Update: Stop Losses
After the late July/early August waterfall decline, it became obvious that some sort of stop loss criteria was needed to protect capital and prevent huge losses. After some testing and research, the best choice was a partial stop loss when my model flips modes in the opposite direction of the current position. For example, if the system is long but flips to unhealthy mode, 50% of the position will be stopped out at the closing price on the day of the flip. Similarly, if we are short and the system flips to healthy mode, half of the position will be stopped. The other half of the position will be given the opportunity to ride it out until the next signal.