Big change in the handling of user-defined assets from Oct to Dec version...why?

I started using the results from annually-traded stock screens to define assets in ESPlanner Plus. In the October 2010 version I did an exploratory run using only an asset that had a real average return of 17.8%, a relative variance of 128% and a beta of 100%. In the December 26 2010 version, the same inputs gave recommended total spending of about 25% less. (I think the inputs were the same, but the database was corrupted by the upgrade and I had to create a new database and enter my data from scratch.) I don't know if you've created a major bug or fixed one, but the difference lowers my confidence in the current results. What can I do to find out what changed? I've compared the "inputs and assumptions" section for the two runs already. By the way, my standard-asset portfolios seem to be giving similar values to the October version in the new one.

Thanks,

Rick