1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Thoughts on Autodealing's Future

Discussion in 'General Forex Talk' started by euroforex, May 22, 2010.

  1. euroforex

    euroforex Private

    Nov 21, 2008
    Likes Received:
    Any discretionary trading strategy can be regarded as a "signal processing analysis" that uses external information to measure probable results. The pertinent essentials of this external information are incorporated in a defined methodology that induces structure in historical prices, pip ranges and trend formations. The hypothesis of any efficient trading strategy is that it must instantly and correctly adjust to reflect new information. The complexity of these robotic analysis models/algorithms can encompass nonlinear calculations and/or focus on the effect of trend following and value investing behavior assumptions that market price clearing is a "bifurcation structure" and such conditions of market dynamics are constantly chaotic. The core challenge to this task is that you incorporate both economic and pragmatic data, all based on the basic principles of the Forex market inter banking trading. But knowing how to convert the "a priori" of possibilities into accurate speculation is the way to minimize unavoidable trade losses.

    As Upton Sinclair said, "It's hard to get a man to understand something if his living depends on him not understanding it."

    With computer information processing there are two scientific systemic "top down" and "bottom up" designs in computer software development. The one that is most prevalent in Forex trading is based upon neuroscience and psychology as a bottom up design, for example your attention is drawn to a flower in a field. Your attention was not contingent upon knowledge of the flower but the outside stimulus, contrasted by top-down which means you had an idea before of the flower you wanted to see, making it salient upon recognition. Computer modeling for robotic/autodealing in Forex is based upon parallel computing that allows for faster results (Dr. Duka, founder of Dukascopy who explains the "ping" effect for nano second trading) in the process of problem solving with simultaneous coordination with parallel random access machines (PRAMS).

    Consequently the workload of a clearinghouse computer server (compressed transmission collection into what is called a "processor farm" that passes a myriad of topologies) the pipeline parallelism benefits with added processors, that apply implicit parallelism and explicit parallelism with the techniques of performance based on load balancing. Amazingly, the outcome, a mirror of our own complex adaptive system to life, appears as a pattern, the biological inheritance to see patterns (trends) and construct schemata in our minds.

    With the Internet, Forex trading is electronically executed by what is called "straight through processing (STP). The entire trade lifecycle with STP minimizes settlement risk and clearing simultaneously. A fully realized STP will provide asset managers, broker/dealers, custodians and other financial services players a reduction in settlement risk and with the emergence of "business process interoperability (BPI).

    For the Forex trader to interface with the clearinghouse exchange for information an application programming interface or source code interface is developed for coherent interface consisting of several classes or several sets of related functions or procedures and/or a single entry point such as a method, function or procedure. ECMAScript is one program that consists of sets of functions, procedures and variables and data structures that relieve the programmer from needing to know how the functions of the API relate to lower levels of abstraction.

    What is the point of all this? What is encompassed in an robotic/autodealing system, in part, the variety of computer generated techniques called, "predictive analytics" that takes statistics and data mining and processes them to make predictions for future events. Predictive analysis is used in making customer buying decisions and credit scoring.

    There are two commonly used forms of dynamic modeling path structures to improve forecasts; the "autoregressive model" (AR) and "moving average" (MA). The Box-Jenkins methodology (1976) developed by George Box and G. M. Jenkins become the cornerstone of stationary analysis in time series with their autoregressive integrated moving average (AIMA) used to describe non stationary series; producing a pronounced trend without a constant long-run mean or variance.

    In recent years, the time series model has become more sophisticated used for financial time series. These are the autoregressive conditional heteroskedasticity (ARCH) and generalized autogressive conditional heteroskedasticity (GARHC) that are used to understand inter-relationships among economic variables represented by systems of equations using "vector autoregression" (VAR) and structural VAR models.

    Ironically this all goes back to the BASIC language problem as the "fixed point" program that Steve Jobs wanted to make into a "floating point" that Gates already had - revolutionary step in development of replicating human cognitive interactive relationship with a computer that today is mathematically myopic with all our computer technological advancements, but still based on this own binary DNA code as the archetype.

    Thus identifying what Jobs so desperately wanted, the "floating point" with robotic/auto trading software has always been at the core of the issue of getting an edge on the market with leading indicators and oscillators, since these are Lagging Indicators to economic Leading Indicators. Indicators, though the latter is considered market sentiment and can change with geopolitical policies and economics news.

    Searching for the floating point - incorporated by the right combination of real time data is key to a robust autodealing program.

    For example, Google, which revolutionized the Internet search engine, is fast becoming obsolete as it is based on counting links back to authoritative Web pages. Dmitry Skyarov (wrote "The Advanced eBook Processor (AEBR) software and was arrested in the US in 2001 because the FBI determined his software violated copyright protection of the Adobe eBook Reader - charges were later dropped) states, "...it's a hypertext corpus of unfathomable intricacy, and it's expanding faster than a flat universe in a cosmologically significant vacuum energy density." It's invisible to the naked eye, but its there, a string of binary code, submerging us in a sea of information that isn't necessarily pertinent to what we need to know.

