The whole goal of combat in B/X, and by extension, OSE, is NOT to fight fair.
Okay. How is this relevant? Players making mistakes is an expected part of play. The notion articulated by the OP is that players having the ability to survive
small mistakes, but still suffering if they make major/severe mistakes, is a net positive for the experience overall.
If players are dying because they keep rushing headlong into fights, the lesson they need to learn is to plan the battle and stack the odds in their favor. <snip> And if the players are just outclassed by opponents or monsters that are too high a level for them, they need to learn how to run away and come back later.
Or the DM needs to get better at providing useful information to the players such that they actually understand the threat level they're expecting to face (as you noted in the bit about traps, which I have removed.)
It's a two-way street. I've seen WAY too many DMs who think their "subtle hints" are more than enough when they absolutely, emphatically ARE NOT, and only a mind reader could actually divine the intent from what they signal to their players.
That’s why the best settings are sandboxes:
As already noted: the best
for you.
so players can go where they want, find out where the more approachable dungeons are and skip the harder ones until they’re higher level.
Sandboxes are not the only way to do this. I do not run a sandbox game, yet every single thing you've just said applies to at least parts of the game. The only major difference is that I actually make sure to give the players real, recognizable opportunities to
figure out whether the danger is "approachable" BEFORE they go there, rather than the typical "hard" sandbox game DM, whose allegedly "subtle hints" include weird NPCs with unusual features (missing limbs, grizzled look, whatever) or vague signals easily misinterpreted for mere background color.
And sometimes, the only way to figure out which dungeons are the hard ones is by going in and getting jumped by something mean and running away.
The problem is, "getting jumped by something mean" almost always means "instant death" in older-style games. So there is no actual learning that occurs. You simply die, and have to start from scratch. Over. And over. And over. And over. And over. And over.
That's why funnels were invented--to skip past the tedious, time-consuming process of dying over and over and over, while still
having the "earn your success" experience. There are other game design solutions as well, if one wishes to have a different default state--as noted in my first post in this thread.
Sure, bad rolls happen, but good planning and role-playing trumps bad dice rolls all the time.
Except they don't--
in older-style play, where incredibly swingy dice tend to dominate, something numerous fans of the style have explicitly told me repeatedly. Indeed, it's the whole point! That's what you
mean when you say "stack the odds in your favor"; what you are actually saying is, "never actually FIGHT at all, but instead destroy without triggering the fight rules." Which some--note,
some--people find really fun! Most people, however, find that that gets really boring after a while, particularly because a very large number of players do not have DMs who are endless fountains of creative beyond-box thinking, so things tend to boil down to one of a handful (at best, half a dozen) old reliable tricks.
That, that right there, is literally the reason things like ear seekers, cloakers, "gotcha" cursed items, etc. exist in older-style D&D. The DM cannot be infinitely creative, but they're engaged in a DM-player arms race. Hence, they must at some point engage in impossible-to-foresee "gotcha" stuff, resetting the playing field until the players develop a new SOP which cuts through that particular gordian knot. And the cycle repeats, over and over and over.
This is why a number of other players have more fun with something where the rules are clear and
not meant to be constantly subverted. Instead of the rules being a deathtrap to be avoided, they become an interesting obstacle to surmount. Instead of being a
puzzle, where (because DM creativity is finite) there are usually only a small number of actually productive solutions, it becomes a tug of war, where it is actually possible to
outmaneuver the enemy, rather than simply obliterating them or getting obliterated because you couldn't figure out how to obliterate them.
In the absolute ideal case, where it really is true that infinite possibilities are available to the players, then there can be a lot of fun in throwing
allegedly insurmountable odds at the players and watching them surmount them anyway. But all too often, instead of being infinite possibilities, all it becomes is "can you read the DM's mind and find one of the three solutions they'll actually accept?" That falls far, far short of the ideal. If we're going to complain that clear, consistently-applied rules have cases where they don't produce the desired result, we should absolutely be making apples to apples comparisons. If we hold systematic rules to the worst possible applications thereof, we should compare it to the worst applications of the alternative as well—or we should recognize that the actual result from both things is neither ideal nor worst, but somewhere in-between, and ask what the
typical performance is
And It bink you'll find that rules designed to actually work, and actually be used as they are (albeit with sensitivity and discernment), rather than being designed to be ignored and subverted constantly, produce much better average performance despite sacrificing
theoretical maximum potential.
Why are do we make PCs only good at 2 of 6 saves or 1 of 3 saves then complain that PCs are weak to too many saving throws?...then nerf to compensate effects?... then say the game is too easy to win but PCs are too vulnerable to bring hosed?
I mean, in 4e you were always good at at least one, and generally two because most classes wanted ability scores from two different sets (e.g. Str/Cha, Con/Cha, or Str/Wis for Paladins). There were a few exceptions but they were uncommon, and could have been solved in the same way that (for example) Str-based Sorcerers, such as Dragon, got AC bonuses based off their Strength.
