We roughly estimate the p-value of a recombination as p_V + p_J. But... this is not a probability.
Something such as 1 - (1 - p_V)(1 - p_J) is a bit better (and it's still an estimation), but... it's almost equivalent when these probabilites are small.