Adaptive Modeler 1.2.0

Adaptive Modeler 1.2.0 offers an effective tool which helps you to create agent-based market simulation models for price forecasting of real world stocks, ETFs or other securities. Thousands of trading agents are provided with real world market data and use their trading rules to compete and adapt on a simulated market. Their collective behavior is used to generate forecasts and trading signals. Models evolve and adapt incrementally (walk-forward) in real-time without repeated optimization or overfitting on historical data. This results in better adaptability to changing market conditions and more consistent and reliable performance.
Adaptive Modeler features an easy to use drag-and-drop user interface, real-time charts and plots to visualize model evolution, behavior and performance, a user configurable genetic programming engine for trading rule creation, (custom) quote intervals ranging from 1 millisecond to multiple days or variable, Trading Simulator with hedge-fund style performance report, data export function, batch mode, User?s Guide, Tutorial, examples, context-sensitive help and much more

How does Adaptive Modeler differ from other Trading Software?

1. Most conventional trading software based on technical trading rules supports the user in finding or creating a (mostly static) rule-based trading strategy by optimizing or training on historical data. If one searches long enough, this approach will always produce a trading strategy that seems highly profitable on historical data. This however doesn't mean that this strategy will also perform well in the future when price behavior may be different. The apparent past success of the strategy has in fact merely been caused by repeatedly optimizing or training on the same historical data. This tends to lead to overfitting (or curve fitting) and is likely to produce trading rules that fail when exposed to new price data.
2. More advanced software may provide adaptive trading rules that automatically adapt to price developments using neural networks, genetic algorithms or other techniques. However, one adaptive trading rule will still not be able to capture the complex price behavior of a financial market caused by the interaction of various heterogeneous investors, and this approach still carries the risk of overfitting.
3. Agent EvolutionIn financial markets no single trading rule or strategy continues to beat the market for any long period of time. Financial markets are constantly changing and new trading strategies come and go, affecting price behavior and each other’s returns. As the market evolves, trading strategies need to evolve as well in order to stay profitable.
4. Instead of optimizing one or a few trading rules by back-testing them over and over again on the same historical data, Adaptive Modeler lets a multitude of trading strategies compete and evolve on a virtual market in real time. This means that every historical price is only used once for "testing" the trading rules (as in the real world). This process is sometimes also called unoptimized, walk-forward or out-of-sample tested. The overall behavior of the virtual market is the basis for trading signals.
5. Though technical trading rules still form the basic building blocks, Adaptive Modeler automates the process of creating new trading rules to adapt to market changes and also diversifies the risk of a single trading rule by using many different trading rules simultaneously to generate trading signals.

Major Advantages

1. No overfitting or curve fitting of historical (training) data
2. Trading signals are based on multitude of trading strategies instead of only one or a few
3. Trading strategies are constantly adapting to market changes instead of being static
4. Puts user in charge of high level model evolution control instead of low level rule programming
   

Major Features:

