Google “Kelly criterion” for an introduction
lol I don’t need to google. Few Kelly’s ideas were in link that I gave and my example was simplified idea of the approach. However that might be different ways for self-tuned MMs that are based not only on winning percentage. It might be expected profitability of a trade. it might be median value from monte carlo simulation. It might be… My main idea was MM that self-tuned for a system that works with more than one instrument.
Eu
I whole heartedly seconfd the motion. I also propose that we find a way to rate systems according to their relevance to real world. I know we have realism factor but I am thinking along the lines of risk management.
I know that most systems on c2 are probably for research and development, but the key is to be able to develop a tradable system for general public can be extreamly difficult. For example to convince a novice trader that 20% drawn down is probably the norm for any type of decent return, and why it should be trade in such a way can be even more difficult to convince anyone to trade with such and such strategy. Therefore, I think any trading systems on c2 or anywhere on the market probably would most likely be part of a money management firm rather than individual traders.So c2 is a based to define for research and development. But I think to implement trading systems on a large scale for an average investor, it will be best to leave it to the professional traders.
For example to convince a novice trader that 20% drawn down is probably the norm for any type of decent return
IMHO The kind of “usual” DD is a way to hell. There are two different things. One is actual “around-trading” business and C2 is part of it, when you have to convince somebody to your pictures to make few coins from your con and it very normal for the kind of business. Second thing is more complicated. There might be something and in the case it’s over profitable. (I don’t sell myself. Gosh). But C2 offers probability of risk and if you’re right it offers high reward. Anyway… It’s only my mumbling.
it will be best to leave it to the professional traders.
Sure. I’m giving the link for fourth-fifth time.
http://autumngold.com/
They’re “professionals”. Choose any of them lol. Only one difference for me is they don’t blow up every few weeks as C2 systems. But you have to pay for that. Otherwise, have a look on equity curve from professionals. They nearly match C2 lol.
Professional in trading is somebody who retired from the business. Hopefully with some sum of money. Until you’re retired from the rate race you’re nobody 
Eu
"Professional in trading is somebody who retired from the business. Hopefully with some sum of money. Until you’re retired from the rate race you’re nobody "
You have a point , but you are mising the point, it is not enough to let a beginner trader to jump into a trading system but one should explain the nature of trading , the management of a particular trading system ,and making sure that the trader can follow the system .
"I have discovered the Hunter Memory: in a minutes chart, trace an Exponential Moving Average with length 55 and displaced 21 minutes on the right (future). It seems that prices cannot go too apart from such an average. This is an about sure rule"
So, are you suggesting that occurence happen in all market or only in particular maket such as fast or slow moving market. Why do you believe that Hunter Memory exist in the time frame of only 21 minutes and EMA with length 55?
"I have discovered the Hunter Memory: in a minutes chart, trace an Exponential Moving Average with length 55 and displaced 21 minutes on the right (future). It seems that prices cannot go too apart from such an average. This is an about sure rule"
So, are you suggesting that occurence happen in all market or only in particular maket such as fast or slow moving market. Why do you believe that Hunter Memory exist in the time frame of only 21 minutes and EMA with length 55?
"I have discovered the Hunter Memory: in a minutes chart, trace an Exponential Moving Average with length 55 and displaced 21 minutes on the right (future). It seems that prices cannot go too apart from such an average. This is an about sure rule"
So, are you suggesting that occurence happen in all market or only in particular maket such as fast or slow moving market. Why do you believe that Hunter Memory exist in the time frame of only 21 minutes and EMA with length 55?
"Professional in trading is somebody who retired from the business. Hopefully with some sum of money. Until you’re retired from the rate race you’re nobody "
I disagree. One can never get out of the game. It is true in any business, for eg., in the pure or empirical sciences research and development or in trading the financial markets etc.,
Productiveness constitutes the main existential content of virtue, the day-by-day substance of the moral life; as such, it is a responsibility of every moral being, whatever his finances, Even if a man has already made a fortune, therefore, or inherits one or wins one in a lottery game, he needs a productive career. A rich man may choose, if he has a legitimate reason, to pursue a kind of work that brings him no money. But he still must work. A bum is not a person living a man’s life, even if he has no trouble paying his bills.
