In this issue:
I. NeuroShell Trader Advanced Seminar heads west
II. Commentary by Steve Ward, CEO
III. Offices closed Labor Day
IV. eSignal 8.1 is a problem
V. One way to stop this newsletter
*******************************************************
I. NeuroShell Trader Advanced Seminar heads west
In response to requests from NeuroShell Trader users who live on the west coast, the NeuroShell Trader Advanced Seminar will be offered in Tucson, Arizona, from October 9 – 11, 2007. The informal seminar will be led by Steve Ward, CEO and Chief Technical Officer of Ward Systems Group, and Marge Sherald, COO of Ward Systems Group Inc. and author of many articles on artificial intelligence and trading.
The seminar is planned to be interactive with questions and answers, consistent with being able to cover all of the material. The size of the seminar is limited to enable everyone to take part, and the special rate on hotel rooms is only available for a limited period of time, so please decide soon if you want to attend this event.
For complete details see our website at http://www.neuroshell.com/seminar.asp
Why should you attend? We think one of our attendees from the March 2007 seminar says it best:
“Thank you for another excellent seminar – the best yet. I won’t be able to get to working with all the techniques you gave us for awhile, but with the excellent documentation you sent with us, I will be able to implement any that I wish whenever I have the opportunity. Even though I began using NeuroShell first on a 286 12 MHz computer, I found this seminar to be extremely valuable, and I expect to easily recoup the cost in improved models. While I sensed that even very experienced users found much to learn at the seminar, less experienced users were able to see not only where they were going, but how to get there. This seminar was a very significant boost up the learning curve.”
*******************************************************
II. Commentary by Steve Ward, CEO
1. Surfers. Have you ever watched surfers from the beach on a day with pretty good waves? On those days have you ever noticed that some of the surfers ride out, then wait a few hours for the perfect wave, and, not finding one, ride back in to try again another day? Neither have I. It doesn’t happen; surfers don’t wait for the perfect wave to surf. But for some reason, many of our customers wait for that perfect model to trade. The perfect model runs very profitable into the future for years, with negligible drawdown. Like the perfect wave, I don’t think it exists. So what I do is build an odd number of good models and then go with the “majority” signal. Actually I have an odd number of daily models, and odd numbered sets of intraday models set for different times of the day I like to trade.
Have you ever seen surfers measuring the waves with a yardstick before they decide to go out in the water? Ok, you know where I’m going here. Market data and neural net models are not precision watches and they are not an accountant’s spreadsheet. They are estimates and you shouldn’t go nuts if your data isn’t perfect or some statistic isn’t exactly what you think it should be. Get on with your trading life.
Invariably after I give lectures like the ones above, I catch myself doing exactly what I am preaching against. I think the tendency towards these things is what keeps analytic people from becoming successful traders.
Many people, including many of our best customers, say you should trust your models implicitly and dispense with your judgment. I’m not totally convinced of that yet, even though my business is models. Even if you have great models it helps a lot to become a good trader if you are going to use your models intraday on margin, including futures and Forex of course. Even the best models are sometimes wrong, and it is hard to implicitly trust your model when the market starts moving in the wrong way fast on margin! I have lost money even when my model was right because it started out the wrong way and I bailed out too fast. That would argue for trusting. I have also made money when my model was wrong because it started right and I got out in time. That would argue for becoming a good trader. Post your thoughts about this on the forum, because I think it is an important open question.
2. Tucson Seminar. I really enjoyed the last one and so did most if not all of the attendees. I liked meeting the people I’ve talked to over the years, and the class enjoyed interaction with other traders like themselves. Everyone I talked to felt they got more than their money’s worth, not just from what they learned from me and Marge, but also what they learned from others. Before the seminar, during breaks, and well into the evening after the seminar there were little groups exchanging ideas everywhere. Every seminar we have ever had has been either in Florida or Maryland, but we are now bending to the desires of our western customers at least this once. See you at the Marriott in Tucson University Park in October!
*******************************************************
III. Offices closed Labor Day
The Ward Systems Group offices will be closed Monday, September 3, in observance of Labor Day.
*******************************************************
IV. eSignal 8.1 is a problem
It appears that there is an 8.1 eSignal Data Manager available, that does not include the eSignal charting software (the standard Data Manager that comes with the eSignal charting software is 8.0). We have obtained this release and have found that it does not work properly with the NeuroShell Trader software. In fact some users have told us it doesn’t even work right with eSignal charting. Do not use this release; continue to use eSignal 8.0 until the problems are resolved.
*******************************************************
V. One way to stop this newsletter
It is really easy. Just change your email address and don’t tell us.
*******************************************************