how much will the following imbalance skew the bayes algorithms (if at all)
Very little.. It will bias the scores very slightly towards higher bayes scores, but the chi-squared combining tends to make this effect not very noticeable unless the training imballance gets severe.
