There are several competing considerations in a DRV debate. The interest in finality is substantial in a DRV of a deletion debate - in most cases, substantial community time and energy have been spent already. On the other hand, the overriding interest in building an encyclopedia is paramount, and in many cases overrides the interest in finality.

Thus, to me, the question at DRV is not what I would have !voted in the first place, not even what I would have done had I been the closing admin, but whether the closer exceeded the bounds of their discretion, a demanding standard. Therefore, I will !vote to endorse a closure, unless:

  • The closer committed a significant procedural error, for instance:
  • Inappropriately closing a discussion early,
  • Misinterpreting the relevant guidelines and policies, in a way that no reasonable person would have interpreted them,
  • Deleting a page out of process,
  • Wholly failing to explain the rationale when requested;
  • and the error was material: either
  • there is a reasonable probability that, but for the error, the result would have been different;
  • the error obliterates my confidence in the discussion or the closer's assessment of consensus, such as a substantially malformed listing or a conflict of interest; or
  • The closer's assessment of the consensus was clearly erroneous. That is, the closer's conclusion must "be more than maybe or even probably wrong", but must "strike[] [me] as wrong with the force of a 5-week-old, unrefrigerated, dead fish." [1];
  • New developments since the closure call into substantial doubt the result of the previous discussion. I would normally !vote relist in such cases; or
  • In cases involving serious irregularities in the AfD process, the closer committed clear error in handling the issues.

In addition, I regularly !vote to speedy close or speedy endorse in cases involving

  • blatantly unfounded attacks against the closing admin, an AfD participant, or a DRV participant;
  • actions that demonstrate a manifest disregard for Wikipedia policy or the DRV process, such as the nominator unilaterally recreating a deleted article while the DRV is open.

Explanation for these very narrow criteria:

  • DRV is not AfD redux. Once an AfD discussion is closed, the question is limited to the procedural and substantive propriety of the closure.
  • Procedurally, since Wikipedia is not a bureaucracy, there is no point in !voting to overturn unless the error was material.
  • Substantively, the issue is whether the closer properly read the consensus. But whether an assessment is proper is in the eyes of the beholder. Assessing the relative strengths of the arguments at AfD, especially lengthy and contentious ones (those, after all, are normally the ones sent to DRV), involves delicate weighing and balancing. Respect for my fellow admins, the great interest in finality, as well as recognition of the fact that I am not infallible make me reluctant to second-guess a closure simply because I might disagree with the closer's conclusion.
    • For example, if, in an AfD on Weapons to be used in World War III, there are four !votes to delete invoking WP:CRYSTAL, and four !votes to keep saying WP:ILIKEIT, and the closer closed as no consensus, that would be a clear error in judgment and I would !vote to overturn.
    • But, if, in an AfD, there are 6 !votes to delete, 5 !votes to keep, each side presented reasoned arguments grounded in policy, and at first glance no side seems to be particularly strong or particularly weak. The closer carefully assessed the strengths of the arguments and decided that the !delete arguments are stronger, explained themselves, and closed as delete. I would endorse the outcome, regardless of how I would have read the consensus.

Notes

edit
  1. ^ Parts and Elec. Motors, Inc. v. Sterling Elec., Inc., 866 F.2d 228, 233 (7th Cir. 1988).