The capability that games offer for self-expression is exciting, but can also be onerous. In much the same way as we are made uncomfortable by articles telling us what the way we move or dress or speak says about us, finding out, abruptly, that we are communicating when we had no intention of doing so can be a surprising burden.
We are told a number of things about our capacity for self-expression through play which are not, at times, entirely consistent. That is, games offer us a choice, or something which they say is a choice, but often refrain from supporting that choice through game mechanics.
This isn't inherently a problem. It's entirely possible to support player choice while still maintaining the possibility for the player to make bad choices. Indeed, it would be nearly impossible to do otherwise, and even if it were possible it will still be, in most cases, not even particularly desirable. The foundation of most gameplay in the first place is making choices, having them turn out well or poorly, and then understanding why those choices were good or bad.
A death wish is a viable choice, insofar as it is a choice, but should not be equally viable to others if survival is the end-goal.
Balance is something designers often discuss, but that term contains some implicit assumptions which are rarely brought to light. 'Game Balance' supposedly refers to setting up the game in such a way that no choice is more viable than another– except that obviously makes no sense, because if that was the case there would be no game– or at least not a very interesting one, just a variant on rock-paper-scissors.
In order for mistakes to be possible, balance cannot be absolute.
Which raises the question: Which choices should be balanced to be generally viable, and which are do-or-die mistakes?
Is the dividing line between the scale of the choices? The big choices should all be supported, and the small ones can either be valid or invalid on the basis of case and context? Sometimes. A good example of this would be in fighting games, where all of the characters are ideally balanced against each other so choosing one over another is never a mistake, but making individual decisions within the battle can be. This seems, anyway, to be what most fighting game players desire. However, in strategy games big decisions are frequently mistakes, such as choosing to invest all of one's resources into culture while a barbarian horde bears down on your capital city.
Is it how self-expressive that particular option is? Being able to choose your character, your history, your skills, your basic personality, these are all incredibly important to us. It's quite frustrating to be playing an RPG, find a skill you like, and find out later that it turns out to be completely useless in the late game. Then again, what about trying to play a pacifist, or a monk who fights unarmed? Surely these are expressions of character, and just as surely, in most games which don't explicitly support these styles, these would be incredibly difficult and suboptimal ways to play. But the players who embrace these styles don't mind... do they?
Is it how long choices take to come to fruition? Being punished for a choice which you made long before you knew what effects it would have is surely frustrating. In the strategy example from before, what if there were no invading army, but investing in culture was still, due to the mechanics of the game, a consistently suboptimal choice in the long run? No one would enjoy that... Would they?
It mostly all comes down to communication– but, then again, what doesn't? If we know we're making a suboptimal choice, then we don't resent the game when that choice doesn't work well for us. Balance discussions rarely happen around something which is explicitly considered to be a joke weapon or character. However, if everything the game tells us leads us to believe that this choice is supposed to be equally viable to its alternatives, then we feel betrayed and angry when that isn't the case.
Of course, that only addresses part of the problem. The thing is, even if you make those choices which are suboptimal explicitly suboptimal, that will raise some questions, or rather one question which contains many: Why? A good example of this would be, in a few games, choosing the sex of one's character. It's an unpopular design choice nowadays, for obvious reasons, but in some older games choosing a female character always results in lower strength (and, sometimes, to 'balance' this out, higher intelligence or dexterity or whatever). You could trot out whatever statistics you like about the average characteristics of the human body, but the fact is most games deal with decidedly un-average human beings, and often beings which are not human at all, so most arguments on that basis are suspect.
Now, character sex is a particularly touchy example, but the fact is that any time you want to make a choice explicitly inferior in some way, or even just explicitly different in any way, you're going to have to justify it somehow. This gets a bit tricky for a lot of games, since they regularly and intentionally divorce themselves from reality by having ballerinas out-wrestle 300 pound trucker-assassins, it makes it extremely difficult to make a convincing justification for any particular choice being suboptimal.
It's up to you to decide which choices within your game are vectors of self-expression, which choices are vital pathways that must be balanced if the game is to be fulfilling– and, conversely, which choices it's possible to just fuck up.
Not all choices are equal: If they are, they're not choices.