I completely disagree with your conclusion. If lethality can be mitigated by the players of the game to the point where it becomes extremely rare, then the game is not terribly lethal in the first place.
Does not follow. Cf. Player who can beat Contra w/o Konami code vs. Contra's inherent difficulty and Japan's low maternal mortality vs. childbirth's inherent lethality.
IOW, if the characters are rarely dying, then there is a low chance of death in that game. Why there is a low chance is irrelevant.
Why there is a low chance is, in Emerikol's game, the
whole point. We can flip this around by using a different failure state other than death. For example, in games where people prefer playing more of a heroic fantasy, or where character goals have more primacy than mitigating player death, their goals are often achieved, I daresay at a rate greater than 50/50. That doesn't mean that characters will achieve their goals automatically without doing anything, but rather that through their actions they mitigate the chances of failure.
The fact that it is the players who can mitigate this to the point where the game is rarely lethal leads me to believe that the game is more about following Emerikol's lead than anything else.
It is about following Emerikol's lead. But that's neither here nor there. The same effect can be found in games that rely on rules as mediation instead of GM mediation. In 3e or 4e, for example, encounter difficulty can be mitigated by optimized character builds and/or tactical combinations.
But, yes, this is completely besides the point. Pulsiver is in no way discussing difficulty in his blog post at all. He's comparing sandbox to linear games, calling sandbox games like Sid Meier games Interesting Decision Games and linear games Wish Fulfillment games and pointing to the fact that the majority of top selling games fall into the latter category.
I don't quite see it that way. "Linear vs open world" is actually mentioned as a separate way to look at design. He refers to "wish fulfillment" as "having an experience". The experience doesn't have to be linear, it merely has to be foregrounded to a greater degree than choices. In his paradigm, card games fall under "interesting choices" (and provide poor "experience"), yet game play is quite linear. Specifically, he calls out D&D and RPGs as "bridging the gap". That's because you can have interesting choices
and "an experience", regardless of whether you're playing a railroad or a sandbox.
Which rolls around to Emerikol's problems with his player. It sounds like the player is looking for a linear game where you know what's going on, whereas Emerikol is running a more sandbox game where the players are expected to drive the game.
I don't quite see how you come to that conclusion. Emerikol says (my emphasis),
The player I ran into basically said that he felt my game was too challenging. I kept the group too stressed out fighting for their survival. He enjoys a game that while still having combats and so forth is not one that worries him especially. The fun for him is using cool powers to do cool things.
Whereas, other players want to fight and claw their way to every goal. The fighting and clawing though is the real goal. The artificial reward goal is just symbolic of the completion of the journey.
It sounds to me that the player was seeking a different experience (Big Damn Heroes) from the rest of the group (Fantasy F'ing Vietnam) and/or sought different interesting decisions than the one Emerikol's game offered. That's entirely unrelated to sandbox vs. linear. I will say at the same time, though, that I don't necessarily agree with Emerikol that the problem was that his game offered interesting decisions and his player wanted wish fulfillment.