Altreva News
20 Jul 2020 - Adaptive Modeler version 1.6.0
- Added a Bitcoin (BTCUSD) example model.
- Agents can now (optionally) trade and own fractional shares.
- Agent Window now also shows the agent's initial wealth and the number of transactions it has done.
- Some minor changes (see the Release Notes).
28 Jun 2020 - Adaptive Modeler version 1.5.6
- The gene statistics (introduced in version 1.5.5) slowed down model evolution
somewhat. Therefore they are now only being tracked when any gene statistics dataseries
are present in charts or in the current values window.
- Bugfix: Since version 1.5.5, models with 20,000 or more agents sometimes crashed at random moments.
This has been fixed.
- Some other bug fixes and minor changes (see the Release Notes).
5 May 2020 - Adaptive Modeler version 1.5.5
- Gene statistics data series that show the number of times genes occur in trading rules and how often they get
evaluated. This can be helpful in selecting genes.
- "Return" gene that gives the agent's Breeding Fitness Return. This allows agents to observe their
own trading performance, based on the assumption that the behavior of market participants is affected by their
own returns.
- Genome window now has buttons to browse through all genomes. If an
agent gets replaced while viewing its genome, the new genome will
immediately be shown. The window size and position are remembered now.
- Some bug fixes (see the Release Notes).
17 Jun 2019 - Altreva now accepts Bitcoin and other cryptocurrencies
We now accept Bitcoin, Bitcoin Cash, Ethereum, Litecoin and USD Coin. Cryptocurrency payments are processed by Coinbase.
A Coinbase account is not needed. See our store.
4 Jan 2019 - Adaptive Modeler version 1.5.4
- Larger population sizes: The Evaluation Edition now support 5,000 agents, the Standard Edition 25,000 agents
and the Professional Edition 250,000 agents.
- Better use of multi-core CPUs for parallel evaluation of trading rules of large (single) models (10,000+ agents).
- EMA genes (exponential moving average) with fixed period lengths of 10, 20, 50, 100 and 200 bars to make it
easier for agents to observe and compare short and long-term price developments using commonly used EMA periods.
- Several bug fixes
See the Release Notes for more information.
10 Oct 2017 - Adaptive Modeler version 1.5.3
This version contains some bug fixes and minor changes (see Release Notes).
28 Mar 2017 - Adaptive Modeler said to offer highest computational modeling strength and scalability versus model development effort
In a review study published in Computer Science Review which compares 86 state-of-the-art agent-based modeling and simulation tools,
Altreva Adaptive Modeler was found to offer the highest computational modeling
capacity and scalability level versus model development effort.
The article
is available online free of charge.
7 Apr 2016 - Adaptive Modeler version 1.5.2
- Optional generation of cash signal at market close to avoid overnight positions.
- Spread and slippage can now be specified either in points or as a percentage of price.
- New * gene to multiply Change values with Lag and Direction values.
- Some bug fixes and minor changes (see Release Notes).
7 Dec 2015 - Adaptive Modeler version 1.5.1
This version contains some bug fixes and minor changes (see Release Notes).
3 Feb 2015 - Altreva awarded for best Agent-Based Market Simulation Models
Altreva has won the
Acquisition International Hedge Fund Award 2015 for Best Agent-Based Market Simulation Models!
5 Jan 2015 - Adaptive Modeler version 1.5.0
- Multiple custom input variables are now supported (up to 100)
(Professional Edition only).
- New "mrlim" and "tflim" genes for generating order limit prices, based on mean-reverting
and trend-following price expectations.
- The Monte Carlo Simulation data series now uses a student's t-distribution with customizable
degrees of freedom to better simulate the returns of stocks and other securities.
- Some bug fixes and changes (see Release Notes).
15 Sep 2014 - Adaptive Modeler version 1.4.3
- Average Trade Duration added to Performance Overview.
- Trades statistics are now also available as data series so that they can be exported.
- Assignment of initial position values of agents can now also use a uniform distribution.
- Several small changes and bug fixes (see Release Notes).
6 Apr 2014 - Adaptive Modeler version 1.4.2
- Tournament selection for selecting agents for breeding with adjustable
tournament size to control the selection pressure.
- Adjustable crossover operator to control the average size of subtrees that are being exchanged.
This can affect the genetic evolution rate and granularity (Professional
Edition only).
- Some bug fixes (see Release Notes).
