Look, this discussion can go one of two ways depending on your philosophy towards life.
On the one hand, you can be from the school of hard knocks, where failure is expected and grit is chief among the key ingredients to success. That road to success is paved with abuse, setbacks, put-downs, failures, and outright disasters. A common sentiment to those near the top of the mountain (roughly Base Camp 1 in Everest terms) seems to involve throwing rocks and hurling insults down to those slowly making their way up the road. Their predecessors on the road to mastery did the same; why should they not avail themselves of the opportunity, take some comfort in the elevated position in which they now find themselves? It pays forward the valuable teachings these people once learned, and that lesson must be ingrained in the next generation, even if it takes a rock to the head.
Another possibility are those who prefer a gentle, guiding hand, one that seeks to elevate and inform, not to insult. This involves censure only when dangerous mistakes are made, to make absolutely clear the importance of never repeating it.
For me, video games that fall into the second category work best. This applies not only to the gameplay structure but also to the community surrounding the game. The fighting game community and I strongly disagree for that reason. For an example of a highly-regarded philosophy of fighting games (and it uses other games as examples), see Sirlin’s Play to Win. One comment: I find the author’s tone insufferable and deeply disturbing in its lack of empathy. I’ve tried to read that book a few times and could not get past the pseudo-intellectual presentation (especially in the frequent appeal to authority in the form of aphorisms and quotes). Seek at your own risk.
In this vein, I prefer games that make it possible (though somewhat unlikely) for a skilled gamer to finish it on one playthrough. A player that has mastered the basic skills should be able to demonstrate that mastery through play. Certainly there are games that make failure an intrinsic part of the gameplay, such as roguelikes. Particular examples of roguelikes, like FTL, seem to display a level of balance and give sufficient opportunity to experiment that offsets the common end for most playthroughs: a swift and sudden end due to a poor roll of the dice.
By contrast, games that punish you from proceeding by forcing you to fail the first time don’t feel quite as engaging to me. Despite a generally favorable experience, I’m finding Hollow Knight irksome in this regard. This Metroidvania title features the usual hallmarks of a great indie title: a sweeping, atmospheric soundtrack, distinctive art, a unique story unfolding through text, and some pretty poor performance. However, for several of its bosses, the handful of attacks each boss features forces the player to customarily die at least once before understanding all the patterns. It’s not really the setup per se; what I recall of Mega Man games, as one example, didn’t always make it so difficult to beat bosses on the first try. It’s a mixture of the speed of attacks, the relatively short invincibility period before your character can take more damage (though there is an item to extend that), and occasionally the surprise appearance of boss or mini-boss encounters.
It’s not really a bad thing in the larger sense of the game’s design. Your penalty for death is a temporary loss of accumulated wealth and a trek back from the last save point you chose to use. You essentially lose time, but it’s something about my bruised ego that I think is at the core of why this set of design decisions leaps out at me. I don’t feel a sense of loss when I die, merely annoyance. I don’t feel that I’ve surmounted an obstacle when I finally defeat the antagonist. What I feel mostly is that I’ve been set up to fail, that I was not given a reasonable opportunity to exercise my skill at the game. There’s something about the skill progression in roguelikes where this situation doesn’t arise. I think it has something to do with making expectations clear to the player; you will die a lot. However, you will learn to counter these obstacles and more.
Hollow Knight doesn’t seem to follow a similar progression. One might learn to navigate the labyrinthine underground areas fluidly, countering enemies easily. New abilities will expand the map and grant access to new areas. However, the moment you encounter a new boss battle, you’re probably going to get your ass handed to you until a round or two to establish the four attack patterns of the given boss, at which point you’ll wrestle the controls to dodge attacks sufficiently well to land your own attacks (seriously, I’m pretty sure the controls have either not responded occasionally to my input or done the exact opposite of what my inputs should have done, like dashing left when I’ve clearly held the joystick to the right). This difficulty deepens with the “dream” variants of bosses that become available to challenge after a boss’ defeat.
In contrast, the staple of the Metroidvania genre (and the ‘vania portion of the pairing), Castlevania: Symphony of the Night, subtly hints at a boss battle by providing an obvious Save Room before the encounter, allowing the player to prepare and even choose to not to fight the boss just yet. Even then, I seem to recall ample opportunity to learn a boss’ behaviors within the first encounter and thus counter, as opposed to being killed outright, which is an opportunity I don’t feel afforded in Hollow Knight.
On the face of it, I can admit that maybe I’m just not as good enough at the game as I think I am. It’s also possible it’s an intentional part of the design of the game in an effort to balance out the difficulty and not make the game feel too easy. That said, I’ve played a lot of Binding of Isaac and FTL, with a fair side of Demon’s Souls thrown in for good measure, and I never quite felt as annoyed at this kind of encounter in those games. It might be an effect similar to that of the latest Prince of Persia game with regards to failure in that game. The game designers in that case argued that a system of “lives” followed by a “game over” screen with a reload from a save was just too time consuming. Their solution to this was to make retrying an encounter in the game, both in exploring the world in the game’s signature parkour style or in fights with enemies and bosses, diegetic as opposed to an artificial “video game” trope. In cases where you fail, your magical companion will pull you away from the danger and give you another chance to complete the fight or the obstacle. While functionally similar to the autosave where you reload from the beginning of a designated area, something about the structure of this setup made the cost of failure seem too low to provide a challenge to players, even if in regards to behavior and time lost, it was nearly identical.
Perhaps it’s a matter of dashed expectations more than any perceived loss or inconvenience. I expect to have a fighting chance after many hours of play, and my hopes are often dashed with the first encounter with a new boss. Again, it may easily be possible that it’s just that I can’t do it, or for some it’s perfectly acceptable to die the first few times against a boss character before “getting” it.
I think it has more to do with the way a game presents adversity to a player, with a reasonable growth in difficulty over time, allowing the player to adapt and feel accomplishment. While it may make some sense to erect a wall in front of the player at some points (and it might really be more realistic in terms of capturing adversity in life), as entertainment it often is the point at which the player no longer feels engaged or challenged by the game. In my book, that’s more a failure on the game’s part than the player’s part. While it’s impossible to cater to everybody’s tastes and behaviors (and an argument can be made that an attempt to do so only leads to safe, bland, junk-food equivalents of games), an attempt should still be made to a degree to ensure the game feels like it’s going somewhere for the player.)
Or it could be as simple as this: a game with a narrative and high difficulty are inherently incompatible. Narratives do not traditionally end and restart when a character dies, especially the main character. Certain games like Heavy Rain try to continue the narrative after a player character perishes, though players may still feel they’re “missing out” on something when this happens. This might have been the intent with the aforementioned Prince of Persia; unify the gameplay mechanic of difficulty and failure with a narrative element that introduces consistency across both.
Aside: After writing this piece, I changed my opinion about the Hollow Knight controls to a degree. For the PC version, a new option was added to use an “experimental” controller API, that does seem to reduce latency somewhat. Also, the Nintendo Switch version of Hollow Knight seems distinctly easier to control than the standard PC version (without the experimental controller API flag enabled). The game has been somewhat more enjoyable as a result.