In science, for e.g., the Second Law of Thermodynamics was formulated long before most scientists were convinced that matter was indeed composed of atoms. Even the mid-1800s onward, some scientists had been taking the atomic theory of matter more and more seriously, while others remained unconvinced of the the reality of atoms. These critics felt that while the suggestion that matter was composed of atoms was a useful idea that simplified many of the calculations about the properties of fluids and gases, it was nonetheless meaningless to ascribe physical reality to entities that were too small to ever to be seen by the naked eye. Many of the elder statesmen of physics at the time, most notably Ernst Mach (in whose honour the speed of sound in air, the Mach number, is named) held this view.
Nevertheless, the atomic hypothesis eventually won out, via the same strategy by which all revolutionary ideas succeed. As Max Planck, himself a young Turk of the quantum revolution, once remarked, "A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die and a new generation grows up that is familiar with it."
A key development that convinced these younger scientists that atoms were real, regardless of what an older establishment claimed, was in accounting for the jitter that small objects underwent due to their random bombardment from still smaller atoms and molecules striking them from all sides. This phenomenon is termed “Brownian motion” after Robert Brown, the botanist who observed the excursions of a pollen grain in a drop of water using a new scientific instrument of the time, the microscope. While Brownian motion had been known since 1828, it was not until 1905 that a satisfactory theoretical description was provided in the Ph.D. thesis dissertation of another young upstart, Albert Einstein.
This world is neither hostile to innovations in any business nor does it lead to an intolerable rat-race of innovations. Men who are immune to facts and logic have no alternative but to traffic in fantasy.
Brownian motion
It’s interesting that you mention Brownian motion, because I think there are trading system using the Brownian motion to trade the financial system, however I am not sure how profitable in the real world.
Einstein was able to quantitatively calculate the excursions of a pollen particle due to the collisions with the water fluid in which it was suspended, and also relate the magnitude of the fluctuations to the temperature of the surrounding medium. The close agreement between his calculations and experimental observations convinced many physicists that the atomic hypothesis was indeed correct. While his thesis research was not as revolutionary as his subsequent investigations of relativity (published the same year), he would have been well known to scientists even if his only contribution to science were his elucidation of the statistical nature underlying Brownian motion.
Similarly, it is possible to elucidate the statistical nature underlying the financial markets; i.e., it is possible to predict the future price of a financial instrument based on its present price with certainty by calculating the forces acting on it.
One can’t afford to forget the past because one would not then recognize the future for what it actually is (better or worse), for the future is only the past again, entered through another gate. So, until the time we discover celestial gateways (wormholes etc.,) through which we can time-travel to and return from the future, I believe that only a few minority can consistently and accurately determine the future based on the present but no one can be correct about it all the time.
Eventhough no one can perfectly predict the future based on the present, one can assemble an objective base of evidence that can help any investor make more informed decisions away from the clouds of emotions.
So, in order to have a profitable trading system it would mean to control the present factors so that it can be profitable in the future. But the delima is if we control the present market at 100% (efficiency market theory) which mean that future value would only reflect distribution. Therefore, any reflection of the current trading system is the reflection of distribution and some systems wil fall into positive distributions. And because of the on going adjustment in the market , any major adjustment in the present market will break the system apart. So any adjustment in the present is actually for the future distribution, and I guess that’s why computer based trading system can be profitable if we know the future distribution. Because computer based trading system has no emotion it is much easier to make decision. Therefore, is it possible to create an artificial intelligence to trade the financial market? And what parrameters would have to be involved (neuro-net trading system) to implement such a complex task. I think rule based trading system works as long as you are on the right side of the market. Basically , with the current trading system , we can say ok if we are in the bull market then apply this system, bear market trade this system, so all the back testings show how the market perform in various market but it does not prove anything , unless the future market condition mimic the previous market conditions.
PS, has anyone try to trade Neural Network Engine?
Therefore, is it possible to create an artificial intelligence to trade the financial market?
