# What should we do with challenges with unfortunate win conditions?

This has now happened three times in recent history:

All questions have technically correct and unambiguous win conditions. However I strongly believe that their way of scoring is anti-fun.

In all of them it's because the question involves two challenges:

1. Solving the problem.

Actually solving the problem is hard in of itself, and fun. Because solving the problem is hard, the code to solve the problem is going to be big. Then having to golf that big code is not fun. The amount of work and juggling of parts required to optimally golf a piece of code exponentially increases.

However, the scoring requirements only rewards the non-fun part, and you effort in actually solving the problem goes totally unrewarded.

For me this means that for these questions I lose interest after actually solving the problem, so I just post my huge ungolfed solution, or not post my solution at all if there's a shorter one. I'm sure I'm not the only one feeling this way.

### Because actually solving the problem is so hard, the scoring feels arbitrary, and optimizing your answer for the best score is anti-fun.

• Well there's always the option to make it fastest-code. But ultimately it's the OP's choice if he thinks golfing or optimising will be more fun. I'm also sure that there are several people on the site who do enjoy golfing challenging problems. So I'm not really sure we should do anything with challenges with "unfortunate win conditions", because that sounds like a very subjective thing to me. – Martin Ender Apr 21 '15 at 2:56
• That being said even problems which are hard to solve don't necessarily result in convoluted algorithms. Sometimes there is a really elegant and golfy solution that's just really hard to come up with. – Martin Ender Apr 21 '15 at 3:11
• I posted Pearls as a golf because there was not (IMO) a more satisfactory criteria. I guess I could make it a timed optimization challenge, but that would require a bunch more test cases, or much larger ones. I agree that golf isn't optimal for the challenge (and am open to ideas, at least until answers show up), but it was the best of a "bad" bunch to my eyes. – Geobits Apr 21 '15 at 19:07
• I agree that for some puzzles (mine included), it's probably more fun to "just solve" than solve + something. But since challenges need a winning criteria, that's not really possible. – Geobits Apr 21 '15 at 19:09
• @MartinBüttner, stackoverflow.com/questions/1433263/code-golf-decision-tree is a favorite example of mine. Many of the answers ended up translating the input to perl/ruby/python and evaluating – gnibbler Apr 21 '15 at 22:11
• I really didn't like my Klotski being a code-golf either. However, I didn't want to make it fastest-code, as that would require me to either test it or make it purely algorithmic-based. Potentially we could start using popularity-contest for these? – Nathan Merrill Apr 22 '15 at 16:57
• I really like these kinds of questions, and I had a great time working on Pearls and one about lasers and mirrors. I didn't complete my Pearls solution because I hit a wall and had other distractions, not because it wasn't fun to think about or golf. Maybe language of choice has a significant impact (2000bytes of semi-golfed C# tends still to be very readable, perhaps this isn't true of golf languages, and personally I wouldn't go near more than 20 lines of python). I also don't feel as compelled to mindlessly golf these things, as a (reasonably compact) solution in of itself is postable. – VisualMelon May 4 '15 at 6:51

Communicate your idea to the author via comment or chat ping. Modifying the criteria of a valid challenge against the author's will is not an option in my opinion. (If the challenge is invalid, then that is a totally different story.)

If the author rejects or ignores your idea, and if changing the scoring system would significantly change what constitutes an acceptable answer, then you can post your own challenge with the modified scoring criteria.