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

[mynd66]

As a relatively new trader I've been carefully reading this thread and absorbing all of the insight provided by everyone. But I'd like to respond even with my limited knowledge.
If the markets are random as some of you suggest then how can you explain this... if every trade ever placed by someone to go long or short had a 50/50 chance of moving in a certain direction then everyone’s win/loss ratio would be just as random right? If you made a chart of win/loss ratios of everyones balance sheet would they look like the chart does?

How bout this. When you look at the randomly generated chart can you imagine if the history of the DOW was just a tiny little upward section in that chart? If history shows a market that has gone up for reasons more likely than not such as progress than I think it will take more than randomness to push us back down to zero. As long as the majority of people want improvement in life the market will have upward bias. It couldn't possibly drop below where it started (well extinction), humankind will always err on the side of progress. At times there is regress and a downward bias. If the markets did not improve then we would not have the technology we have today. Markets tend incline (more often then not) and decline (less often then not). So can tomorrow be the start of a random gradual 12,400 point drop in the DOW? Ok I may not really be making sense here but can someone create a chart that generates random data where it starts at zero and moves up on the first tick yet never falls to zero again yet keeps generating random up/down ticks? What are the odds that it will get to 12,000 without falling to zero once?

There are human caused reasons as to why the market goes up and down. If it has a particular bias on a large scale then it must also on a micro scale. If downside is limited and the upside will inevitably fall back sooner or later than is it really impossible to develop an edge? If the market goes up or down, unless it will drop to zero or rise to infinite the time till it reverses gradually becomes less and less. Why can't we keep narrowing this down to come up with some sort of edge. Even if one happens to be wrong, his actions will have been a non random influence in the market.
I’m not sure where I’m going with this… let me know if I’m crazy but its late I’m tired and I just wanted to participate in this discussion(gotta get my posts up, spend too much time reading). It has been a good learning experience.

 

[heypa]

If the markets are random how come they seem to drop at a faster rate than they rise. Perhaps it is because I am seeing a pattern that doesn't exist because of my flawed ability to visualise.

 

[ProfLogic]

Quote from marketsurfer:

human psych is not predictable in the micro which is all that is relevant in the second prior to decision. the above statement is irrelevant,misleading, and naive' it is often used but rarely questioned in the TA world by the true believers.

everyone is trying to game the system, making the these trend ideas ancient legend.

read "equity markets in action" for market facts not myth.

surf

I need to set up a time when you can have a heart-to-heart with my computer because there doesn't seem to be any hesitation the second prior to it systematically triggering & executing a trade. Maybe it could learn something from your vast amount of research on this phenomenon.
It bothers me that my computer is creating profit irrelevantly and being naive. How gosh.

 

[MAESTRO]

This might help to understand where I am coming from. I wrote this short article for one of my university lectures.

 

[Thunderdog]

Quote from Rahula:

The thing about surf that I don't understand is that he constantly says there's no such thing as trends (patterns) - he then uses that as a justification for countertrend and mean reversion trading.

If there is no such thing as a trend then why would you countertrend trade a trend that doesn't exist?

And if markets are overwhelmingly random then mean reversion is a meaningless strategy - is it not? If the mean is a random number and deviations from the mean are random, than what 'edge' can mean reversion strategies offer?

Quote from dtrader98:

I think that one very important and salient point you are missing is that although markets have an unpredictability component in the short run (variation about the mean), the mean itself is not zero. Therefore, the long term bias is not random, rather the central tendency favors those who buy and hold long term (similar to a dc offset in electronic systems). If it were zero, then there would be no hope for anyone, including buy and hold over the long run. The fact that the mean has a slightly positive offset over the long run is what allows those who buy and hold to benefit over the long run.
This in itself is a mathematical edge that anyone can benefit from with much less stress (less risk, albeit lower reward). The more diversified the portfolio, the more valid the mean offset becomes (i.e. indexes vs individual stocks, non-correlated assets vs. correlated).

It's also one reason why people who look at long term charts can not fathom the chart to be one large random phenomena (according to their expectations of what random should look like).

The long term charts tendency to drift upward can be expected, even using random models, if the entire model is correctly applied (including a non-zero mean).

A very good exchange.

Perhaps I am just a hayseed, but I am inclined to agree with Rahula, based solely on personal observation and experience. However, dtrader98 has certainly given us some food for thought in this thread. One thing I can't quite grasp is the argument that the short term is necessarily random (the "variation around the mean"). If that is indeed so, then how is it that there is a fairly large number of successful short-term traders who have been successful for years and who continue to be so? Are they all being fooled as a group by randomness, year after year after year? Eventually, when an increasing number of monkeys begins to type all of Shakespeare's works, we may have to conclude that perhaps not all of them have done so purely on the basis of chance alone. (Not a perfect analogy insofar as the numbers are concerned, but you get the point.)

 

[ProfLogic]

Quote from MAESTRO:

This might help to understand where I am coming from. I wrote this short article for one of my university lectures.

I speak only for myself but I fully comprehend where you are coming from. It is simply what all financial and economic educational environments teach. What you believe is what you were taught and then you created a way to elaborate on that. It tells me you were a good student and then progressed your education to ITS next step. Well done!

