The issue is getting the trader to establish an expectation (and thus a trading philosophy) of future price movements that are based on the highest probability for a range of price patterns to occur over a defined range of time. The statement: "Predicting future prices...." is somewhat incomplete. The statement: "Predicting a defined range of price behavior over a defined range of time...." might be a more optimal approach to the concept of making good price projections.
Assuming that we are only talking about directional trading and not non-directional trading (options strategies), I would agree with you. The moment a directional indicator "always makes sense," it stops being a mere indicator and it takes on the role of "perfect indicator" and I don't know of any directional based perfect indicators that meet the profit requirements of most traders.
What does "Price" really look like? And, when is "Price" instantiated? My observations over the years teaches me that the concept of "price" currently in use today, is less than optimal. In other words, my experience tells me that price is not a singular data point in the market. Even in a more practical sense, both buyer and seller are dealing with different versions of the truth about price, as one deals in the truth about the Bid, while the other deals in the truth about the Ask. However, that's just one reason why there is no genuine Price Singularity. Sure, there are "official prices" that the industry uses to demarcate one thing over another. However, from the trader's perspective and from the position have being in the business to net out profits each day, the concept of price is better seen as being continuous and not singular.
Some traders (especially the new traders) see price as some kind of singularity in time, or some kind of event on a chart. Research tells me that price is more like a clustering of 'apparent' random data points having a specific dimension (length, width, height), that travel in a certain direction and at a certain speed. Inside the price cluster, the existence of the Bid/Ask component at a particular location, has a certain probability for existing in that area of the cluster, across a certain span of time and with a certain rate of speed upon arrival and departure from that area of the cluster. Yet, regardless of the fact that a particular area of the cluster might be void of Bid/Ask "now," it has the potential (probability) to contain the presence of Bid/Ask at some point in the future. This is how I view the concept of price.
I agree. Which is why I find it more advantageous to view price as a clustering of data points with a specific shape, or form. I call it a Mode Shape, but that's a bit beyond the scope of this thread. If I force price into a singularity, then I really make it hard for me to capture it. But, if I allow price to expand across its normal range of frequencies and let it take on the shape that it so desires, then it leaves a much bigger signature on my radar, allowing me to better approximate its dimensions, direction and speed. If I know that information about "price," then I can place an order somewhere inside the cluster, knowing the probability that both Bid/Ask will will show-up (on schedule) in the right place and at the right time relative to my entry order.
This way, I'm not chasing a prediction. Rather, I am allowing the prediction to simply unfold through my entry and limit levels. Price is going to behave a certain way, whether I like it or not. I might as well learn to map out its behavioral patterns as optimally as possible, to take advantage of what it is going to do, with or without my participation in the market. If I can only see "price" as a singularity, then I'm trying to throw a dart while blindfolded at a donkey that moves as fast as a pure bread race horse. But, if I can see "price" as a continuous cluster (so to speak), then I can take a shotgun and shoot fish in barrel.
If I'm given the choice of either throwing darts at donkeys (while blind folded) that move as fast as race horses, or shooting fish in a barrel, I'll take the fish fry any day of the week.
But, in that case, what's the premise for the underlying lognormality? The underlying premise here is the word "random," and I think this is part of what separates highly successful traders from the rest of the pack.
When we typically say "random price movements," we are not usually talking about the same kind of chaos mode randomness to be expected in a physical system involving the lognormal distribution of say, molecules in a gaseous state. Price, unlike the molecules found in gas within a specific system, can only go Up or Down. There are no transverse paths for price to follow in a three-dimensional system. Molecules in a random gaseous state within a specific system do have transverse paths to follow. Even in a so-called "side-ways market" however, prices are constantly in a vertical flux (up or down only) while being driven through the dimension of Time. Therefore, we have to think very carefully about the kind of mathematical tools we bring to the market, in order to calculate the "probability" for specific price behavior and not necessarily a specific price point. Thus, the "randomness" which appears on the surface of "price" may indeed be lognormally discharged (distributed), but the structure of price is no more or less unpredictable because of it.
Think of light (from our sun) as price. Now, think of the physical components of light, photon Particles, traveling as waves. Price, has its complimentary particles, too. They are called Data Points within an OHLC stream. Well, they too, travel as Waves in many respects, but they do so as Clusters of Cycles. Within each Clustering of price action, you will find a Cycle of price action. So, you have to convert the Wave concept which encompasses the light's photon particles, to a Wave Cycle concept and then divide the Wave Cycles by the Time-Frames, or the size of the Bar of data in question to derive is Mode Shape. Once price is viewed in this way, its definition takes on a whole new meaning and the entire game becomes one of tracking Clusters (waves and cycles) instead of trying to track an elusive Price point (photon particles).
