First of all - wow! Those results! That majority! And I voted in that majority, but even I'm surprised how uniformly we all voted, given how a lot of those options are
reasonable ways to design combat. I do think there's some range of "good combat design" depending on the tone of the game and the intended audience.
What I love most in combat is battles throughout that provide a difficult but
fair challenge - I really like losing some battles along the way (even if it provides me with an annoying Game Over) as long as
I feel that I could have done something differently to win that battle. That's the essence of fairness - if I lost, it's my fault and not because the game handed me a bad beat and threw me something impossible to overcome.
Star Ocean 2 is probably the game that's done this the best in my eyes - the high general difficulty and occasional wild difficulty spikes meant that I would sometimes encounter half a dozen Game Overs just trying to get from one town to the next, but nearly every time it happened I said to myself "if only I hadn't gone in and attacked..." or "I should have been able to dodge that". Winning these random battles felt really, really good. Each one felt better than beating a boss in most RPGs.
It's easier to do in "skill" based systems like Star Ocean's ABS or even something like a Match-3 battle system, but even turn-based systems can inspire this feeling. Persona 3 definitely did, at times - your ability to know when to play attack and when to play defense often meant the difference between survival and defeat, and if you did lose (which was annoying because you lost out on a few levels of progress on Tarturus), you usually learned something about the enemy so that you could play it different the next time and win, even without any extra grinding.
There's also the challenge inherent in some games (moreso the past, less nowadays) where each battle provides minimal risk in itself, but making it through the entire dungeon with enough resources left to take on the boss represents a genuine challenge. This is not my thing but it can be really good game design when done well.
What is
not good challenge is for enemies to randomly use skills that are powerful enough to kill your party members, and you can't to anything to predict, dampen, or counter it.
===
One final thought - I found these two (similar) options very interesting:
- They should be a puzzle for me to solve. Once I know the solution, they should be fairly easy to beat. (4 votes [20.00%])
- They should have a hidden weakness or gimmick that I can exploit to easily beat them. (0 votes [0.00%])
I really like when games do these with some or most of their bosses - however, and this is really important - a player who doesn't figure out the puzzle/gimmick needs to be able to beat the boss anyway with skillful ordinary play. If everything you throw at the boss does 1 damage until the player figures out that you need to break the glowing gem in the ceiling, the player who doesn't recognize this is never going to be able to progress. That is bad!! It's much smarter to do something like give the boss double the normal HP, and if the player "solves" the gimmick and recognizes that the gem is powering the boss, drop the boss' defense down to an extremely low level so that the player can finish off that boss really easily.
If you insist on making it "unbeatable" without solving the gimmick, start to give the player hints as they hang tough in the battle without solving the gimmick. After two minutes or five turns, have a party member shout that there must be something around giving the boss all that endurance. After ten minutes or twenty turns, have the party member "find" it for you! "Hey! I bet it's that glowing gem on the ceiling! Let's take it down!" This can feel slightly condescending and make it obvious that you're railroading the battle, but it's still a lot better than leaving the player completely stuck and frustrated!