Playing with my encoder

Doing preliminary research and I use my Huffman  encoder to measure information content.  Not ready for prime time. I just started looking at information content in stock prices for the SP100.  I take all stocks and reduce the resolution of the prices.  Run them through the encoder,  and look at the amount of  coding work.  Then raise the resolution, repeat.  I am looking for the point at which any increase in resolution yields no better information content.  This is a simple minded exercise in channel simulation, sort of a first glance. I am looking for the minimum set of channel components that describe the corporations in the index without mutual interference.