5 Dec 2013 - Profitability in Forex trading
Altreva Adaptive Modeler has contributed to new evidence of technical trading profitability
in the forex market. Independent scientific research has demonstrated economically and
statistically significant out-of-sample excess returns (after transaction costs) for the six
most traded currency pairs, which are superior to returns achieved by traditional econometric
forecasting models. The monthly returns (unleveraged!) for the six currency pairs ranged from
3.8% to 5.9%. This research from the University of Newcastle has been published in
Economics
Bulletin and is
freely accessible.
4 Dec 2013 - Adaptive Modeler version 1.4.1 (build 42883)
A bug in version 1.4.1 has been fixed that caused the compounding interval of some
data series to be changed when opening models that were created by older versions.
Users who already installed the previous 1.4.1 build are advised to install the
latest build and no longer use the previous build.
30 Nov 2013 - Adaptive Modeler version 1.4.1
- The Performance Overview now includes trades statistics including the number of long/short trades,
winning trades%, average return per trade and Profit Factor.
- Some minor changes and bug fixes (see Release
Notes).
22 Oct 2013 - Altreva included in Money 101: Top Resources for Finance Majors
Altreva Adaptive Modeler has been recognized as one of the 101 top resources for finance majors
in this list
by Online Finance Degree
which contains major global economic magazines, blogs, financial institutions and financial modeling tools.
The site offers a guide to online finance degrees and aims to help finance students to
profitably leverage the information they can gain from a wide variety of sources.
15 Mar 2013 - Adaptive Modeler version 1.4.0
- An additional custom input variable can now be included in the quote file and used by agents.
This can be any variable that could help forecasting such as fundamental data, an economic indicator,
another price series or a custom technical indicator (Professional Edition only).
- Batch dialog now has a button to select a folder with quote files.
- Bug fixed that caused models in a batch to sometimes get the exact same random seed.
- The exact date/time format can now also be specified in a quote file header when date and time are combined in one column.
- Some minor changes and bug fixes (see Release
Notes).
11 Jun 2012 - Adaptive Modeler version 1.3.0
- New 'spread' gene that returns the spread on the Real or Virtual Market.
- New 'barrange' gene that returns the high/low range of a bar on the Real Market.
- Some bug fixes and minor changes (see Release
Notes).
30 Sep 2011 - Adaptive Modeler version 1.2.9
- Genes have been added that check what day of the week it is so that agents can use this
information in their trading rule. Because it is often suggested that markets may have some
degree of 'weekday effect' and since the day of the week is always already included in
the input data, we thought it useful to add these genes. It turns out that selecting all the
'day' genes (IsMon, IsTue, IsWed, IsThu, IsFri) causes remarkable improvements for the S&P500
index and some other markets when compared to the default configuration.
- PopulationShares data series that shows the total (net) number of shares in the population.
This can be used to see the amount of shares that flow into or out of the population and to
what extent the model is an 'open' or 'closed economy'.
- Some minor changes and bug fixes.
11 Jun 2011 - Adaptive Modeler version 1.2.8
- New data series for historical averages and standard deviations of Population Distribution series
like Position, Trade Duration, Beta and Volatility. This enables easier monitoring and exporting of
historical distribution data i.e. as proxies for the heterogeneity or homogeneity of stylized agent behavior.
- Some bug fixes (see Release Notes).
16 Dec 2010 - Adaptive Modeler version 1.2.7
- Improvements in agent charts: all agent replacements are now tracked in agent charts, not only the last one;
agent wealth chart now stores history of all previous agent in chart when model is saved; last value of previous
agent is no longer connected to first value of new agent.
- Startup splash screen with progress bar.
- Some minor changes and bug fixes (see Release
Notes).
21 May 2010 - Adaptive Modeler version 1.2.6
- Agents can now (optionally) also place market orders (besides limit orders).
- The Market Depth window now provides a summary table with the number of buy and sell orders grouped by limit and market orders.
- Percentage of agents buying and selling (through the Orderbook data series)
- Virtual Market return and volatility data series added for better analysis of Virtual Market
behavior.
- The Standard Edition now support populations of up to 10,000 agents,
manual exporting up to 99,999 bars and an unlimited number of models per
batch.
- Bug fixes (see Release Notes)
26 Feb 2010 - Adaptive Modeler version 1.2.5
- Trading Signals and Right/Wrong Forecasted Price Changes are now also shown in
charts with longer periods (up to 5 years for daily quotes).
