In this issue:
I. Commentary by Marge Sherald, CEO
II. NeuroShell Trader Links to Metatrader4
III. NeuroShell Trader Network Version sale extended
IV. How to stop this newsletter
***********************************************************
I. Commentary by Marge Sherald, CEO
1. ChaosHunter finds neural network inputs
We were doing some consulting work for a company that was trying to determine processing load by the hour. They were using NeuroShell Predictor but were having limited success because of a small number of inputs. So we tried the problem in ChaosHunter with the expectation that ChaosHunter would discover some additional inputs based upon combinations of the original inputs. Since the original data file had rows of processing load data by the hour for several years, we treated the problem as time series data. This led us to select technical indicators such as lags of the current production and momentum (the difference between the current hour’s production and production in the previous hour(s)).
Because we were using the optimization goal function to minimize mean-squared error, we added a column which was the lead of the current hour’s production and used that as the predicted output. To add the new column we simply copied the column which contained the current production values and pasted them into another column in which the cell values were shifted up one row (an Excel option when pasting cells).
Note: If you are building a trading model and select a goal function from the Trading Strategies section on the optimization tab in ChaosHunter, leading the output value would not be necessary because ChaosHunter presumes you are predicting one time period ahead for trading models.
We ended up with a ChaosHunter model that made far better predictions for processing load than the original NeuroShell Predictor model that did not include the lag and momentum inputs.
What was even more interesting was that if we used the lags and momentum input values identified in the ChaosHunter model as inputs in the NeuroShell Predictor model, that model improved significantly, even when we removed the constants and additions and subtractions that appeared in the ChaosHunter formula.
ChaosHunter proved once again that it is an excellent tool for discovering the correct inputs for a model, whether the model is for manufacturing, scientific, or trading purposes. The inputs found by ChaosHunter are also useful in other programs such as NeuroShell Trader and NeuroShell Predictor.
ChaosHunter is on sale until April 19, Check out the details below.
2. Timing reduction for NeuroShell Trader Network version
I was updating a MESA91 example chart that uses a crossover of the MESA91 Sine and MESA91 LeadSine cycle indicators. The model trades the S&P 500 E-Mini when the LeadSine crosses above the Sine indicator for a long entry and crosses below for the short entry. The other rule in the model uses the MESA91 Trend Vigor indicator to determine whether the market is in trend mode. When the market is in a strong trend, it is not advisable to trade a model based on cycle indicators such as MESA91 Sine and LeadSine. The MESA91 indicators are calculation intensive, so they are a good test for the network version of NeuroShell Trader.
Using 5 minute bars from January 2, 2013, to March 20, 2013, to train the model, it took 1 hour and 8 minutes on my 4 core processor computer. When I distributed the processing to my original computer and an I7 with 8 processing cores for a total of 12 processing cores, it only took 28 minutes to optimize the same model. The NeuroShell Trader Network version let me test many more model configurations during the day compared to the non-network version.
We’ve put the updated MESA91 example on www.ward.net. Just remember to log in with your MESA91 serial number and go to the Examples section.
*************************************************************************
II. NeuroShell Trader Links to International Brokers and Banks with MT4 Bridge by ZagTrader
You can now combine the point and click ability of NeuroShell Trader to easily build custom trading models with access to brokers and banks from around the world through the MT4 Bridge by ZagTrader.
MetaTrader 4 (MT4) is the trading platform known as the industry leader for dealing with international financial markets. Depending on the broker, you can obtain Forex and/or futures data from your broker and send trades all within the NeuroShell Trader software using the MT4 Bridge by ZagTrader.
When you sign up for a ZagTrader RT account, you have access to the MT4 Bridge, as well as value analysis, technical analysis, screeners, up to date market and company news, risk reports, and peer group analysis. You can take advantage of fundamental and technical analysis covering every single traded symbol. With your ZagTrader RT account, it’s like have your own personal financial analyst available 24 hours a day.
For more information on the MT4Bridge by ZagTrader, contact ZagTrader
North & South America: +1 (514) 661-1585 FREE
Europe, Middle East, Africa: +971-4-433-4385 FREE
www.zagtrader.com
*************************************************************
III. NeuroShell Trader Network Version sale extended
Since we announced the release of the network versions of NeuroShell Trader we’ve had a lot of questions from our customers. The network versions are designed to distribute OPTIMIZATION of a model over your local area network (the distribution does not currently work on VPN or remotely accessed computers). The result is a significant decrease in the amount of time it takes to create and update trading models as market conditions change.
The Trader Network versions are especially useful when you’re using the multiple template optimization and backtesting feature of the NeuroShell Trader PowerUser versions. The network versions also beneficial when you’re building models with multiple symbols and/or complex indicators, either from the Trader add-ons or custom indicators you’ve created from Trader’s original indicators
Let’s say your local network includes three computers: your main computer with 12 cores (the server) and two other quad core machines. That is a total of 20 cores, all of which can be searching for your optimum trading model simultaneously. The 12 core server will handle a higher share of the load.
You can choose from three different network versions depending on the power and speed you require. For example, the NeuroShell Trader Home Network version can distribute optimization processing to your server computer and up to two computers on your network (and even to multiple cores/threads on each of those computers). The server computer is the one with the NeuroShell Trader user interface, so that is the machine you use to set up your models. The other two computers are only run as clients so they do not need a separate serial number and do not have a user interface. In other words, the network license lets you activate your serial number on only one computer at a time and the other machines can act as clients only.
The NeuroShell Trader Office Network version can distribute optimization processing across ten different computers (one server computer and up to 9 clients).
The NeuroShell Trader Corporate Network version ramps up the power to a max of 25 computers (one server and up to 24 clients).
The resultant speed increase is dependent on the speed of the processors and clock speed on all computers as well as the overhead of controlling, setup, and communication with each distributed thread.
It is recommended that network optimization only be used on complex trading models that have a large number of parameters and which take more than a minute or two to optimize on a single computer. Smaller trading models can optimize more slowly across a network due to the control, setup and communications overhead and might not utilize all the available processing cores if more cores are available than needed for genetic optimization of smaller models.
Ward Systems Group will continue to offer pre-release pricing of the NeuroShell Trader Network versions until April 19, 2013. In order to purchase any of the network versions of NeuroShell Trader, you must already own either the NeuroShell Trader® Power User or NeuroShell DayTrader® Power User for the server computer. We’re also offering the PowerUser versions at a 20% discount until April 19, 2013.
Prices for the network versions listed below are in addition to the purchase price of either the NeuroShell Trader® Power User or NeuroShell DayTrader® Power User.
NeuroShell Trader Home Network (up to 3 computers) Upgrade price from Power User: $995.00 Pre-release Price: $ 796.00
NeuroShell Trader Office Network (up to 10 computers) Upgrade price from Power User: $2995.00 Pre-release price: $2396.00
NeuroShell Trader Corporate Network (up to 25 computers) Upgrade price from Power User: $7,495.00 Pre-release price: $5,996.00
To take advantage of this pre-release pricing, order only from the following web link: www.ward.net/60.asp or give us a call at 301 662 7950 or Skype us at wardsystems.
***********************************************************************
IV. ChaosHunter Sale
If you want to use ChaosHunter to find the best inputs for your trading, business, or scientific models, we’ll give you a chance to discover ChaosHunter at a 20% discount, so your cost is only $1196 compared to the list price of $1495. To take advantage of this sale, order only from the following web link: www.ward.net/60.asp or give us a call at 301 662 7950 or Skype us at wardsystems.
***********************************************************************
V. One way to stop this newsletter
It is really easy. Just change your E-mail address and don’t tell us.
***********************************************************************