In general the higher the resolution of data (in terms of time frame, the higher the period) the more "predictable" the market becomes — not only in terms of correct guesses of price direction in certain intervals, but also in terms of at least standard metrics used for assessing the performance (like pf, win/loss ratio, avg. winner/avg. loser and so on). There's a clear explanation to this phenomenon: just consider which market participants work on which resolutions, and what they do, and the picture becomes apparent.

Alex,

The topic is interesting and unsolved and has the right to be classified as "old as the world".

"Market is where people come to exchange something for money. It cannot be chaotic, random walk or predictable."

I feel very comfortable with the first sentence.

The second sentence look to me too categorical. If it is a point of view based on a belief, then it makes sense to notice that we have means to study markets and with certain definitions of chaotic, random walk, or predictable check to which degree it is so or not so.

If the second sentence is a deduction (conclusion) from the first one, then it is difficult to agree in full by the following reason. We know systems, much simpler than markets, where their deterministic organization does not exclude their chaotic, random walk, or predictable/unpredictable behavior.

Consider a vessel with atoms of an ideal gas interacting in accordance with Newton's laws of mechanics. While trajectories obey deterministic laws, we should recognize that evaluation of a particular trajectory becomes very sensitive to initial state of the system. This exposes certain chaotic properties. While such atoms move deterministically, this does not exclude that certain distribution on velocities takes place and this allows application of probability laws and limit theorems. The latter circumstance makes certain properties such as pressure under constant temperature and volume of the vessel quite predictable. We get a situation where certain properties are insensitive to individual trajectories. If we get all spectrum of such properties from a simple (not very simple) system, then how can we exclude such effect for much more complicated systems in addition involving human beings.

We are left, in my opinion, with definitions, means, and theories (models), where we can test their correspondence to different market faces. However, as you notice this also may be "useless for actual trading".

Best Regards,

Valerii

Valerii, can a coffee shop be chaotic? Can a butcher shop be predictable or random walking? Same here. These terms refer to price time series, not the marketplace itself.

Overall I tend to agree that we are left here with definitions and opinions, but without them the discussion resembles an argumentation between a blind and a deaf.

Alex,

The terms "chaotic", "random", "predictable" refer to a behavior of a dynamical system.

Chaos theory studies the behavior of dynamical systems. Such systems can be deterministic without involving random elements and still remain unpredictable. Often, the unpredictability in such systems follows from their high sensitivity to initial conditions, where small changes in the initial conditions cause big changes in the evolution of the system.

Probability theory, statistics, and theory of stochastic process study static and dynamical random systems. While many associate the term "random" with unpredictability some properties in such systems are predictable. The form of prediction for a quantity combines formulation of the range and probability with which the quantity is expected to be found within the range. Under certain limit cases such random systems behave like deterministic and become very predictable.

Depending on which properties one wants to consider a coffee shop it can be found chaotic, random, deterministic, predictable, or unpredictable dynamical system. The word "dynamical" simply indicates that one is interested with what happens with the coffee shop in time.

I agree with you that members in a discussion can come to more productive conclusions, if they speak the same language. By the latter I mean clarification of definitions so that they are known to all participants. You may be noticed from a few other discussion, when one jumps on using the terms such as "random", I ask to give, if not a definition then at least indicate important features of what he or she wants to consider.

Clarification of the definitions is also important because it usually leads to means and tools, which can be selected and applied for studying a subject. Then, we can experiment, test, check, accept or reject our initial assumptions.

For example, "random walk" can be a mathematical term, where many properties are known. Such a random walk (symmetrical or non-symmetrical) is viewed by many as something simple. In fact, it has several not-obvious and, even, shocking not-obvious properties. But if one indicates in the behavior of a dynamical system some similarity with such rigorously defined random walk, then he or she can also apply the means, tools, and already available knowledge to expect certain behavior.

If the same term is used as equivalent of a random process without specification of details, then it becomes less attractive, less interesting, less constructive.

Best Regards,

Valerii

Alex: "Market is where people come to exchange something for money."

Valerii: "I feel very comfortable with the first sentence."

Alex: "... can a coffee shop be chaotic?"

Rereading the first sentence and comparing it with the last confirms for me that I prematurely agreed because inexactly understood the accent. It is not so that I argue that "market" can be considered as a "place" where people gather to trade. However, this is also not with what I agreed so comfortably. My "quick" reading created an impression for me that the phrase is close to what I wrote in my publications. Those sentences were

"Market is nothing more than people trading something" and "A market is people and programs trading something" and "Market is a result of cooperation of modern technology and human being consciousness governed by partly unknown laws of nature".

These sentences put an accent on the process, where a physical place conducting the process is, even, not considered. They consider the market not as a physical place but as trading, a process involving human being consciousness. I use the term 'market' in those sentences in order to denote this process and not the place.

Best Regards,

Valerii

Here's a good read re: EMH from Woody Brock http://www.sedinc.com/index.php?eID=tx_nawsecuredl&u=0&file=fileadmin/user_upload/reports/March_2014/March_2014_SED_PROFILE.pdf&t=1420140638&hash=3217ec6f3fce0f080a8d649b9ef66697

+1 Valerii, all your posts above are excellent and well motivated, as usual.

Alex, you first post is also excellent except for this sentence : "It cannot be chaotic, random walk or predictable."

This sentence is clearly an opinion and is highly debatable as Valerii just explained. It even surprised me as you also state the following : "Market is where people come to exchange something for money".

If you are dealing with variations of a system explained by real people's actions, then believe me, it can get quickly chaotic or predictable, as people can definitely become crazy as hell thanks to abnormal reactions and attitudes.

Can a coffee shop become chaotic ? Yes, if a couple of customers start a fist fight, you can end up with a general brawl and maybe a reaction of panic as the owner of the shop pulls a weapon to protect his property. So yes, any system based upon human behaviours can be driven into a "chaotic" state as soon as panic or emotions kick in.

Likewise, if I pay 100.000€ to anyone who digs a 1 cubic meter hole in my garden, can you predict what will happen ? I guess so : my garden will be a massive hole and I would be much poorer after I paid all that money.

The tricky part is the following : to model chaos is difficult (you can model the deterministic chaos if you increase the number of parameters, but as of today, you cannot model stochastic chaos.... maybe later, but today, nope, it is beyond our reach). Likewise, transition phasis between chaotic and "normal" stages are also very difficult to describe.

So I end up with the same conclusions as yours : the all encompassing model that describe a system throughout its "quiet" and chaotic stages is yet to be found. Some models will have great metrics in certain market conditions, but poor in other. All you have to do is to know when to turn off/on a given model (which is also tricky). A model that fail in given conditions is not necessarily a bad model, but it would be a poor trader that doesn't know the limitation of his model.

Valerie defines the market, to how I trade and think of the market, perfectly. Great post.