Filters are only as unintelligent as the people who program them.
For spam on twitter to get results, it would seem to require meeting a couple criteria:
1) Unless the person you want to spam is following you, it has to be directed @somebody so it will show up in their mentions, or the target of the tweet will never actually see it.
2) An actionable link for the user to click-on once they see the tweet.
So, there are literally billions of messages sent on Twitter every day. An enormous percentage of them do not include an "@", which means you can almost certainly discount that tweet as spam. So the "sameness" thing really comes down to the URL... So how hard would it be to write a rule that says:
"If person XYZ posts more than XYZ tweets @somebody in XYZ period of time and all of the tweets lead to the exact-same (or less than 10% different) URL, its likely spam."
Answer: It wouldn't be that hard at all. And is 100% necessary, as the decade-plus-long failure of various "sue spammers" campaigns can attest to.
Yes, by all means sue the bastards. But don't expect the judge to solve the whole problem in perpetuity throughout the universe--instead use the judge to extract a penalty from the spammers after the fact.
In other words, it isn't "either/or" but "both" that you require for an effective solution.