How to identify partisan voting groups and effectively get rid of the continuous problem?

Question

Let us say for instance, there is a community which has a problem of it's voting system being abused by the users. Users who come together and vote on each others posts or posts of a certain topics only. Such a group would start out silently, so when the moderators realize something suspicious is happening, some damage has been done.

When the moderators deal with the issue with whatever tools they may have at their disposal, another different partisan voting group appears same as the first. Let us just say this community has a problem with partisan voting. So how can the moderators identify when partisan voting starts (from the beginning)? and how (if possible) can they take actions to prevent future partisan voting from happening?

If the above is not possible, then is there any way the moderators can minimize the damage done by these groups (current and future ones) to the community?

My Question is referring to a StackExchange-like setting.

Answer 1

Your question goes in multiple directions, and has a lot of ways where you might hit system boundaries.

Voting on each other's posts

This is usually very clear voter fraud, intended to directly benefit the voting ring - in a Stack Exchange like setting that usually even means reputation/score gain. Assuming there is no system in place to catch such scenarios (such as SE's serial voting check, which probably would not catch such a situation because it involves too many people), you should definitely act. Start by talking to the users in question, show them the pattern that you noticed, and make sure they know that gaming the system is not okay.

If they keep going, and you can clearly say that their behaviour is gaming the system, the disciplinary measures should start to kick in: If possible, undo their score gain/votes, make sure that is reverted and they are informed as to why that happened. Escalate the punishment over time: Undo the damage, suspend, suspend for longer.

Voting only on certain topics

I think there might be some deeper idea behind this, but if it's just that - voting only on certain topics - that should not be your concern. Users should be free to choose what and how to vote, as long as the votes are genuine and honest (rather than the above exploit scenario).

If this behaviour is actively disruptive to the community (say, a controversial reddit post getting upvoted despite it being completely unfit for the subreddit), you should first evaluate what they are voting on. Maybe the post itself is the problem, and the support it gains is just a secondary issue. You can, however, talk to the people involved just as well.

Noticing this behaviour early and preventing it

I know, I've been saying this a lot: "There isn't much you can do here." Assuming you have no control to change the system that tracks these votes, you can't really find these voting rings if you don't happen to stumble upon them (or they trip some other mechanism by accident). You can educate your users to be vigilant and look out for suspicious voting, you can try to deter people from that behaviour with PSAs, you can even try to keep tabs on it all yourself, but at the end of the day something will slip through the cracks - that's part of the point of an exploit, after all: not getting caught doing it.