We are different because I approached every math course I took with an understanding that it couldn't be the be definitive. Discovery was important to me, not from a standpoint of notoriety but from a standpoint of progressing traditional thinking. The first Logic course I took in community college ignited my thurst for information. The study group I led was responsible for finding 38 new or shorter answers for test questions in the instructors guide. The author was flabbergasted at what we discovered. For me this was just the beginning. I then created a solid career for myself solving problems that run-of-the-mill corporate types said was impossible to solve.

I look at every problem in this market environment as a challenge you look at every problem as already having a definitive answer.

 

[marketsurfer]

Quote from ProfLogic:

I need to set up a time when you can have a heart-to-heart with my computer because there doesn't seem to be any hesitation the second prior to it systematically triggering & executing a trade. Maybe it could learn something from your vast amount of research on this phenomenon.
It bothers me that my computer is creating profit irrelevantly and being naive. How gosh.

your trading skill/luck is not the issue here. i have no reason not to believe your success, you have been very consistent and direct over a non trivial amount of time.

with that said, i'll return to the issue at hand. What I take as a given, apparently, is not understood, so I will attempt to explain--- this human psych being predictable therefore TA works since human psyche is reflected in price, which is illustrated by TA. right?

It's time to knock down this myth---- the first and most obvious reason the above makes no sense is--- the market is NOT driven by the masses, the masses are BUY AND HOLD only--the masses are NOT traders---nor DECISION makers---

the market is driven primarily by large sums of capital controlled by a relatively few people/teams who are constantly trying to GAME the system to create edge--- no one knows what these guys are thinking or why they are thinking it---however, the desire to game the system circumvents any hope of edge created by TA.

hope this makes sense.

surf

 

[Corey]

Quote from Rahula:


Now if the markets are random and the market closed up,up,up,up,up,up,up,up,up,up,up,up 12 days in a row - most people would likely be talking about it as a n-sigma event or a "black swan" even though it is just as unique and random an outcome as a up, up, down, up, down, down, down, up, down, up, down, down 12 day series of closes.

First, I apologize for not reading the entire thread. If this has been touched upon already ... then shoot me.

I only skip ahead because this quote is just plain wrong. An event configuration may be UNIQUE, but it doesn't make it rare. What makes the up,up,up,up,...,up market rare? It is the probability of getting that many similar labels in a row.

Imagine a graph. The x-axis is number of throws of a coin. With each heads, we increment our count by 1. With each tails, we subtract 1. With a 50/50 probability of heads or tails, we would expect (and be correct), to assume our mean is 0. Thus, over the long run, we would expect to see fluctuations around 0. Your first series strays so far from 0, it could be considered an 7-sigma event (assuming this is a binomial distribution where n=12 and p=0.5). Your second series, AT MOST, strays from the mean by 2 (either +2 or -2). That is entirely reasonable to expect...and well within two standard deviations.

What you are mistaken in assuming is that just because any given configuration of heads and tails is equally unique for n trials (so while saying "two heads and one tail" is more likely than "three heads", the configuration "hth" is just as likely as "hhh"), what is important is the likelihood of seeing one label a number of more times than the other ... so much so that we can call it a 'significant' event ... because at the end of the day, I don't care what steps you took to get there, I care that you got there.

Thus, 'reversion to the mean' makes sense. The real question is ... in an ever changing market ... where is the mean?

Mind you, I know this is a discussion about trend -- which initially seems entirely different that basic probability. This would be true ... but in my mind, the probability distributions are not gaussian, but rather binomial, with a skew in a given direction, which over time defines the trend. In my opinion, this skew in the overall market is defined by the prevailing vote as to economic conditions. Intra-day, in less liquid markets, the skew can last as short as a few seconds while temporarily more buyers exist than sellers.

A skew can be created from self-fulfilling chart patterns -- especially in illiquid markets. If a majority of traders believe the market will trade higher, then a skew has been created. No longer is the probability of an uptick and a downtick equal.

Support and resistance lines do the same thing -- at that price level, the skew changes.

Trends themselves can change the skew as traders become more confident in the trend's existence.



Be advised, the above is all strictly opinion from empirical study...except for the basic probability part...

-C

 

[Thunderdog]

Quote from marketsurfer:

...the market is driven primarily by large sums of capital controlled by a relatively few people/teams who are constantly trying to GAME the system to create edge--- no one knows what these guys are thinking or why they are thinking it---however, the desire to game the system circumvents any hope of edge created by TA.

You continue to speak in absolutes. Was it not one of your heroes, Larry Williams, who said that giants leave footprints, or some such? As you know, I have absolutely no respect for your Mr. Williams, but are you now contradicting him?

Happy trails.

 

[fluttrader]

Quote from Thunderdog:

You continue to speak in absolutes. Was it not one of your heroes, Larry Williams, who said that giants leave footprints, or some such? As you know, I have absolutely no respect for your Mr. Williams, but are you now contradicting him?

Happy trails.

Is Marketsurfer Larry Williams ?