Assume the prices appeared in a time period are normally distributed. Then we can model the probability that a price goes over the take profit line. On the other side of the distribution is the probability a price goes below the stop lose line. The offset is to specify skewness of the distribution lean toward the take profit side. If the offset is 0, this means probability of movements to either directions are distributed equally, and we really just have 50/50 chance on either side.