3 ms·

No detection mechanism is 100% reliable so better to ban 100% of the ones you know aren't false positives than risk them trying again and being a false negative

• codezero 12y ago
  Not really. Devaluing a single vote insitu once in a while shouldn't be that negative on the system as a whole, especially if there is suspected fraud. The problem with banning fraudulent users is that they are tenacious, thats's a big factor in their committing fraud in the first place, they are usually on a "mission," see: https://en.wikipedia.org/wiki/Wikipedia:Single-purpose_account https://en.wikipedia.org/wiki/Wikipedia:Single-purpose_accou... (not really the same as a fraudster, but this is a common archetype) They will just make another account or increase their efforts, better to keep them in their container within the system.
  • opendais 12y ago
    You seem to believe they won't to notice the effect on their "container within the system". They will eventually, just like a shadowban. At which point, they create a new account. > while shouldn't be that negative on the system as a whole, especially if there is suspected fraud. Assuming it isn't computationally expensive to do this on a per-vote basis. If you have to do any sophisticated analysis, that may not be true.