Do we want weekly challenges?

Question

From this post on the Mother Meta:

A handful of sites have conducted a variation of the "weekly topic challenge" idea. I, myself, have become quite fond of them since they do seem to be effective at increasing asking rate and can be a lot of fun. I've personally initiated challenges on several sites (most recently on History and Philosophy) and have worked out a simple process.

The first step mentioned is to ask for topic ideas, but first we have to decide whether we want these challenges at all.

Here's a summary of how it would work:

1. A meta question is posted asking for topic ideas. You can suggest as many ideas as you want.

2. Every week, a topic is picked to be the subject of the week. (Other sites have done biweekly challenges instead, but PPCG is reasonably active enough to have one per week.) This would most likely be done by picking the highest voted submission.

3. Results are tallied, rinse and repeat. (The results are meant to analyze how the weekly challenge went, and will include how many posts were generated as a result of it, how many votes and answers they got, etc.)

Since PPCG is more of an... unconventional Stack Exchange site, the weekly challenge wouldn't be asking for just questions. It could ask for a certain type of programming challenge, perhaps...

• one that you feel is underrepresented on this site (such as boggle)
• one that fits a certain theme (we could have winter-themed questions in December or January)
• one that you want to see users here work together on optimizing or golfing further
• or just something that you think would be interesting.

This could be effective at increasing our low and ever-decreasing question volume, but then again, quality != quantity. So what do you think? Post your thoughts on this below, and if most of us agree that it'd be a good idea for the site, I'll post the meta question mentioned in step 1 and the challenges will start shortly (probably on Friday, as suggested by the MSE post).

Answer 1

In principle, I think this is a nice idea. However, if we take the concept without any modification, it might not work as well on PPCG as it does on other sites. Some general concerns:

Sandboxing

As you said, quality ≠ quantity, and if we just tell people "hey, pose loads of boggle challenges this week", I think the results are going to be rather poor. Therefore, these challenges (which are posted for the sake of increasing question volume) would probably benefit from the sandbox even more than usual. So I suggest we should take sandboxing into account for the schedule. E.g. once the topic is decided, people get a week for coming up with challenges and sandboxing them, and then they are posted the following week when they're ready. This should help weed out duplicates, generally bad specs and questions that just weren't well thought through.

Themes wear off

I'm not sure if having several questions about the same topic/concept/theme in a short time is a good thing on PPCG. It doesn't really matter on other sites, but people put a lot more time into answering some questions here than on other SEs, so unless you're really into the current theme, it might wear off, so that people get enough of it after one challenge. We definitely need to keep the topic generic enough to allow for a fair bit of variety.

Quality first

Peter suggested in chat that something like "a challenge a week" would be a better fit for PPCG, and I think he has a point. (Disclaimer: what I'm suggesting below is completely unrelated to what he had in mind, but his comment sparked the following idea.) What if, instead of having a graph-theory week, where we get a surge of graph theory challenges, we take the week to collaboratively work on an excellent question on the topic and post it the following Friday. We could tackle particularly tricky specs, challenges that need controller programs, or even test out some new territory.

This would definitely emphasise quality, and not do much for quantity. But if a bunch of people are working on a challenge, throwing ideas around, I'm sure there'll be enough material that doesn't make it into the challenge but could be reused in the future (or even in parallel, if it's sufficiently different).

This is how the time frame could work:

• Week 1: We collect ideas for a theme or a vague challenge concept. Because this is intended for only a single challenge, this can (but doesn't have to be) a bit more specific. Like "a King-of-the-Hill surrounding genetic algorithms", or "an optimisation challenge for [mathematical problem X] in hopes of getting some interesting new results".
• Week 2: A theme/concept is chosen from the suggestions and those who are interested get cracking on sandboxing it, sorting out the details of the spec, writing a controller if necessary. Of course, if any additional challenges are conceived of in the process, those should also be sandboxed by all means.
• Week 3: The challenge is posted (we'd have to decide by whom, but I'm sure we can agree on something) and run, and at the end of the week, we do an evaluation of how well things went.

And then, repeat. In fact, these three stages could be pipelined. While we're working out the details of one challenge, we could already vote on the next round, and work that one out while the first one is being run on main.