Comments on: Stock NeuroMaster Video Tutorial 1 http://www.neuralmarkettrends.com/2008/03/27/stock-neuromaster-video-tutorial-1/ Using Rapidminer to model trends in the financial markets Thu, 04 Dec 2008 01:48:12 +0000 http://wordpress.org/?v=2.6.2 By: Tom http://www.neuralmarkettrends.com/2008/03/27/stock-neuromaster-video-tutorial-1/#comment-1847 Tom Thu, 27 Mar 2008 23:00:32 +0000 http://www.neuralmarkettrends.com/2008/03/27/stock-neuromaster-video-tutorial-1/#comment-1847 eric: I used about 15 bars of data that was OOS and in the 30 bar backtest the system did show a negative return. If you backtested the entire model the signals would've been profitable. The model tries to give you the best signal statistically it can but the market forces sometimes go against you so its always wise to have a stop loss. eric: I used about 15 bars of data that was OOS and in the 30 bar backtest the system did show a negative return. If you backtested the entire model the signals would’ve been profitable.

The model tries to give you the best signal statistically it can but the market forces sometimes go against you so its always wise to have a stop loss.

]]> By: eric http://www.neuralmarkettrends.com/2008/03/27/stock-neuromaster-video-tutorial-1/#comment-1846 eric Thu, 27 Mar 2008 17:05:26 +0000 http://www.neuralmarkettrends.com/2008/03/27/stock-neuromaster-video-tutorial-1/#comment-1846 hi tom, thanks for the tips! does it concern you that the model shown in the video is not making money? i find it difficult to build models on the indices that make money in out of sample (OOS) testing; building models with profitable OOS appears easier with stocks. on the other hand, your XOM model appeared to test well (was this OOS) and the trade itself seems to be working out well as well! hi tom, thanks for the tips! does it concern you that the model shown in the video is not making money? i find it difficult to build models on the indices that make money in out of sample (OOS) testing; building models with profitable OOS appears easier with stocks. on the other hand, your XOM model appeared to test well (was this OOS) and the trade itself seems to be working out well as well!

]]>