I think that deaths are good, useful tools to give a player a sense that their decision are important, when used in moderation. That said, I think it’s better for a player’s failure to end in losing something that they wanted, where possible. When a player is faced with a death, what do they do? They either a) give up (a very bad thing, as it means the story is not compelling them to continue working at it) or they b) restart. Well, what happens when they restart? They get a second chance, and just refine their choices to the ‘correct’ ones. This leads to a more linear experience where they just tweak what they do to ‘win’.
However, let’s look at the other option: Allowing a player to continue with a loss. Let’s break down what (generally) happens with the easiest solutions first, that is the loss of a stat. Let’s say the PC gets into a gun/sword fight and loses, but you don’t want them to die. One obvious solution in they get a ‘wound’ which lowers their fighting ability (and/or other stats). Well this only is useful so long as that stat is called up again, as they stats are only there to facilitate the story, so let’s look at what happens when we call the stat again: Either the player wins (via whatever way you want to give them, whether it be doing something to get over the wound, or using some unconventional tactics), or they lose.
Let’s break down what a loss means for a player first: A loss, later in the story, is better than a loss earlier in the story for two reasons: 1) because they have experienced more of the story, and 2) because they have played longer and are therefore (hopefully) more invested in the/their PC. So if they lose the second encounter, they get more out of it, but what if they win? Well, the obvious is that it’s more compelling to have won with a disadvantage, right? Think about it, how often in any given story is the PC/MC just an average person? It’s easier to connect to, and therefore is more likely to allow a player to project themselves into the roll. So any disadvantage a PC overcomes (within reason) is more compelling, right? Well, there is also another side, which is namely, the author of any interactive story only writes, at most, half the story. The rest is for the player to fill in. When you give a player a series of choices, they generally choose based on two critiria: 1) what they think is ‘correct’ and 2) what they what to be ‘correct’. When we can help these to line up, but removing an ‘incorrect’ path and allowing that to turn into a possible ‘correct’ path, we’ve made a player happier with the story (in so far as we don’t cheat a player out of a good story by making it obvious what we’re doing and breaking immersion). So when a player picks an ‘incorrect’ path that doesn’t lead to and end, I think it ultimately makes for a more compelling story.
Well, I think that’s long enough for the moment, but there is one thing I haven’t covered: decisions that are ‘incorrect’, but instead of giving a penalty to something later, give a loss to some specific story goal. That is a bit more complex of an idea.