Our User Group discovered that the Stop Price used for stop ascending sorts is based on the Split-Adjusted Stock Price, not the Actual Stock Price. This distorts the back tests done with such searches.

Let's take an example. If you run stop ascending search on February 19, 2010, the number two stock that pops up on the list is VFC, and that is one of the stocks that the back tester would have bought on the following Monday if you started using this system on 2/19/10. If you graph it, you will see a Stop Price on that date for VFC of $17.47, which is the price used in the sort. However, VFC split 4 for 1 on 12/23/13, so the actual Stop Price on 2/19/10 was $69.88. In other words, if you were back in time and actually running this search on 2/19/10, the Stop Price you would have seen for VFC w ould have been $69.88, and VFC would not have been in the top 5 using the Stop Ascending sort. In fact, three of the top 5 stocks returned by my strategy search on 2/19/10 had splits subsequent to that date and thus had their Stop Price lowered for purposes of the back tester.

How does this affect back test results? Most likely it overstates rates of return, possibly significantly, since it disproportionately buys stocks that have a split in their future, and splits generally occur after a pretty good run up in stock price. Now if VectorVest could develop a signal that tells us which stocks were going to split in the future, we could all make a lot of money.

Please address this because it is a major programming issue that prevents actual, accurate backtest results using the stop ascending sort for any search.

Thanks
Bruce Berger