Interesting range of responses so far, a lot of it simulationist. I'm going to look at it from a gamist perspective instead. In brief: succeed 65-70% of the time; have something interesting to do 100% of the time.
1- Trying something and failing is no fun, but always succeeding isn't fun either. So: Everyone should succeed 65-70% of the time when they're doing something their character is good at. Fighters' swords hit 65-70% of the time. Wizards' spells hit 65-70% of the time. Rogues save the party from deadly traps 65-70% of the time. Etc.
2- Waiting patiently for your turn, then whiffing, is no fun either. So: Everyone should have something interesting to do on their turn 100% of the time. Fighters can move tactically 100% of the time, although they may choose not to. Rogues can notice tantalizing details about their surroundings 100% of the time, although they may not think it's important at the moment. Wizards can apply their spells in creative ways 100% of the time, although they may not be feeling creative at the moment.
An interesting corollary is that classes aren't set apart by how often they succeed. Instead, the "feel" when playing a character depends on what they can do that's interesting and the "feel" when watching someone play a character depends on what they're particularly good at.
So you see the fighter, wizard, and rogue, and you think: "wow, that fighter's really good at hitting enemies. The wizard's really good at breaking out the bag of tricks. And the rogue's really good at exploring." But they all have the same chance to succeed... just at different things.
And you when you choose to play the fighter, wizard, or rogue, you think, "I really love the fighter's tactical approach," or "I love planning my spells for a day and coming up with creative strategies," or "I love finding clues and solving puzzles." But they all have something interesting to do all the time.
The other corollary is that challenges should scale, but chances to succeed should not. So a level 1 critter might have an AC 10, 4 hit points, and do 1 point of damage (a minor threat to a level 1 character), and a level 10 critter might have an AC of 10, 60 hit points, and do 15 hit points of damage (it hits your level 1 character, you instantly die).
And this should be true for all three pillars, for every class. Always something interesting to do in combat, exploration, and interaction; and always something to succeed at 65-70% of the time in combat, exploration, and interaction.
1- Trying something and failing is no fun, but always succeeding isn't fun either. So: Everyone should succeed 65-70% of the time when they're doing something their character is good at. Fighters' swords hit 65-70% of the time. Wizards' spells hit 65-70% of the time. Rogues save the party from deadly traps 65-70% of the time. Etc.
2- Waiting patiently for your turn, then whiffing, is no fun either. So: Everyone should have something interesting to do on their turn 100% of the time. Fighters can move tactically 100% of the time, although they may choose not to. Rogues can notice tantalizing details about their surroundings 100% of the time, although they may not think it's important at the moment. Wizards can apply their spells in creative ways 100% of the time, although they may not be feeling creative at the moment.
An interesting corollary is that classes aren't set apart by how often they succeed. Instead, the "feel" when playing a character depends on what they can do that's interesting and the "feel" when watching someone play a character depends on what they're particularly good at.
So you see the fighter, wizard, and rogue, and you think: "wow, that fighter's really good at hitting enemies. The wizard's really good at breaking out the bag of tricks. And the rogue's really good at exploring." But they all have the same chance to succeed... just at different things.
And you when you choose to play the fighter, wizard, or rogue, you think, "I really love the fighter's tactical approach," or "I love planning my spells for a day and coming up with creative strategies," or "I love finding clues and solving puzzles." But they all have something interesting to do all the time.
The other corollary is that challenges should scale, but chances to succeed should not. So a level 1 critter might have an AC 10, 4 hit points, and do 1 point of damage (a minor threat to a level 1 character), and a level 10 critter might have an AC of 10, 60 hit points, and do 15 hit points of damage (it hits your level 1 character, you instantly die).
And this should be true for all three pillars, for every class. Always something interesting to do in combat, exploration, and interaction; and always something to succeed at 65-70% of the time in combat, exploration, and interaction.