Yes, one can create artificial intelligence to trade the financial market successfully because markets are only frequently efficient, not always. One can mimic the brain using a computer. The world chess champion Kasparov was beaten by the big blue.
Deep Blue defeated Kasparov, because the equivalent of dozens of chess expert’s knowledge and strategies were loaded into it, it had immense computing power, and computers don’t suffer the occasional lapses that people make.
It is not necessarily a good comparison, because A.I. only excels in extremely-constrained domains, where the rules are crisp and relatively easy to code (like chess playing). They are still strugggling to make an autonomous vehicle (The military has competitions for the purpose of building and demonstrating unmanned vehicles that can operate out in the real world, in mixed terrain, but it has been dismal so far)
The reality is, that a cricket has more general-world capability than any A.I. program or robotics (A.I. was my minor, when I got an MS in computer science).
>It is not necessarily a good comparison, because A.I. only excels in extremely-constrained domains, where the rules are crisp and relatively easy to code (like chess playing).
But you don’t need A.I. to trade. Indeed many of the best trading systems
are “extremely-constrained and the rules are crisp and relatively easy to code”. You don’t have to code for everything, just a few winning situations.
Or better yet for just one:
If A + B = C then buy;
Exit on close…;
It can be more like checkers than chess for that matter. The problem is
most traders want it all. Some simple way to win most of the time isn’t good enough.
"But you don’t need A.I. to trade. "
Correct. I have seen very few A.I. packages that had any lure potential to my advisory-seeking dollar…
I think, the question is not why would you need AI to trade the financial market , but can you create an AI to trade the market. Of course anyone can trade the market just flip a coin and manage your trade probably would do better than most trading system. The point is we are looking for ways to improve our trading process wheather it’s the moon that affecting our decision making or we just lost 50% of our porfolio should not dictating our research. I think pure research can be very challenging, it gives us the ability to stay ahead and drive us to do what we do best. As I mention before, back testing to me doesn’t mean a lot but when I used the system to forward testing and come out positive , that’s in itself is an archievement. So we should not disreguard any research that can improve our daily life. Yes, I know that AI is probably a very large project in the financial market, but what if we can create such a dynamic language that could teach AI how to process data from the beginning to the end, and able to adjust on its own. So far I have found one company that has involved in neuro net , but I am just currious as to how many researchers are actually extensively looking into these projects. I notice that most systems on C2 are based on action-reaction and there is nothing wrong with that. I am just currious if there are any systems using a different methodology beside moving average, momentum, ect… to trade the financial market. I think at the end, AI would have to think like a trader, basically with a very tight and simple set of rules.
My opinion:
1. A feasible neural network is much less intelligent than a worm. There is no reason to expect that it will outperform any other programming method.
2. The behavior of a neural network depends entirely on (a) the number of layers and other structural features of the network, and (b) the data that you feed to the network. So asking “can a neural networks learn to trade” is imho like asking “can we make a good trading system with Java”, the difference being that neural networks are more like an extreme low-level program language. The answer is probably “yes”, but the question is: how?
3. Even in simple examples (like ‘find the point with the smallest smallest distance to a set of points’) ordinary statistical methods are much, much more efficient than a neural network.
AI was introduced 50-odd years ago to model the way humans learn and extrapolate from the knowledge they glean. In a broad sense, AI is an attempt to model human learning and decision-making. In recent years AI techniques have been applied to a diverse number of activities, including trading. A number of trading software packages that traders can purchase today are based upon AI techniques.
Since its introduction in the 1950s, AI has split into two camps - expert systems and neural networks. Although these two terms are often used interchangeably, they are distinct in their approach to modeling human decision-making processes and, therefore, in their applicability to trading.
An expert system differs from a conventional computer program principally in that it solves problems by using knowledge that the system already contains rather than knowledge obtained through the implementation of a mathematical algorithm.
For example, a conventional computer program would use an algorithm to determine the annual volatility of say, the Euro, but an expert system would already have this information encoded in its knowledge base. This a priori information structure allows the expert system to successfully deal with problems for which no clear algorithmic solution exists.
