Why do I see "Trends" in Randomly Generated Data?

Quote from MAESTRO:

You would be surprised to know how close you are to one of our most reliable strategies. The only difference is that a Poker game has a fixed probabilities distribution, where the markets do not.
More...

A fixed probabilities distribution can created if the chart environment is layed out & set up consistently.
 
Quote from ProfLogic:

Ah, the master of "take comments out of context" rears his ugly head again. He who is incapable of doing his own research is first to cast doubt on the hard work of others.

I am a student of Einstein like anyone that follows the theories of any respected and brilliant theorist. I, unlike you, respect the research of others by trying to advance the search for knowledge not stonewall it like yourself.
Why is it you must start arguments everytime you enter a thread?
I also said I spoke in one of Einstein's classrooms in Zurich. I presented a paper on my research there a few years ago. I never said I taught there. Of course anyone could see how that could be misinterpreted . . . . NOT.
More...

aligning one's self with einstein is simply too much, even for an ego like mine...

pretty freaking funny!


surf

:D :D
 
Quote from marketsurfer:

aligning one's self with einstein is simply too much, even for an ego like mine...

pretty freaking funny!


surf

:D :D
More...

I respect the theories of Einstein and that is funny to you. Amazing.

You are a seemingly intelligent man that can't formulate or test original ideas for himself, so you ridicule others for their choice of higher psychological mentors, their determination in formulating new theories and their focus on discovering new problem solving formulas.

How sad for you.
 
So this has me wondering, how do you profit from random movements/data? :confused:

And where can I learn more about this subject? A lot of the links and .pdfs in this thread are confusing. I'm just finished up my first statistics course and this info is WAY beyond that.
 
Quote from Sniemiec:

So this has me wondering, how do you profit from random movements/data? :confused:

And where can I learn more about this subject? A lot of the links and .pdfs in this thread are confusing. I'm just finished up my first statistics course and this info is WAY beyond that.
More...

What you have to understand is that the markets aren't random in the classical sense of the term or to be more precise they are non-linear systems that exhibit both apparent random and probabilistic behavior. If they were random in the classical sense of the term, consistent long term profitability significantly different than their random bias would be impossible. It isn't impossible, just very, very difficult and requires thinking outside the box.

Below are a few links to get you started.

http://www.turtletrader.com/chaos.html

http://www.investopedia.com/terms/c/chaostheory.asp

http://www.imho.com/grae/chaos/chaos.html

http://www.zeuscat.com/andrew/chaos/


However you will find few published detailed how to explanations. The reason being that most who figure something out in terms of the markets tend to keep the details close to the vest for obvious reasons.

I know someone who monitors published academic material on non-linear systems and computational approaches in dealing with them. Academics will periodically focus on the markets as they are sexy and exciting, at least compared to predicting the impact of pollution on fishery stocks.

My contact told me about a researcher engaged in the financial area over some time. He would publish papers and lecture at academic conferences periodically over several years. Each presentation showed further progress in predicting future markets activity. About 3 months after a lecture that gave results significantly beyond what most here would believe possible, he stopped publishing and vanished from academia.

The really interesting thing is that shortly after that the on-line versions of his research disappeared from places like cite-seer and others. This is odd behavior for an academic, as your professional reputation is built in part by what you publish and who chooses to distribute it.

Did he discover something and not want to leave a trail of cookie crumbs for others to follow? It's impossible to know.

Jerry030
 
Jack,

Many thanks for an intelligent and though provoking post.

At times this forum can be a bit like observing celibate clerics giving marriage consoling advice. It may seem to make some sense but in the end you know they haven’t any domain experience.

Jerry030


Quote from jack hershey:

Very cool and the humor is there too. NYT and Nature trump the Lo to be thanked crew of the late 80's. What a list to trump.

It is very good to make successive discoveries of what is at hand; what is really at hand.

Jumping the threshold is a great thrust because it puts to rest so many games that just waller under this line in the performance sand.

Wavelets, at least, are more sensible that the brownians on which they ride. Parrondo, though, is only an exercise that is way away from the markets. The reasons are imbedded in the Rahmann (2002). For these reasons, well given, come the seeds to realy climb out and above the performance so far cited.

The players are not A and B or two other pairs that play under other defined rules. Changing the rules insitu is only an interseting thing to play with.

Part 3 scopes and bounds what Rahmann wants to consider. By Part 5 we find there is no tooling to deal with getting someplace.

So dtrader thanks for getting us back on the jerry030 track of dealing with significant performance above the Maestro and surfer buddies levels.

What is the pertinent substitute fro A and B?

What is the set of seeds that can be seen in Rahmann (2002) by inference?

The players in the making money game are two in number and each has a strategy that can be revealed by the rules they play by. What is the change in the consideration of rules that Rahmann infers?