    The Internet browser, Stumbled Upon, mimics Skyarov's theory, allowing you to set your preferences for random searches of information you might not have gotten if you Google the subject matter. Could a Forex autodealing program do the same?

    This is the same case for Forex robotic/autodealing software. In this context, the randomness of pricing chaos based on the market sentiment is what technical analysis tries to tackle with a complex system of moving averages, futures and open orders all tugging at each other for price control, simultaneously. It is the hierarchy of global economics that must be incorporated.

    Sklyarov writes, "From here, you simply fire up things by banging out an orthonormal basis of eigenspace for each and obtaining the eigenvectors for the matrices in question. That's it."

    Thus we are moving toward the qubits computer memory processing, with the binary matrix replaced by unitary matrix based on quantum circuitry.

    The evolution of autodealing began in the 1990s - bank-to-bank market systems developed by EBS and Reuters. Today, these systems provide electronic brokering for foreign exchange transactions, allowing member banks to trade various currency crosses with one another by way of electronically posting bids and offers an striking at various price levels.

    Latency issues still arise: technological improvements have provided clients with the ability to access dealer's liquidity through a variety of channels, such as electronic communication networks (ECNs), technology vendors, bank graphical user interfaces (GUIs) and application program interfaces (API). But there is a growing challenge as the banking industry's infrastructures and the client's trading infrastructure must remain cohesive both at the interbanking and retail traders levels. Because the foreign exchange market is global, dealers that provide steaming liquidity need to consider co-locating intelligent and dynamic pricing engines alongside their main sources of liquidity and take into consideration the rate sources that reflect the strongest volumes in a given trading location and currency pair.

    The self-regulatory nature Forex and global span of its entire suite of products will continue to ensure that Forex is on the forefront of innovation within the world's capital markets. Autodealing will likely continue to be an important and vital part of robust innovation.

    Thus it is imperative that market participants be cognizant of the challenges and cutting edge opportunities that autodealing present for all parties. Nothing short of implementing artificial intelligence (AI), it would seem.


    To wit: I ask myself how does one deal with the "fight for liquidity" that becomes the "fight for milliseconds" on the FX topography during real-time trading sessions as a means of producing a highly efficient product that would revolutionize this classical FX model of the three tiered system marketplace that is more centralized than decentralized. Of course, MetaTrader/Expert Advisor et al are currently the means to the end; but these (as I have tested many - still carry a co-dependent laggard signal base function because of their archetypal limited language format that cannot identify potential risks in time. And checking risk before execution is what makes or breaks a trade. Moreover, MetaTrader, for example, developed with proprietary algorithms, can capitalize on the aggregate view of the FX market; but if all FX traders are working off of the same programming hybrid, the programming language differentiate is compromised. Compounding the issue is the fact that the growth in teh number of FX autodealing venues is leading to an increase in liquidity fragmentation. If pre-built algorithms do no achieve the "edge" for retail traders to compete with Tier One - EBS and Reuter D2 API interbanking exchange. As well, the competitive edge is diluted back to a "zero-sum" outcome for the majority of retail traders. This is not to say there are not credible autodealing programs already used with favorable success; but the accentuated fact that the market is continuously evolving makes for building and customizing a constant endeavor as a key strategy in capitalizing on the FX market.

    What needs to be engineered is a "cross-diagramming expertise" with applied cognitive psychology to the foreign exchange market that obtains "the single version of truth by analyzing the supply chain data to be proactively managed as a decisive and reliable leading indicator.

    What I am formulating at this time is what is known as "enterprise application integration"(EAI) coupled with message oriented middleware (e.g. Currenex, Lava, Hotspot, etc.) that empowers the retail trader with a top-down format of information streaming, subjugating the interbanking "ping" data feed that outwits the three tiered system. By delivering queries for data source with an "enterprise information integration" (EII) that taps into real-time operational time sensitive activities, price, volume, sentiment and indices data movements are culled out by strength and weakness that overcome historical data warehousing for real-time objectives. As well, what has been noted is that non-correlated liquidity providers will have a significant effect on the spread size and also elevate the lack of liquidity problem.

    By embedding EAI and EII to organized multiple versions of the "truth" scenario the proprietary "ping-time factor" signal becomes the power of choice for the spot trader by advanced liquidity information including contra-side orders in millisecond execution there is the issue of spread fluctuations during buy/sell transactions). This is attained by developing an info set comprised of "end user data clusters" that has an intra-web system sensor data matrix. Using an "amoeba" 3D graphic depiction (augmented reality would be nice) the nature of currency pair movements are illustrated by a liquidity "pod" distinctions for computer monitor visualization - signaling the dominate currency that can then be traded accordingly to the respective currency pairs. It would completely eliminate the current trading terminal "chart".

    The liquidity pool - decentralized in this case - can be performed with the least amount of latency by designing an aggregate liquidity pool (ALP - POD) to gain a unified view of the interbanking FX exchange.

    For example: The top 3 bids are on EUR/USD from EBS, whereas the next best order bid could be on FXAll Accelor (bank provided liquidity pools are Hotspot; TradingScreen; Currenex; 360T, Integral FX Inside).