(again, well beyond this topic -- I'm only giving an example of how I differentiate your concept of price, from my concept of price)
Within the variable changes in the structure of price, there exists an underlying constant in its structural form. That constant, not the variable change that envelops it, is my primary concern as a trader. I'm concerned with penetrating the superficial chaos with analytical tools, in an attempt to locate the form's structural core. That core is what I'll target. I may not get all the pips available in the entire structure, but if I target only the core, I can reduce risk, increase accuracy and grow capital at an alarming rate of speed. But, if I focus only on the variability, I will never be aware that such variability has a rooted core.
While the frequency and timing of patterns do indeed vary, the underlying structure and form of price behavior remains fairly constant enough to benefit the observant trader (Harmonic traders already know this). The Matrix for this "core" is found in the fact that we impose "order and stability" on the markets by pushing its price points through an OHLC filter and by sub-dividing that primary filter across multiple dimensions of Time to arrive at Bars of data. Thus, forcing a template of structure on the market, which in turn reveals many things about the markets behavior - if we study the template carefully enough. Ironically, therefore, one of the biggest Indicators in existence, is the OHLC structure itself! Sitting right there under our noses all the time, is the biggest hint about what the market goes next.
From OHLC - we derive everything else we know about "price." OHLC is the grid that makes all other technical indicators possible. Yet, this very same grid has been overlooked and underutilized by many over the years. It contains more information than most traders realize. In fact, the number of useful indicators that can be driven off the grid are myriad and none of them have anything to do with the typical mathematical averaging of historical data points. But, that is not how most people learned to trade. Most people were taught to pick-up a chart and start plotting lines. Few people were ever taught to examine the raw OHLC data to look for patterns not involving mathematical averages. This is what gives my signature line (below) such frothiness and relevance. It takes a while for us to finally figure out what we don't know, or to realize that what we though we knew, is less than what is optimally available, if we simply looked at the data through a different lens.
Only true, if one has not identified the existence of something other than 40 years old indicators. As long as that remains unidentified, then it cannot be seen for what it is. There isn't just a class of new modern indicators out there - there's a super-massive class of new modern indicators out there and none of them are premised on the mathematical average. Now, their output can be averaged - sure. But, their baseline calculation and underling logic is definitely not average based. Harmonic Patterns are one example of this truth. Delta Patterns are another example. None of these are average based. Combining the two creates a powerful new approach to trading the markets. Most people have already heard of Harmonics already. However, few have heard of Deltas and for good reason - I created them.
A different [u]concept[/b] of price, as opposed to another averaging transformation of price.
I agree partially, here. I tend to think that all information that drives price, is incorporated into the price by the time you are able to respond to it as a trader for the most part. Even certain external news drivers are often times baked into the price action already. Extending this idea of using outside sources far enough, and you end up being more of a fundamental trader.
Where I disagree is in the notion that we cannot mathematically evaluate the market. Actually, you can evaluate the markets mathematically - we do it all the time, in fact. The question is whether or not we are doing it optimally. For the past 40+ years, the dominant methodology has been based around averaging. However, patterns in market behavior do exist - but only if one has the right set of tools to observe them. It is no different than studying the night sky with the right equipment from the optimal setting of an observatory and realizing that other galaxies do in fact exist and that they do in fact have identifiable structure, just like our own Milky Way. I am sure that from the view of someone sitting in an observatory somewhere inside the Andromeda Galaxy, if they don't have a powerful enough telescope to see sufficient detail, that they would conclude that our Milky Way galaxy was to chaotic to support carbon based life. Yet, the very structure provided by our own galaxy, is partly what makes life hear on earth possible.
I've trading equity options long enough to realize that Volume can be very misleading - depending on who is doing the buying and who is doing the selling and for what reason. Open Interest tells me little to nothing, unless it is juxtaposed (properly) against something of relative or equal valuation and how often can I make that determination with any degree of precision when trading options (hardly). And, what do Accumulated Ticks tell me when there is a Bear Run on a stock (little), or when a market maker and a large scale brokerage decides to dip the stock like a chocolate bunny (even less), or when that lady who once had coffee spilled on her six (6) weeks ago, decides to hold a press conference with her attorney while standing directly outside of the establishment where she was "injured," to announce her billion dollar lawsuit in front of a CNN camera (virtually none).
I agree with you that obsolete indicators are not an optimal way to trade. However, I don't for one minute believe that we have tapped the full potential of what's possible, in terms of generating new ideas for a more modern approach to optimizing the trade with indicators that don't readily make themselves known with a quick glance at a price chart, or the smoothing of a nearby (handi-dandi-indi) Stochastic.