When 666 quotes that I do not believe backtesting is valid, he makes the point Rahmann infers. Silvermotion can only concieve and percieve that he knows what is going on after the event (SCT's tooling doesn't work until afterwards). What these people believe is out of order and who cares.

The two players in the game are the trader (like jerry or cutten etc and not like Maestro or Surfer boys) and the MARKET.

The scope and bounds of each's rules has to be considered as Rahmann looked at the rules of Parrondo individual losing gamers A and B in order to see if he (Rahmann) could do a work around to get above prior mathematical levels of success (the Part 3, 4, 5 thing).

I changed the players to the trader and the MARKET. I injected the consideration that the rules used are those of the two players and not a common set of rules.

All these rules, it turns out, do not have to be applied all of the time. The poker calculator uses the same deck and one set of rules and just what is showing (using 1 million combinations) to only do a probability. It does not work for trading, as suggested, because the deck is not constant nor are the rules of the game that are in play at any time. Similarly, as jerry notes, to win at ruolette you do not play the odds unless you are the house; instead you play the machine by dealing with the fact that two velocities are involved and when the ball leaves the rim to bounce around, it does this at a specific location (different each cycle) and all the places where the ball gets captured are fixed in stone around a physical never changing arrangement. Do you know the order? Maybe not but college students do. We are not flipping coins. Put the coin away.

The use and application of rule subsets from two different sets of rules is the other part of the opportunity.

The comment: Where is the trading fractal?" does not refer to just a chart of a certain duration of bar. What the comment means is: what part of the MARKET model is in use at a given time? It also means; what part of the trader's strategy is being applied at a given time? Both players A and B can change what they are doing at any time. Backtesting if it were of any utility would have to do this according to the action-reaction of player A to B and B to A.

The MARKET pace-volatility matrix is dear to me because it embodies most of the considerations. constructing it monthly in equi-deci-divisions of pace crossed with 5 minute bar market volatility informs me of how the player called MARKET is playing. the rows and cloumns are normal distributions that are ONLY truncated (columns only) by the market's CAPACITY, a transparent truth always available.

So any trader can make his trading rules and pair subsets of these rules to the MARKET trading fractal at any time. As was ponted out people who measure the markets often have vested interests.

My vested interest is pool extraction from the viewpoint of a trader. I feel most people who trade want to make as much money as they can. I recommend looking at what is offered and taking it. That is what the pace-volatility matrix tells me. At any pace (immediately measured) the market offers a given volatility distribution and I know what the distribution mean median and mode are in that five minutes. And I also know (imdeiately measured) what the market IS doing volatilitywise. I also know all day long everyday what cell the market is operating in on this chart. So I frontrun the other participants in the market to take what is offered segment by segment.

Trends operating on the market fractal (this is not the 5 minute bar) is what researchers need to get around to. Randomly genrerated data is not good enough to work with since it is not reality. Neither as dtrader pointed out is the zero crossing game. Jerry points this out as well and in terms of his personal performance which is way beyond the threshold of the literature. All of the people who are classified as great trader you have never met" also pass being well above the threshold test. Surfer's buddies do not know any of these people and neither does surfer he tells us. He says he does not know ANYONE who is a player who is "out of the box".

Research, NYT or Nature articals do not hack it so far. The inherent academic and financiing sources biases do not support understanding what is going on or what is possible. Crayola 101 is humor to be sure, but it does allow a person to use a string to add up the day's offering. Four inefficiencies contribute to not reaping the offering. Today, in modern times, what jerry is talking about measuring and understanding is where the intellectual action could be focussed.

The trader only has to take care of being in the market and always being on the right side of the market. What I mean is just not possible to understand (as a passive onlooker) until you are practicing doing this.

A recent poster mentioned that he looks at a naked bar. Suppose he knew, always, the instantaneous market pace and volatility; his boat would be floated. He can't get out of this framing; he is IN the market fractal. All he has to do is collect the remaining infomation that explains the MARKET's playing and then USE his rules to extract the offer. The market is transparent as it unfolds.

Two losing players can win to some extent. Beyond this extent is what is important. Make the MARKET a winner (obey the market because it is always right) AND extract all that is offered by staying in the market AND on the right side of the market.

It looks like someone has to start funding these biased people and get them biased to what jerry suggests. Academics are cheap that's for sure.
More...
 
"Why do I see "Trends" in Randomly Generated Data?"

Why not? Since you can't even define a trend, what makes you think it can't be generated by both random and non-random processes? The fact is, they are: generated randomly and nonrandomly. This would be obvious if you gave any serious thought to what a trend is.

Your problem is that you conclude that because some trends occur randomly, then all trends must occur randomly.

This is like saying that because I look up into the sky and see a hot air balloon or a helium balloon, that all objects I see in the sky must be lighter-than-air objects. It's illogical to jump to that conclusion.

Some trends are random. Other trends aren't. The traders making the most money have figured out how to tell the difference.
 
You must log in or register to reply here.
Back
Top