wmGO

Whitmark Genetic Optimizer 2.5 for NinjaTrader 6.5

(Click on button above to see exhibits full screen)

Description

The Whitmark Genetic Optimizer (wmGO) takes advantage of a genetic algorithm (GA) to to put your research and development efforts on overdrive. It significantly improves your ability to conduct iterative backtests and optimizations of your trading strategy parameters. These parameters can be either inputs that drive the behavior of an indicator or a series of switches to activate specific trading rules or setups within your trading strategy.

The wmGO is a much smarter and more efficient means to select the best settings, particularly as the number of parameters of your strategy grows or the number of values you want to test for each parameter. The wmGO integrates with the NinjaTrader Strategy Analyzer and offers many features not found elsewhere. I, as well as several of my clients, have found the wmGO to an invaluable tool to test the viability of new strategies and dramatically accelerate strategy research and development efforts.

Benefits

Your ability has evolved to design and develop your own sophisticated trading strategies so you need more advanced tools in the quest for the most robust trading strategies and parameter settings. "Exhaustive" or "brute force" optimization is vastly inefficient when it comes to searching large solution spaces and is just not smart enough nor efficient enough to give you the kind of throughput you need to expedite your research and development. If you find that you are severely restricting the range of values you seek to test or optimize to keep current run time reasonable, then the wmGO can vastly improve your productivity.

Features

Smart Search. The wmGO is an alternative to the NinjaTrader Strategy Analyzer default "exhaustive" search optimizer that can be used for both optimization and walk forward operations.
Multiple fitness functions. Go beyond simplistic measures of performance to guide the optimization process to search for the preferred equity curve behavior . . . an upward sloping equity curve with modest drawdown. Fitness function selections can now be made directly from the wmGO control panel. Available fitness functions include:
- Sqrt(Trades)*Avg(Profits)/StdDev(Profits). Best for strategies that generate higher win rates and modest profits per winning trade.
- TotalProf*Wins/MaxDD. Best for strategies that generate lower win rates and have large profits per winning trade.
- Net Profit.
- Net Profit per Trade.
- Sharpe Ratio. Must have at least one month of trading activity to evaluate this fitness function. Uses the standard Sharp Ratio system performance statistic found in NinjaTrader.
- Expectancy per Day. Made popular by Van Tharp in his book, Trade Your Way to Financial Freedom.
- Other fitness functions to be added in time. Proprietary fitness function can be incorporated on a customized basis.
All Markets, All Timeframes. Works for all markets (e.g., equities, futures, and forex instruments) and all timeframes (e.g., tick, volume, minutes, seconds, range).
Efficient Operations. Some genetic optimizers will revisit the same set of parameter values and reprocess these iterations thereby wasting valuable processing time. The wmGO will not reprocess duplicate parameter sets to promote a speedy convergence to the optimal solution.
Specify Parameter Increments. Some genetic optimizers do not use parameter increments when selecting candidate parameter values, resulting in a search method that could get "stuck" in a local optimum. The wmGO has an option to specify whether you would like to use the increment values to obtain better coverage in the solution space.
Bootstrap Current "Best" Parameters. So, you think the parameter values you have identified through arduous trail and error are pretty good? When the Bootstrap mode is selected your "best" parameter values will be forced into the initial generation when submitted to the optimizer as a "Min" value. The "Max" value will be used to specify the value range (Max-Min) to test above and below your "best" value submitted in the Min field. In other words, if your "best" lookback period is 13, and you've entered 13 in to the "Min" field and 18 into the "Max: field:
- The parameter value of 13 will be forced into the initial generation of the algorithm and be allowed to influence the solutions of future generations.
- The range to test above and below the value 13 will be defined by the defined "Max" less the "Min" or a value of 5 = 18 max -13 min.
- The new "Min" will now be 8 = 13 min - 5 range and the "Max" will remain at 18.
- This approach is the only way to force in a "best" scenario into the initial generation given the current NT 6.5 user interface.
Repeatable Seeded Runs. Optimization runs can be "seeded" which allows an optimization to be rerun with the same series of "random" values. This can be useful when testing alternative versions of a strategy to ensure performance differences are due to changes in the strategy logic verses the stochastic nature of the genetic algorithm methods.
Optimization Settings Output. The settings used to control the optimizer are written to the output window to provide an audit trail to ensure this important information is keep together with your optimization results. These fields include:
- Strategy: Name, Parameter Count
- DataSeries: Primary Instrument, PeriodType, PeriodInterval, PointValue, TickSize
- TimeFrame: BackTestMode, BackTestFrom, BackTestTo, ExcludeWeekends, SessionBegin, SessionEnd
- General: IncludeCommission, MinimumBarsRequired
- Historical Fill: FillType, Slippage= 0
- Optimize: KeepBestResults, Method, Type, OptimizationPeriod(WFO), TestPeriod(WFO)
- OrderHandling: EntriesPerDirection, EntryHandling, ExitOnClose
- OrderProperties: DefaultQuantity, QuantityType, TimeInForce
- GASettings1: GASeed, PopSize, MaxGen, ReproPct, CrossType, MutRate, MinTrades, Threshold
- GASettings2: BootStrap, UseIncr, SaveResults
- GASettings3: FitType, FitName
- Iterations: PrimaryBarCount, MinBarOrTickCount, ExhaustiveIterationCount, GOIterationCount
Tool Tips. All input labels and fields are provided with tool tips to assist you in managing the settings. Just hover your mouse over each to read the content.
Optimization Statistics Output. In addition to the run summary and results in the output window, the parameter values for each iteration including the fitness function and standard trade performance statistics (see below) can be saved to the wmGO folder in a comma-delimited .csv format to be examined in a text editor or MS Excel for further analysis. If you have ever tried to parse out the parameters in the Strategy Analyzer, you will find this to be a big timesaver.
- Each Parameter Value Parsed
- Fitness Function Value
- Fitness Function Type
- Total Net Profit
- Long Net Profit
- Short Net Profit
- Gross Profit
- Gross Loss
- Commission
- Profit Factor
- Cumulative Profit Pct
- Draw Down Pct
- Draw Down Amt
- Sharpe Ratio
- Total Trades
- Pct Profitable
- Winning Trades
- Losing Trades
- Avg Profit
- Winner Avg Profit
- Loser Avg Profit
- Win Lose Ratio
- Max Consecutive Winners
- Max Consecutive Losers
- Largest Winner
- Largest Loser
- Trades per Day
- Avg Trade Duration Minutes
- Avg Bars in Trade
- Profit Per Month
- Max Time to Recover Days
- Avg Maximum Adverse Excursion (MAE)
- Avg Maximum Favorable Excursion (MFE)
- Avg End Trade Drawdown (ETD)
- Iteration Number (duplicate iterations excluded)
Reference Strategy Included. Includes a reference strategy, wmGORefStrategy1 (featured in the video below) that demonstrates how the wmGO can be used to select the most desirable:
- indicators or sets of indicators
- Trade entry and exit setups
- Trade filters
- Money management methods
- parameter values used as inputs to drive all of the above
Additional Frequently Asked Questions (FAQs) can be found here.