The following 2 users say Thank You to JetTrader for this post:
First of all, thanks for your interesting post. I think we agree on many of xour statements, so I will not comment everything as it has a lot of merits on its own. I will just add a few points, where I think it worth adding something.
Regarding the concept of Mode Shape, I would be really interested, if you like to share. But it is certainly too heavy for this thread, so why don't you start an own thread on your concepts, I think it would be worth doing, and that there is a grateful audience here.
I am further interested in your concepts of Mode Shape.
We know that markets are most of the time at least similar to lognormal distributions. Obviously price has only two dimensions, if you compare it to molecules of an ideal gas. Sometimes markets change from lognormal distributions to other types showing fat tails.
I am very careful with applying cycle identifiers. Price action is mostly non-linear and there are few cycles. However, cycles do make sense, where introduced by human behavior. This applies to the daily cycle, with its regular expansion and contraction of volatility. It applies to the end of month phenomenon, which can be traded, and it certainly applies to (annual) seasonal cycles. Also news events can be considered as an external disturbance that translate into a dampened cycle. But otherwise, I do not believe in stable cycle periods.
I am very skeptical of harmonic trading. Although I use it, I believe that it mostly relies on self-fulfilling prophecy. Fractal dimensions do not typically take harmonic values, this is a fairy tale.
Why don't you explain us, what deltas are? Again, I would appreciate if you presented your concepts in a separate thread!
The following user says Thank You to Fat Tails for this post:
I wrote about as much as I can on what I do, over in the Introduce Yourself thread, as requested by BM when I initially registered. If you read that post, then you will understand why I won't be able to launch a thread with full details on Mode Shapes. But, I do appreciate the interest. I could probably try to cover some basics, but without connecting the dots on the lower level concepts, which I would never be able to flush out, it will be difficult at best to glean much from the higher level ideas. but, I might be able to try.
That's gets right back to the question of: What is price anyway. If you look at it as a decoupled non-linear event, then finding cycles will be difficult. But, if you see price as a coupled to extra-dimensional substrates between the OHLC regions, then you'll have no problem spotting the waves and their cyclical moments. See - this is the whole point about the standard definition of "price" and it is one more reason why I believe that in the development of my system, I stumbled upon a new definition with real substance. It is very difficult to explain without a lot of lower level territory being covered that I cannot expose. Also, graphical presentations always speak louder than words and I don't have anything like that on hand.
Bottom line - in order for your version of price to move from the Open to the High (as just one example of the coupling to an OHLC substrate) it must move there, and in the process it must also create a delta between the Open and the High. That delta has a value and it is what provides the substrate upon which "price" is connected to all other points within the OHLC in that bar - AND - in all subsequent bars behind it. In other words, I have found that there is a real connection between each data point within each bar of data - across a finite range of bars. That range is what provides the Mode Shape of the data. It is simply a way of expressing the "status" or "condition" of the OHLC data, as it passes through Time, from the current (real-time) bar to all bars behind it.
Like I said, this is not ADX easy language type stuff that can easily be placed into a few posts on a forum. It is years of research (almost ten years now) culminating in a different way to interpret market behavior over a specified time period. I basically take the "stuff" (empty space) between the Open, High, Low and Close, and then I treat it as though it is real data - actual market data. Well, (lol) when you do that, you get access to a whole different universe of possible paths for analysis of market behavior. Why am I able to do this? Well, because the Open cannot remain the Open for the entirety of the Bar's life-cycle. If that were to happen, then there would be no market to trade.
Therefore, I capitalize on the fact that your "price" MUST move. It has no choice. And, what's so eloquent about it, is the fact that it can ONLY move to a High, or to a Low. The Close is ubiquitous and only has relevance at the end of the Bar's life-cycle, where it is nothing more than the Open, simultaneously. Thus, that leaves only two data points to consider - the High and the Low. Now, that applies to the mechanics of just one (1) bar of data. But, when you couple that bar mathematically to the bar directly behind it, you discover that there are in fact not just four (4) transverse substrates (O:H, O:L, L:C and H:C), but there are in fact a total of twenty (20) additional substrate connections, between just two bars of data!
So, lets put that into perspective. The typical trader who can only see the traditional idea of price, across two (2) bars of data, say: bar  and bar , will claim that bar  is connected to bar  through only the  Open to the  Close. Thus, being able to establish only one (1) relationship between the first bar and the second bar of data. Conversely, with my idea of price, those same two (2) bars of data now yield sixteen (16) direct relationships, plus four (4) additional indirect relationships, for a total of twenty (20) mathematically derived relationships that can be uses to take the pulse of the market.
