Neural Market Trends

I got my ass spanked in the markets over the last year.  My 401k portfolios are surviving after some severe drawdowns (25%) between 2008 and 2009 and my trading capital in my Forex and Stock accounts got destroyed.  Luckily it was only risk capital, money I could afford to lose, but it still hurts anway.  So what did I learn from this hell of a market?  I learned that I’m not as smart as the market and that I tried to second guess it too much.  When times were good my system worked, when times got bad It hammered me. 

Perhaps the biggest thing I learned in this market crash is that nearly all my neural net trading models FAILED.  Yes, the models that were designed to alert me to impending market doom failed for the most part.  My volatility and timing model did OK but they had some major issues as well which I’ll discuss below. 

So today’s post is really a cathartic one for me.  It’s a mea cupla of sorts, a post in which I take my lumps and start to exorcise my trading demons.  It’s about the things I did wrong over the past year and hopefully a reminder to never repeat again.  Plus, it’s another way to ease myself into trading and posting to this blog again. 

As I rebuild my portfolios in 2010, I expect not to solely rely on neural net models again.  I realized that in a crazy market, neural nets can’t adjust fast enough.  As the market was crashing, my models were generating BUY signals when the market was obviously telling me they were wrong.  If only I had listened to the market instead of the models!  I learned that my biggest mistake was something called “confirmation bias.”  I wanted to believe what my models were telling and I traded accordingly.  Why?  Well it always seemed to have worked in the past, so why should this time be different? 

By the time I realized that there was a discontinuity between what the market was TRYING to tell me and what my models were generating, it was too late.  I had to unwind many positions at a big loss. 

I also made some severe psychological mistakes like not following strict money management rules and getting caught up in the market hype.  I thought it would never happen to me but it did.  In the end, I loosely followed money management rules and leveraged myself too much.  I didn’t honor all stops and threw caution to the wind.  In the end, I got what I deserved.

However, a few of my models performed OK.  My market timing and volatility models held up surprisingly well in the face of this Armageddon. A few wheels came off of them when the $VIX hit 80+ but it managed to generate signals at all upcoming market inflection points.  The problem with the timing model was that it generated too many BUY signals, not necessarily at the wrong time, but not at the optimal time.  The timing model hit all the turning points in the market but it couldn’t differentiate which ones were the major ones, the ones where I should’ve moved more money into the markets.  

So I got trapped in positions that in a normal market pullback made sense, but not in a market crash.  As a result of this shotgun approach, my 401K is up 20% for they year, as opposed to the 60% that the market has rebounded too.  Oh well, I guess I’ll have to refine or rebuild those models. 

This learning lesson doesn’t mean that I’m abandoning neural net models completely, I’m just reassessing how they’ll fit into my overall trading system.  I’m also sticking with a trend approach to trading but severely simplifying it.  I’m not as smart as the market and why should I spend time building sophisticated trend models when I can just look at Price and Volume, and trade accordingly.  I will listen to what the market is telling me and avoid the hype in the news.  Money talks, Bullshit walks.

In the end, this market crash taught me many BIG lessons and that I’m taking to heart.  I’m a bit older and hopefully wiser now and look forward to a new year, filled with trading opportunities no matter if we’re in Depression, Recession, or Bull Market. 

In closing this post, I wanted to thank all my readers who stuck around.  My feedburner stats have keep relatively stable since I posted about closing down this blog.  I do plan on returning to posting in 2010 but the frequency and content is still up for grabs, and I plan on providing a separate subscription service (separate newsletter of sorts) in the future. 

Thanks for reading.  I want to wish each and every one of you a Happy Holiday season and a Happy New Year, may 2010 be profitable for you!