In Reed-Solomon codes, the symbols of a code word contains multiple bits. Since the error correction and detection happens at the symbol level, it doesn't matter how many errors there are within the same symbol, it only counts as a single symbol error. Because of this, Reed Solomon codes are considered to be a great candidate for transmissions that are subject to burst errors.
On the other hand, BCH codes are always considered to be a good candidate for random errors.
But is that also true in an apples to apples comparison where you compare total number of message and parity bits (not symbols!) are the same for a Reed-Solomon code and a BCH code? Would there be any way in which the BCH would lose against the RS code in terms of error correcting capability?