Hi,
The post content field nowdays is very complex, it stores all the raw data, including all sorts of shortcodes, blocks and HTML – which is stored as plaintext. Sometimes those include classnames, IDs etc.. where numbers are very common. My guess is, that a stored HTML tag has a class name involving “1200” either fully or partially. But since it’s a text there is no way to differenciate, so it’s matched.
The index table engine resolves that to some degree by rendering the content first then reading everything “visually” and storing that. If however the issue is resolved, then I would recommend keeping the setting as-is.