Plugins with few support threads penalized in search results.

[jsmoriss]

UNSTABLE plugins with many "resolved" support threads are given higher priority in search results.

STABLE plugins with few or no support threads (and thus few or no "resolved" support threads) are given a lower priority in search results.

Instead of the "resolved" number of threads being significant by itself, the RATIO of open vs resolved threads vs active installs should be the significant ranking factor instead.

[dd32]

Instead of the "resolved" number of threads being significant by itself, the RATIO of open vs resolved threads vs active installs should be the significant ranking factor instead.

Alternate opinion: The sheer existence of forum threads indicates that the plugin is more likely to be useful, as it can be a proxy for "do users interact with it" although #2686 would be better for that.

[gibrown]

Ya, 2686 (number of reviews combined with ratings) is better for weighting than a count of support forum threads IMO. Part of the thinking with focusing on resolved was that it encourages better support which should end up in a better user experience when they run into issues. It also isn't a super strong weight.

The wilson scores in 2686 requires some new fields on the index (I started something in D57367 on wpcom) and also figuring out how to reweight the search results to get it working (which is the harder and more time consuming part).