How does it handle tied hits? What e.g. there are 5 docs with the same score and

_text_match(bucket_size: 3)

? Do they still get divided or will all with the same score go in the same bucket?