1. Market data retrieval
   • support for quote intervals ranging from 1 millisecond to multiple days
   • support for variable intervals (i.e. for constant range bars or tick data)
   • flexible and intelligent (CSV) ASCII file reader that automatically accepts a wide range of format variations such as those used by most charting and technical analysis software packages
   • use of open, high, low, close, bid, ask and volume data fields
   • automatic detection of quote interval, market trading hours, number of decimal digits, etc.
   • automatic detection of date and time formats (in most cases)
   • automatic detection/handling of missing quotes and changes in market trading hours
   • accurate calculation of actual market trading time during a period for accurate compounding of returns, volatilities, etc.
2. User configurable model parameters
   • market trading hours
   • population size
   • initial agent wealth distribution method (equal, Pareto, Maxwell-Boltzmann)
   • initial agent position distribution method (equal, Gaussian)
   • stepsize of agent position values
   • transaction costs for agents
   • random seed value
   • minimum price increment
   • forecast source (Virtual Market Price or Best Agents Price)
   • Best Agents group size
   • breeding frequency
   • minimum breeding age
   • parent selection group size
   • mutation probability
   • maximum genome size and depth
   • minimum and maximum initial genome depth
   • functions and terminals to use for creating trading rules
   • optional uniqueness requirement for creation of new trading rules
3. Model creation and evolution
   • saving and loading of model configurations
   • Mersenne twister pseudo random number generator
   • pausing and resuming
   • step mode
   • multi-threading (model evolution continues during most user operations)
4. Available output data (data series)
   • several return calculations of security, Trading Simulator and individual agents such as total (excess) return, compounded average (excess) return and trailing (excess) return
   • return distributions (of security or forecasts) with kurtosis
   • weighted/historical volatility of security, Trading Simulator and individual agents
   • Virtual Market price and Best Agents Price
   • bid, ask and spread on Real Market and Virtual Market
   • trading volume and number of trades on Virtual Market
   • number of buy/sell orders in orderbook before/after market clearing
   • agent defaults and margin calls
   • forecast, forecasted price change, forecast error, mean absolute error, (root) mean squared error, right/wrong forecasted price changes, Forecast Directional Accuracy, Forecast Directional Significance, Forecast Directional Area Under Curve (AUC)
   • filtered volatility (volatility during right forecasted bars and during wrong forecasted bars)
   • historical averages and distribution data series for agent values such as age, wealth, position, (excess) return, volatility, beta, trade duration, number of offspring, genome size, genome depth
   • genetic operators statistics such as average nodes crossed, average nodes mutated, number of mutations
   • Trading Simulator data series such as wealth, position, trades, total (excess) return, compounded average (excess) return, trailing (excess) return, weighted/historical volatility, beta, alpha, (relative) Value at Risk, Sharpe ratio, risk-adjusted return, maximum drawdown, MAR ratio
   • Historical and Monte Carlo Simulations of Trading Simulator returns based on user specified parameters such as investment horizon, sample period / expected drift and (filtered) volatility, expected forecast accuracy, wealth, VaR confidence level, etc.
5. Trading Simulator
   • user configurable parameters (allow short positions, broker commissions, spread, slippage, etc.)
   • forecast accuracy filter
   • performance overview including total (excess) return, compounded average (excess) return, beta, historical volatility, (Relative) Value at Risk, Maximum Drawdown, Sharpe Ratio, Alpha, Risk-adjusted return, MAR ratio
   • performance calculated according to user configurable parameters (compounding period (i.e. year, quarter ,month, week), number of periods, risk free rate, VaR confidence level)
   • sub period returns and statistics
6. Charts
   • bar charts and line charts with up to 8 series per chart (real-time)
   • histogram charts (real-time)
   • dragging and dropping of data series into charts
   • horizontal dragging of charts to browse through history
   • transparant data overlay and crosshair for showing current or mouse-over values
   • linking of charts for synchronized browsing and crosshairs
   • lineair/logarithmic scaling (automatic)
   • moving averages
7. Population window
   • scatter plots of 2 agent values
   • colored scatter plots of 3 agent values
   • agent density plots
   • correlation and regression (of agent values)
   • 3 different axis modes (auto, standard deviation intervals and custom)
   • 3 different gridline modes (round numbers, standard deviation intervals and bin edges)
8. Market Depth window
   • visualizes virtual market pricing mechanism by showing depth of orderbook before or after clearing
   • shows matching volume
   • cumulative or non-cumulative volumes
   • price range shown can be adjusted by user
9. Agent window
   • shows agent details such as age, wealth, position, (excess) returns, trade duration, volatility, beta, generation, genome size, genome depth, etc.
   • shows trading rule
   • allows fast browsing through all agents
10.  User Interface
    • customizable user interface (tabbed windows, moving windows, maximizing windows, renaming, maximizing charts, changing colors of chart gridlines and axes, white or black backgrounds)
    • creating multiple window instances possible (for Charts, Population and Agent Windows)
    • saving and loading of Styles (workspace layout)
    • choice between displaying US or European dates
    • automatic scaling of GUI elements to system font and dpi settings to support various screensizes
    • context-sensitive help (dialog boxes, data series tree, gene selection)
    • optional user interface tooltips
    • Startup window with recently used models, examples and tip of the day
    • Getting Started Tutorial
11. Data exporting
    • Automatic (real-time) exporting of any data series values to a CSV file
    • manual exporting of historical values of any data series to a CSV file
12. Batch processing
    • automatically create models for all quote files in a folder (using a given model configuration and Style)
    • automatically create multiple runs (models) for a security (using a given model configuration and Style)
    • automatically export results of multiple models to a single export file (CSV)
    • run models until end of quote file or for a given number or bars
    • automatic naming, saving and closing of models
    • saving and loading of batch settings
    • batch creation through application user interface or command line
13. Logger
    • keeps track of missing quotes, unexpected quote times and other non-critical irregularities in received quotes

Enhancements:

•  Improved market data retrieval. A flexible and intelligent ASCII file reader now accepts a wide range of quote file variations with no or minimal need for conversions.
• Support for high-frequency data (milliseconds) and variable intervals (for constant range bars or tick data).
• Bid/Ask price data supported. Open/High/Low (and Volume) now optional.
• Some bug fixes (see release notes on forum)

Requirements:

1. The minimum system requirements for running Adaptive Modeler are:
   • Windows 2000, XP or NT 4.0
   • Microsoft .Net Framework 2.0 or higher (will automatically be installed during installation of Adaptive Modeler if not present yet).
   • 512Mb RAM (supports up to 20,000 agents; for 100,000 agents 1Gb RAM is required)
2. Other requirements:
   • historical market data of the security to be modeled
3. Recommended:
   • data feed or downloader tool that retrieves market data and exports it to Comma Separated Values (CSV) ASCII text files
   • fast CPU
4. For simultaneously running multiple instances of Adaptive Modeler using the same quote file the following additional requirements apply:
   • for Windows 2000: SP3
   • for Windows XP: SP1

WareSeeker Editor