never I said that non-combat tasks should be gritty for everyone, only that I like them to remain gritty for those PCs which never bother to invest in them
My puzzlement was over why this should be so, when combat is not handled the same way (contrast eg Rolemaster or Runequest or Classic Traveller, where combat ability
is handled the same as other skills).
A combat system where all PCs get roughly equally better by level at fighting, is probably the easiest way to ensure that everyone has something to do in combat. I would not mind a game where a PC who doesn't bother getting better at fighting, ends up lagging behind to the point of not being able to fight, but I would hardly expect another person at the same table to share the feeling. That's because in general when there's a fight, everybody fights, at least in the broad sense (i.e. "healing others" and "controlling the battlefield" qualify as "fighting" for me).
OTOH non-combat skills are most of the time (not always, but most) individual efforts.
And this is the answer to my puzzlement. Thank you.
I personally tend to prefer situations both in and out of combat that engage the whole party - social ("In the court of the duergar king!"), physical ("The temple is collapsing around you!"), etc.
whether the game allows characters of the same level with a large spread of their effectiveness at something.
4e certainly allows this. At 20th level, the lowest skill bonuses in my game (poor stat, no training, perhaps an armour penalty) are around +8 to +10. The highest skill bonuses (good stat, training, item etc) are +20 to +29.
I don't like the image of getting at a locked door without a Rogue in the party, and the players think "who cares, the cloistered/bookworm Cleric healer can do it anyway". I like to get there and think "damn we should have had a lockpicker here, what do we no now?".
The Thievery (= lockpick) skills for the 5 20th level PCs in my game are +7 for the paladin in plate and shield, +8 for the invoker in hide armour, +13 for the fighter, +15 for the ranger in hide, and +16 for the chaos sorcerer, or +18 with his tools (he is the only the PC to carry thieves' tools).
That's a non-negligible gap before we even get to the effects of training or items (none of the above PCs is trained in Thievery). Arguabyl it's too big a gap! I gather that 3E was even worse for scaling gaps, but those gaps can make it hard to set group challenges, though the DC charts and group check rules do their best to work around the big numbers.
A system like MHRP, for instance, scales only from d6 to d12 in ability - that's 4 steps (normal d6, enhanced d8, superhuman d10, godlike d12).
a note on the idea of interpreting the cloistered Wizard's high skill with locks (even tho she never picked one) as representing her other ways of dealing with the problem, e.g. through non-spell magic. It's a very smart way of handling the problem IMO, but once again it raises a gamestyle question, not dissimilar to that of abstracting hit points into "luck, and more", and that question is how much you like your game to explicitly connect mechanics with narrative vs how much abstraction you can take.
Related to what you say here is the following: when the group without a high lockpick rogue wants to get past a locked door, do you want the gameplay to be focused more on "What other mechanical resources, like crowbars or knock spells, do we have access to?" - a fairly classic D&D approach - or do we want the gameplay to be focused more on "What story can I tell about my PC that makes it plausible that s/he is picking this lock?" - which is closer to the MHRP approach.
4e is somewhere in the middle between these two approaches.