The default sorting order is Daily installs, but it's meaningless for the vast majority of scripts on the website. Even for the scripts that have been installed within the last 24 hours, 9 times out of 10, the error margin is larger than the value. This essentially makes the sorting random.

The other download counter, Total installs, has the opposite problem. Any popular script from ancient times, which hasn't been maintained for the last 8 years and has been broken for longer than that, can easily be at the top.

I think Weekly installs and Monthly installs would be much more relevant stats.

Case in point: https://greasyfork.org/en/scripts/by-site/google.com?filter_locale=0&page=2 — Daily installs is 0 on page 2 out of 16 pages, and even on the first page, only 6 scripts have more than 10 installs.

I understand what you're saying, but I don't believe the r###lts would be significantly improved by going to by week or by month. If the first few aren't what you're looking for, you're probably gonna have to put in some search terms.

Going to weekly/monthly would just reinforce the "most installed script gets listed first and so gets more installs" feedback loop.

If what you're looking for is fighting the "what gets to the top, stays at the top" problem, then it's a known problem with a known solution.

All you need to do is to give more weight to the recent votes. Typically it's called "sort by popularity", "hot" etc. and is the default order. It can be something like #### * week1 + 512 * week2 + 256 * week3 + 128 * week4 + 4 * month2 + 2 * month3 + 1 * month4 + remaining / ####. You can experiment with the multipliers and see what works best.

Sorting by popularity solves all the problems at the same time:

• gives more visibility to the recently uploaded scripts;
• keeps popular scripts close to the top, but not at the very top;
• provides meaningful order for scripts with 0 recent downloads.

It would actually be fantastic if you implement this.