Problem you might encounter going live with your system

You work hundreds of hours developing a system. Test it hundreds of times using various markets, Monte Carlo computational tests etc. After six months of developing and testing you decide to go live and events you didn't  foresee begin to unfold.

Let me share with you two of these events:

1. The case of the missing symbol:

Last Friday I got a buy signal for "Greater Bay Bancrop" (GBBK). I had bought the stock and was eager to see how it will perform on the following week. Yesterday, as the market opened, I was watching to stock and something very strange had happen. It seemed the stock was not moving for almost two hours. I searched for news about the company and it appeared the company was bought by another company called "Wells Forgo & Company" (WFC). According to the news article, share holders of GBBK will get WFC shares instead. So far so good. Well, no quite.

As I use a computerized system, my buy signal was for GBBK. Now that the stock is not valid any more, the trade remains open and the system doesn't know it should be closed. I can close it manually by placing a sell order according to the exit rules of my system. Only problem is the system's exit rule doesn't apply to WFC as there is no "buy" signal for this symbol. I think I will just try to cover my trading costs for this trade and close the position.

2. Don't count on free historical data

Following a "buy" signal for "BioDelivery Science International Inc" (BDSI) I bought the stock a week ago. Today, while looking at the trading volume of the stock, something caught my attention. It seems like the volume was considerably low than last week volume. To be sure I run my system on the stock again and to my surprise, no buy signal appeared on the screen. I doubled checked the data and it seemed to be correct. My only suspicion was that while I updated the historical data for my stocks this morning, I noticed some stocks had incorrect data for the previous week. After running the update for last week for the entire list, the problem disappeared. I'm guessing that the historical data for BDSI was incorrect as well, and as a result I received a false buy signal.
The moral from this event is to check and double check the historical data of a stock when a buy signal is received.