact philosophy in the stock market, currencies, oil, bonds, you name it. and it works everywhere, in every timeframe from 5mins to 5years, because the human element remains the same everywhere. so while i dont believe in universal systems, i certainly believe there are universal strategies and approaches that are profitable in every market in every timeframe.
My experience with quant systems (and I loosely use the term quant as it normally applies to option pricing etc; I should probably say algorithm here) is that if you want it to work reasonably well over something like the EUR/USD and the USD/JPY (if I'm taking this approach its probably related to portfolio size) then I would build the system around a USD Index where a single trigger in the system can be spread across the underlying pairs. This makes the model extremely scalable and helps protect the trading process itself; you can also be a little more covert with the position building as well.

The other thing I've discovered and have found support for in other peoples work is; you pretty much have to constantly update the quant model. Randomizing a dataset once and then using the prediction going forward for an extended period doesn't seem to work as well as regressing (curve fitting) from a sequential set of data and updating the algorithm every quarter. That's my experience at least... I unfortunately really don't understand the mathematics behind the process and can't really get a handle on the why. It probably has something to do with the curve fitting process and longer term price action momentum, but that is just speculation on my part.
Trade Small. Be Patient. Learn to Survive.
Quoting merlin
this is a great topic to debate

here is my take...

quant systems... there is NO universal system. however, the more universal the system, the better. its always a good sign to see a system work on the EUR/USD and the USD/JPY, as the pairs have very different characteristics so the system must be robust to be profitable in both. it would be even better to see the system work on a completely different market like crude oil. with this said, most of my quant systems are taking advantage of market inefficiencies that only exist in a very specific place. therefore, i trade very few universal-type systems. in quant trading i feel its more profitable to have a tight focus.

discretionary trading... YES, there are universal strategies that work. as i have pointed out many times, i usually trade the "human element" of the market (ie reading the order flow and guessing what the major players are thinking/doing). i use the same exact philosophy in the stock market, currencies, oil, bonds, you name it. and it works everywhere, in every timeframe from 5mins to 5years, because the human element remains the same everywhere. so while i dont believe in universal systems, i certainly believe there are universal strategies and approaches that are profitable in every market in every timeframe.

Well said Merlin.
Base on the poll, so far, there are at least 24 people who believe in a "Universal System". I just want to clarify that there is a major difference between analyzing the market and a trading system.

Let me give you an example. If you a see a "Double Top" formation then you may conclude that the price will drop very soon. This is analysis. You may even say the extent of the price drop.

But a trading system based on double-top is beyond the analysis. This system answers the following questions:

* Recognition of the patter
* Point of entry
* Profit target
* Stop Loss
* The number of lots to Buy
* Money management settings (for example when to secure the profit)

This is an example. In reality the trading systems can vary and their components can be a lot different from each other.

The bottom line is that a trading system is beyond analysis of the market. Your trading system may enter the market in the wrong direction of the price movement but eventually make money. This means that your analysis was wrong but the system worked.

With this definition in mind I believe that the odds you find a Universal System is slim to zero. You may find a method of analysis that works for different time frames, currency pairs, or even different markets but the odds that you find a system that can trade any time frame plus any currency pair is almost zero. I do accept - as mentioned - by some of you that there are systems which are more flexible and less independent on time frame or the currency pair but in my humble opinion the concept of "Universal System" or whatever you call it is more of marketing nature than reality.
[quote=SeekingLight;1714126]To this and in reply to pretty much all...

So yes, adopting this perspective there's a universal system, if you view a trader as a system(this IS a viable perspective, as imho you must actually do this if you've got several under management - each employed trader over time becomes a "system" with expectancy and risk modalities and so forth..) as I expect he will be able to do it on any market that has bars and ticks

SL,
Being a Discretionary Trader....I agree with your Statement. The trader, trading the "Human element" (see Merlin Post ) with an unbiased slant, though informed of over all Market sentiment....which can change like the weather some days, leaves discetionary trading to become, in effect, the system....and the system being the trader's ability to adapt to shifting sentiment.
Please give the benefit that I am less of a "writer" of thoughts as I am Trader of the charts.
Fxj
Proper Prior Planning Prevents Piss Poor Performance
Quoting SeekingLight
There's been genetic software evolution algorithms for years. They learn and adapt. Some of the algorithm have been showcased by putting them into games where you get creatures that actually "learn" the things you show them / teach them. This includes language e.g., albeit this is all in it's infancy. It's no small wonder they also tried to see whether they could teach those algorithms how to fly a combat jet on their own...(I believe it turned out they could, just the landing bit was a bit complicated for em Then again I don't think that's gonna bother many..) Btw, speaking of machines, FF is completely broken for me. 404s, totally broken formatting, etc...
Well.... let's put it this way... would you rather have your life depend on another human or a machine? I'd rather have another human...
Quoting MoForce
Well.... let's put it this way... would you rather have your life depend on another human or a machine? I'd rather have another human...
*tries to picture a human replacing a respirator machine*
Trust price. Know yourself.
Quoting SeekingLight
*tries to picture a human replacing a respirator machine*
well... there are humans constantly watching that respirator so you're still really depending on humans... any wrong move will have that respirator's plug pulled. lol
  • Post #33
  • Quote
  • Edited 4:17am Nov 20, 2007 3:49am | Edited 4:17am
Quoting MoForce
well... there are humans constantly watching that respirator so you're still really depending on humans... any wrong move will have that respirator's plug pulled. lol
We don't have that sort of staff in hospitals. That's why the machines monitor heart rate, blood pressure and oxygen levels and alert the people themselves.

You might wish that there are humans constantly watching - I can assure you, they are not.

They have far too little time and reason for that. Where's the point in watching someone 24/7 breathe?

It is the machine(s).

It's also ridiculous to so desperately try and save a point. There are plenty of other occassions where human precision would be to slow, inaccurate and biased to trust a human life with and machines and computers are used instead.

And no, I am not going to supply all the thinking for you and put 50 examples as well. Spend a few minutes on the subject and you'll realize it or not.

We trust our lives to machines every day. I don't know whether anyone really WANTS wants it or not I don't know, but that's just how it is. Going by the fact we're not in medieval times anymore, mass manufacture and purchase them I guess most have accepted it.
Trust price. Know yourself.
sometimes i wonder, even if the option is simple, most of the machine do is to do the repetitive job, but not the start (decision),
there is that part which human wish not to be replaced , no?
that tool, no matter how hitech it will be, there always human in charge / maintenance division.
one thing that i remember is, while they do that job, we can do something else (better thing, hopefully). education become more and more important because many simple job are replaced already. compete with machine is silly in some point,
well on continuing this off-topic discussion... check this out...

The video is playing on the right side of the page...just wait for the ads to finish
  • Post #36
  • Quote
  • Last Post: Nov 20, 2007 9:40pm Nov 20, 2007 9:40pm
Off-the-wall off-topic
0 traders viewing now