Video Overview

To see the wmGO in action, please click on the link below to view an 18 minute Flash video:

Overview of wmGO 2.0 using provided reference strategy. (1024 x 786 Flash Video)

Overview of wmGO 2.5 using provided reference strategy. (TBD)

For best results, view full size or full screen.

Requirements

The wmGO works exclusively within the Strategy Analyzer and, like the default optimizer, will evaluate historical fills on-bar-close only. Ninja allows historical fill model options of "fill if touched" (default) and "fill if penetrated" (liberal) for your entry and exit limit orders.
Given the nature of how the wmGO works, all parameters must be numeric values, even if the parameter is not to be used in the optimization. Therefore, all boolean, enumerated, and string parameters types must be converted to numeric types (integer, double). Typically, the changes required to make a strategy wmGO compatible can be done in several minutes by the strategy developer.
To use the wmGO you must be connected to the internet to enable the software license to authenticate your copy of the software with each use. You do not have to be connected to your broker data provider to use this product. NinjaTrader supports the concurrent use of the same NinjaTrader license for machines that are connected to the internet but not connected to a broker.
Currently the wmGO is only supported on NinjaTrader platform version 6.5 (demo or paid).

Terms of Use

The wmGO 2.5 for NinjaTrader 6.5 is available for lease until June 30, 2010.
This product is exclusively written for NinjaTrader 6.5. Although there are no known reason why the wmGO 2.5 will not work with NT7, until the beta NT7 arrives and we are able to determine the feasibility and level of effort to make it compatible, we will not commit to supporting this release of the wmGO under NT7.
The wmGO is for the sole use of the lessee for up to two devices (desktop or laptop computers) owned by the lessee.
Additional End User License Agreement (EULA) terms apply and are found in the installation setup.
Once clients have a chance work with the wmGO, they undoubtedly want to know everything about this technology. To keep the offer price of this valuable tool at a reasonable price, Whitmark Development reserves the right to limit support for extensive inquiries of this technology or its application. While most strategy developers will have no problems making their strategies ready for wmGO, additional paid consultative support is available for strategy design, conversion, and development assistance if needed.

Free Trial Download

The Whitmark Genetic Optimzer (wmGO) is available for a 3-day trial period. To install the wmGO, please click the link below.

Download the Whitmark Genetic Optimizer 2.5 for NinjaTrader 6.5

If you performed a trial of an earlier wmGOv2.X release that has since expired and would like to try out this latest release, please email your Machine ID to info@whitmarkdevelopment.com.
To install the wmGO, simply:
- Close the NinjaTrader 6.5 platform.
- Execute the wmG0 setup application from the link above.
- Restart NinjaTrader 6.5 platform.
- After viewing the overview video above, launch a new Strategy Analyzer window and test drive it yourself with the sample strategy wmGORefStrategy1 or your own compatible strategy.
During your trial, make sure you are testing strategies that will generate 30+ over at least a couple of weeks of data.
To continue your usage following the 3-day trial period, you will need to return to this page to supply your machine ID to lease and activate the wmGO.
The wmGO can be readily uninstalled if you choose not to lease.

Price: $295 USD

Licensing for this product is managed by an online license management system. To purchase the product, you will need to cut and paste the 32-digit Machine ID that can be found from the NinjaTrader Control Center > Help > About dialog box. Click here for additional information on finding your NinjaTrader Machine ID.

Please note, you may purchase this product without supplying the NinjaTrader Machine ID, but we will need to follow-up with you via email to obtain this information before activating the product. Please allow 24 hours for the activation from the time we receive your Machine ID.