    With this progressive model an aggregate view combines this into one Super Book. The autodealing - with manual override - trading platform indicates order placement messages - such as EBS entry or hesitation - based upon venue that it can understand with an integration adaptor layer that is able to translate between events and market data, institutional and interbanking orders. But here's the key component: Because the aggregation becomes more complex as more liquidity pools are added into the mix, everytime one of the pools changes the aggregation algorithm detects this - determining instantly if the change impacts the market view for a particular currency that can either become dominate or subordinate overall.

    The "event" based rule algorithm to calculate the value at risk threshold:

    EUR/GBP moving average - velocity is analyzed (up/down/sideways)
    EUR/GBP moving average is greater than the required Open/Close spread and required liquidity amount (EBS) of EUR is available versus GBP, then;
    Request BUY EUR/GBP if spread of EUR/GBP and EUR/GBP moving average is less than required spread, then;
    Cancel Trade or
    If trade is not completed in 2 seconds, then;
    Cancel Trade.

    This is a basic building block, with actions to auto-hedge the positions through trading that yields advantages in cross-asset trading that can be easily implemented since messages from equities, derivatives, interest rates, and fixed income systems can also be treated as "events" for cross-border trading with multiple instruments.

    Inside Trading
    Shock events or "prefect storms" or "Fat Finger" events (as we saw with the DOW on May 6, 2010 -when it dropped 998 points in less than fifteen minutes) highlights the need for revision.

    At the time of my research in 2008:
    On September 29th we saw a 778 point drop in the DOW. Consequently the major currency pairs reacted in comparative volatility. The following day, the EURUSD posted its largest one day decline since January 1999, dropping 574 pips before the New York FX session closed.

    These are “shock events” that pre-market sentiment fundamentals and technical chart analysis and posted daily economic reports fail the FX trader because the behind the scenes interbanking liquidity data feeds didn’t show the “perfect storm” scenario. There were three combined elements that fostered this outcome: Funding costs, extended short-dollar positions and London fixing buy orders.

    The non-transparent Over-the-Counter derivative market and asset-liability in the major banking centers securitization based their currency borrowing on the modeling strategy “earn the spread” but if you can’t fund the short-term (Eurozone’s greenback supply that is being repatriated to sell out of foreign central banks through the Open Market Operation) there isn’t any long term. The weakest part of the mathematical model is in the assumptions used.

    The Eurozone became vulnerable to the New York FX session; where market players – knowing the short-dollar positions would compromise London market player’s long positions on EURUSD, and overwhelmingly shorted the euro given the financial leverage based on the net capital rule that has allowed for capital leverage from 12:1 to 40:1 in the past years – the excessive weakness has created a paradoxical situation by increasing demand for the USD.

    Oddly, the US default mortgage fueled (8 out of 10 Americans are currently in delinquency payments) credit-deficit crunch is having a contrarian ripple effect on the Eurozone as the financial stresses are historically similar to what happened to South Africa in 1985. South Africa banks had significant short-term borrowing in the greenback, either lent long or converted into local RAND currency loans. A decade later the Federal Reserve opted not to renew these loans that ultimately lead to the closing of South African financial markets dependent upon a two-tiered currency infrastructure. The Eurozone is struggling to raise greenback liquidity in the short term; a dying last grasp of breath in the stormy global financial dire straits.

    Compounding the current global financial market malaise that is strengthening greenback demand is through short-term money markets as US investors “sell any and all international bids in stocks and bonds and [are] purchasing greenbacks to meet their obligations.”

    In this regard, is it imperative that a FX traded automated algorithm provides an aggregate view that has the expert ability to immediately sense best liquidity and price combination's in response to market conditions. It is critical, as we saw on September 29th – 30th that liquidity sensing and smart order routing happens with minimal latency – before the market finds its overall trend. (This example is attributed to Steve Meizeninger, Director of Education at the International Securities Exchange, a leading expert in Forex, equity and index options trading.)

    Note: JP Morgan released a new trading network called, "Dark Books" that doesn't publish quotes in the open market. Bank of America as incorporated a new FX algorithm called "Trade Weighted Average Price" (TWAP).

    Market Potential
    FX autodealing is a highly competitive market, gained by innovated firms that incorporate the latest technologies and algorithmic techniques, i.e. market aggregation, rules-based algorithmic trading, smart order routing, real-time risk management and trading on the economic news/market sentiment - all tailored to the FX Spot market. To date, institutional clients can access liquidity event feeds with over 20 interbanking platforms, with trading volumes having grown 90% with over 11,000 institutional global brokerage firms.

    FX Autotrading Theorem

    Can we make computers think more expertly then the human brain? Will we ever be convicted of murder for trashing our personal computer? Just as the Turing Test feasibility created by Alan Turing since it’s origin in the 1950’s has set the paradigm of technological development into the realm of artificial intelligence; it is just a matter of time that the complexity of the human brain is qua computer. Today, it is not that “man is imitating God” but that technology brings “man closer to God”.

  2. ali syed arsalan

    Dec 23, 2009
    Likes Received:
    at present i think computers are more efficient than humans because of emotionless behavior of computer.

Share This Page