- Position of dialog box for changing data series and performance window parameters is remembered.
- Some bug fixes
7 Jan 2010 - Adaptive Modeler version 1.2.4
- Sortino ratio added (as a data series and in the Performance Overview)
- Models can now automatically be updated from the command line using an
"/Update" switch (opens a model, evolves it until the end of the quote
file, saves the model and exits).
- Window tabs can now be reordered by dragging them with the mouse
- Windows now have a context menu for renaming, removing, etc.
(right-click on title-bar or tab)
- Some minor changes and bug fixes
5 Jan 2010 - Several Example Models added
Several example models for major US and international stock market
indices as well as some stocks and commodities are now available for downloading.
Also, their performance is updated daily and shown online. For more information, see
Example Models.
19 Oct 2009 - Adaptive Modeler version 1.2.3
- Improved possibilities to specify the calculation period for data
series and Performance Overview. Any date range can now be specified,
for instance to compare performance before and after a given date.
Performance Overview now also shows FDA and number of trades during
calculation period and other improvements.
- Some bug fixes (see Release Notes)
8 Jul 2009 - Adaptive Modeler version 1.2.2
- Bar Return data series for quantitative analysis of bar-to-bar returns,
log returns, absolute returns, etc.
- Autocorrelation indicators to measure autocorrelation of returns,
volatility, volume and some other data series.
- Hurst exponent (time series predictability indicator)
- Moving Averages and Autocorrelation indicators can now be shown
without their source.
- Several minor improvements and bug fixes (see
Release Notes)
28 May 2009 - Adaptive Modeler version 1.2.1
- Improved consistency and reproducibility of models across different
machines. The potential effects of different CPU types, Operating System
versions and settings and .Net runtime versions on model evolution have
been reduced. This helps reproducing an exact model from its
configuration file (with a fixed random seed) on another machine and
keeping copied models running on different machines in sync (such as the
Sample model).
- A bug has been fixed that caused models that used the RndLim gene to
be irreproducible (even on the same machine).
- Some minor changes and bug fixes (see Release
Notes)
16 May 2009 - S&P500 model 8 months out-of-sample return is 179%
The S&P500 index sample model has reached
an out-of-sample return of 179% over the last 8 months (since it was
created) compared to a return of -27% of the index itself.
11 May 2009 - Adaptive Modeler version 1.2.0
- 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)
12 Apr 2009 - Adaptive Modeler version 1.1.4
- Getting Started Tutorial
- User friendly Startup window
- Examples
- Recently used models in File menu and Startup window
- Minor changes and bug fixes (see Release
Notes)
4 Jan 2009 - Adaptive Modeler version 1.1.3
- Besides the Professional and Evaluation Editions, a Standard Edition
is now available as well, offering the core functionality including
real-time forecasts and trading signals (see
Compare Editions).
- Batch mode: pausing models at end of batch is now optional.
- Some minor bug fixes.
17 Dec 2008 - S&P500 model 3m out-of-sample return is 127%
The S&P500 index sample model has achieved a true
out-of-sample return of 127% in the 3 months since it was created on
9/16/2008 and 101% since it was first published online on 9/26/2008.
More about the sample model.
12 Dec 2008 - Adaptive Modeler version 1.1.2
- Distribution series now differentiate between zero values and n/a
values. I.e. agents whose return can't be calculated yet because their
age is too low are excluded from the distribution whereas before they
were included as zero values. This improves the distribution charts and
Population window and also provides more accurate calculation of mean,
stdev, correlation, etc. In the Population window n/a values are shown
in green on the "Z-axis".
- "Tip of the day" in Introduction screen.
- Some bug fixes (see Release Notes).
18 Nov 2008 - Adaptive Modeler version 1.1.1
- A bug has been fixed that could cause an exception when model
evolution approaches the maximum number of bars that can be stored in
memory. (100,000 for the Professional Edition and 20,000 for the Evaluation
Edition).
- Market Depth window gridlines and axes are now redrawn after colors
have been changed in Options window.
11 Nov 2008 - Adaptive Modeler version 1.1.0
- Market Depth window: visualizes Virtual Market pricing by showing
depth of orderbook before and after clearing
- Area Under ROC Curve (AUC) test of forecast directional accuracy
- Volatility gene added
- Charts and all other windows can have white background now
- Some other minor changes and bug fixes (see
Release Notes)
This version is backwards compatible with files from version 1.0.1.
26 Sep 2008 - Professional Edition released
Adaptive Modeler 1.0.1 Professional Edition is now available,
offering more agents, more bars, better export and batch
functions and more. Also see the detailed
Edition comparison. The Professional Edition can now be ordered from
the online Store.
26 Sep 2008 - Adaptive Modeler version 1.0.1
- Automatic Scaling to support various
screensizes, fonts and DPI settings. All windows, dialogs and other
GUI
elements now scale well on various configurations.
- Some bug fixes
(see Release Notes).
- Note:
Contrary to earlier notice, backwards compatibility with data files from
version 1.0 or earlier is unfortunately not yet supported due to further
improvements in internal data structures.
11 Jul 2008 - Adaptive Modeler version 1.0b
Version 1.0b has been released!
This marks the end of alpha stage and in effect is meant to indicate
that Adaptive Modeler has reached a more professional deployment level,
including improved speed and scalability, more efficient memory usage,
backwards compatibility with previous versions and more extensive
testing. New features and improvements will of course continue to be
added.
New features, improvements and bug fixes in this version:
Charts:
- Browsing through history of charts
- Crosshair in charts for showing values at mouse
- Charts can be linked for synchronized browsing and crosshairs
Agent-based model:
- Agent initialization options (wealth can be equal, Pareto distributed
or Maxwell-Boltzmann distributed; asset allocation can be equal or
Gaussian distributed)
- Configurable stepsize of agent position
- Support for zero-intelligence trading
- Uniqueness requirement for creation of new genomes is now optional
- Initial margin check before order placement
New data series:
- Forecast error
- Mean absolute error
- (Root) mean square error
- Population cash
Various:
- Choice between US or European dates
- Automatic detection of number of decimal digits in quote file
- Context-sensitive help for genes and types
- Batch exporting: random seed exported with model data
- Creation of initial population can be canceled by user (i.e. when too
slow)
Improved
computational performance and scalability:
- Faster creation of (unique) genomes; is now O(log n)
- Faster Virtual Market clearing; is now O(n log n)
- Faster loading of models, smaller model files
- More efficient memory use
- Faster exporting of historical bars
Bug fixes:
- Order validation was incorrect sometimes causing extreme Virtual
Market prices and agent wealth values.
- Defaults counting was inaccurate in some cases.
- Agent Total Excess Return data series tried to calculate market return
for bar that might not be in memory anymore. Now if agent creation bar
is not in memory anymore, no value is returned.
- Agent Replacement Fitness Excess Return data series returned
inaccurate value when agent creation bar is no longer in memory. Now if
agent creation bar is no longer in memory, no value is returned. In
the distribution data series, the agent's return is set to 0.
- Logarithmic scaling exceptions
- Performance window: when compounding period is longer than memory
size, an exception was caused.
- Population window: exception when drawing density chart or legend and
only 1 value on z-axis.
- Population window: number of gridlines drawn at bin edges could exceed
maximum, causing exception.
- Some data series gave index out of range when MinBreedingAge was
higher than 100.
- Batch started from command line failed to launch multiple models.
- Chart period combobox was not being repopulated when Market Trading
Hours changes resulted in less possible chart periods.
- In long term chart, last bar's value was sometimes not visible.
- Unstable/alternating x-axis gridlines and labels
8 May 2008 - Adaptive Modeler version 0.99
(build 3050)
A number of bugs have been fixed, most notably:
- application no longer starts minimized after running a batch
- very wide charts with many gridlines no longer cause errors
- extreme values in Distribution charts no longer cause errors
- multiple changes to the Market Trading Hours that occurred during a
(data series) calculation period no longer cause errors
12 Mar 2008 - Adaptive Modeler version 0.99
- Creation of trading rules through genetic programming is now
documented in detail
- Trading rules can now be viewed
- Selection of 'genes' (functions) to be used in trading rules is now
user configurable
- Maximum size and depth of trading rules is now user configurable
- Improved Batch mode with GUI (simplifies creation of batch processes)
- Context-sensitive help system
- "Best Agents price" as alternative to Virtual Market price forecast
(with adjustable groupsize)
- Agent default management and improved short positions management
- Volume data is now being used by trading rules
- Improved pseudo random number generator (uses Mersenne twister
algorithm, stores seed value, seed value can be set by user)
- New data series: Right/Wrong Foreceasted Price Changes, Buy/Sell
Orders in Orderbook, Spread, Margin Calls, Defaults, Volume, etc.
- Maximizing of windows
- New chart interval of "4 hours" added
- Performance Panel and Distribution charts are no longer being
recalculated when not visible
- Tooltips in data series treeview with short descriptions
- Various bug fixes and minor improvements
25 Jul 2007 - Adaptive Modeler version 0.98
- Population window (visualizes model dynamics by plotting 2 or 3
agent properties simultaneously in a (colored) scatter chart including
correlation and regression)
- Several new agent time & distribution data series (breeding fitness
return, replacement fitness return, trade duration, volatility, beta,
generation, etc.)
- Agent window to show agent details
- Customizable User Interface (moving windows, tabbed windows,
maximizing charts, etc.)
- Support for multiple Charts windows/tabs
- Return distribution data series with kurtosis (for prices and
forecasts)
- Up to 8 data series can be shown per chart
- Histogram bin details can be shown in data overlay
- Improved quote retrieval (incl. support for MetaTrader4 export files)
- Various minor features, improvements and bug fixes
3 Jul 2007 - Life detected in Adaptive Modeler!
Adaptive Modeler is being featured in a
show
case by Evil Limited, a
company devoted to Electronic Virtual Intelligent
Life, as an illustration of the application of simple agents interacting
and evolving in a
complex (non linear) way to outperform financial markets.
12 May 2007 - Adaptive Modeler included in
ACE Comp Labs and Demos
(Agent-Based Computational Economics Computational Laboratories and
Demonstration Software) directory of Professor Leigh Tesfatsion
(Department of Economics, Iowa State University).
28 Feb 2007 - Adaptive Modeler version 0.97
- New return & risk indicators (alpha, beta, Sharpe ratio, VaR, Risk-adjusted return, Max Drawdown, MAR, etc.)
- Trading Simulator Performance Overview with various return/risk indicators and sub period results & statistics
- Data exporting
- Histogram charts showing agent wealth distribution, simulation return distributions, etc.
- FDA-based filter for trading signals (user configurable)
- In-chart data overlay showing current/mouse-over values
- Automatic linear/logarithmic scaling of charts
- Simple batch functionality through command line (i.e. automatic creation of models from all quote files in a directory and exporting of results)
- New data series (excess return, weighted FDA, total FDA, etc.)
- Several improvements to existing features and bug fixes
14 Sep 2006 - Adaptive Modeler version 0.96
(build 2448)
- Trading Signals are now visualized
in charts with up/down arrows
- Fixed bugs and documented some issues regarding importing Yahoo CSV
quote files
16 Aug 2006 - Altreva Adaptive Modeler version 0.96
- Saving and loading of configurations (model parameters)
- Improved quote retrieval (new quotes are now read whenever quote file
has been updated, timezone independent)
- Faster quote file pre-processing
- Bugs fixed relating to 24h markets (forex)
- Bugs fixed relating to short intervals
24 Jul 2006 - Altreva Adaptive Modeler version 0.95
- Risk Management functions (i.e. historical/Monte Carlo simulations)
- Improved & faster charts (including bar charts)
- Selected data series and window layout can be saved separately and applied to any other model
- Several new data series (volatility, more population/agent/GP info)
- Configurable data series
- Moving Averages
- Automatic handling of missing quotes and changes in trading hours
- Trading Signals
- Logger (to keep track of notifications, warnings, etc.)
- Various other improvements and bug fixes
30 Dec 2005 - Altreva Adaptive Modeler version 0.94
- evolved sample model included with installation
- user friendly introduction screen and tooltips
- improved screen layout
- various minor improvements and bugs fixed
26 Oct 2005 - Altreva Adaptive Modeler version 0.93a
- added support for common ASCII quote file formats such as
MetaStock ASCII and Yahoo CSV.
14 Oct 2005 - Altreva Adaptive Modeler version 0.92a
- toolbar added
- some improvements to agent-based model (i.e. limitation of short positions)
13 Sep 2005 - Altreva Adaptive Modeler version 0.91a
- extensive documentation added
- various improvements and bugs fixed
26 Aug 2005 - Altreva Adaptive Modeler version 0.9a
Initial pre-release version.