That's an increase in probe activity of 20:1. A 2,000% increase in the level of market detail at any given time. I can probe the market with 2,000% more sensitivity than anyone I know. That's got to provide an edge. Increase the number of bars and do the math.
I'm mapping the very DNA of the market with this approach. It is one of the reason why I am typically in positions, before the crowd gets there and out of that same position before the crowd leaves. First in - First out.
This issue confuses a lot of people. The "harmonics" being spoken of, relates to the use of Fibonacci derived sequences across multiple time-frames where specific retracement levels have been struck (within a margin of error) in a specific order. The word "harmonic" got accepted to mean something like "resonance," which is not quite what the originators had in mind, I don't think. The harmony simply comes from the synergistic effect of multiple pattern completions (under specific rules) within multiple time-frames, or multiple bars of data. The more synergy (harmony) at depth the market produces, typically the higher the probability, or chance for the market to move one way or another depending on the pattern type.
The issue with Harmonics, is that often times you will see smaller TFs producing consecutive bullish patterns t completion (for example) inside a larger TF that is producing a bearish pattern to completion. The inverse is also true. In those scenarios, when the smaller TF "harmonics" fail, people tend to throw-up their hands and say: Well, see - they don't work. But, because they don't have their Harmonic matrix set-up correctly, spanning across all time frames, they never saw that big Whale on the larger time-frame that was placing a lot of pressure in the opposite direction, causing the smaller TFs to break-down.
The market does respect the Fibonacci Ratios, but if one is not set-up correctly to see the appropriate harmonic synergies taking place (and they are happening all the time in the market) and fixates for too long a period of time on a singular time-frame, then one will typically conclude that they don't work, or are less optimal than they really are. Harmonics have to be set-up correctly; the appropriate ratios to use have to be understood thoroughly and the trader has to be patient enough to allow enough of the pattern to complete before taking trades.
I just did, but I can only go so far with what I'm willing to release. Check out my post under Introduce Yourself. I Delta is a very simple concept. The distance and magnitude between the Open and the Low, is a simple delta. Yet, there are many other deltas just like that one that nobody ever pays attention to. Those Deltas, when taken in the aggregate and across multiple time-frames, yield Magnitudes. And, these Magnitude values couple the real market values (price points) together along the four (4) basic substrates mentioned above, which in turn helps to produce the overall Mode Shape of the Price Cluster. The Cluster contains your definition of "price" plus the values that I derive from the coupling of the substrates together to form the Mode Shape. Your idea of "price" always travels through Time, with my Cluster of derived "prices" and they travel in the form of a Wave of Cycles, producing a probability matrix that tells me when the Bid/Ask stands the highest probability for existing in some region of the Cluster, at a specific time, with a certain speed and for a specific duration.
I apologize, but this is just one (1) concept within a system that contains 127 such concepts spanning ten (10) worth of research/work. There is simply no way I could do this online, even if I wanted to. Can't be done. However, I obviously understand and appreciate the interest and I hope that helps at least a little bit. I know it is not often that some thing like this comes along - or at least with someone willing to talk to some degree about it. Anyway, just use what you can as inspiration for new ideas of your own. That's what I hope people go out and do - ignite more innovation in this business.
Please register on futures.io to view futures trading content such as post attachment(s), image(s), and screenshot(s).
The following 2 users say Thank You to JetTrader for this post:
This would be like a neural net that can be exploited by a genetic algorithm?
I agree that Fibonacci derived sequences work well as an overlay of multiple patterns and timeframes. The indicator on the ES chart for this morning detects 540 harmonic levels and then extracts confluences zones by adding up the single conditional probabilities for a price reversal.
The following user says Thank You to Fat Tails for this post:
Dudes! so much ragging on indicators! Indicators get a bad wrap because most people just get bag of indicators and start slapping them on to charts, look at them for 10 minutes, and determine they don't work. What you need to do is find a trading concept you believe in, then find indicators that support that concept, then learn the specific indicators well to find how to exploit their value. I do pretty well trading 2 very basic indicators, but I don't use them in the same way that the masses use them. It's generally not the indicators that do not work, it is their incorrect application.
The following 4 users say Thank You to monpere for this post:
Favorite Futures: Index,Currency and Energy futures
Posts: 1,701 since Jan 2010
Thanks: 501 given,
I read and followed your comments with great interest. Unfortunately cannot relate to it as you are not prepared to share with others this fantastic "system" or revolutionary way to trade. I can understand if this is a trade secret and abide by your reluctancy to reveal it to the world, but then why did you bring it to the forum, as at the end of the day it is a futile discussion.
The following 5 users say Thank You to perryg for this post: