Just thinking about a rules idea, prompted by Denis McCarthy's examples in the comments in this post on AD&D Chargen.
He mentioned using Feather Fall to carry an entire party down a cliff or other drop. I thought about this - it's the kind of thing you would expect would take coordination. If you're going to jump off a cliff holding onto a wizard, you'd probably want to practice that - the grabbing, the coordination of jumps, the timing of the spell casting, etc. etc. You'd want to put some time and effort into perfecting this potentially dangerous task.
Same with, say, timing a charge through an area with a fireball. You move, the fireball is thrown ahead and explodes, you move right in and hack and slash and slay the wounded fireball victims. Time that a bit wrong and it's not unlike untrained troops trying to charge in just behind a moving wall of artillery fire - they're either going to go in too fast and take fire too or too slow and miss out on the effect.
I just think, if you're going to do this stuff regularly, you'll need practice. It's not a given that if it can be done once, it can be done always just because you thought of it. Or because you're assuming you must be practicing this in your copious spare time between sessions.
At the same time, you get times when the party is backed up against a cliff and the wizard says, "Hey, I can have everyone grab me, we can jump off the cliff together, and 1/3 of the way down I can cast Feather Fall and we'll all land with no damage!" You'd want that to work. It's a crazy cool plan. You'll want that "I throw the fireball just ahead of him" plan that allows the PCs to pull off a coup de main or turn a potential TPK into a victory to work.
But you might have those reservations I have above - if I allow this once, does that means it's now a routine action that always works?
Stealing an idea from Discworld, maybe it only works because it's 1 in a 1,000,000 exactly. Maybe it only works that first time because, against all odds, you coordinate it just right under stress. That stuff happens - I still tell the story of winning a grappling match with a move I didn't know and still can't do well.*
So perhaps a good ruling basis would be, yes, your crazy plan works. Perhaps a roll for avoiding calamity ("Don't roll an 18" or "Don't roll a 1.") Otherwise, it's either "Yes" or "Yes, but" or "Yes, also." After that, though, it's subject to all of the if, ands, or buts of thinking it all through. Yes, you perfectly pulled off that cliff jumping to safety. Next time you try to coordinate five PCs doing the same thing there is a roll involved to see if you do it properly, unless you put in some in-game effort to justify making it routine. Next time you perfectly time that fireball, we'll apply this penalty I realized should have applies and just roll on the scatter chart for a miss and see what happens. And so on - you get the benefit of the doubt on your craziness or "it just might work!" plan, but you can't use that as settled precedent for it working in all cases even if easier than the first case.
I think some people would have trouble adjusting to this - especially if you're used to negotiating out "works now and forever" and not "now, we'll see about later." Yet I think it's got promise. It nicely severs the relationship between the worry about allowing something now and then seeing the consequences later, rewards trying new things (the less precedent against it, the more likely it's going to work this one time), and keeps things moving. In my case I'd be less likely to worry that I've just allowed a game-breaking change. Or that I'd turned a ruling meant to keep things moving and reward your cleverness into the basis of all action going forward. "We do the Feather Fall thing over and over until the entire army is down the cliff."
I like the idea of making this a standing, known goal when making rulings - I'm probably going to say yes, but it's likely that it's not precedent. I'm saying you try the cool thing that should probably take a lot of time, effort, learning, practice, etc. and everything just worked out this time under stress. If you want to repeat it, and it seems like it should be harder than it was, that's something we've agreed to. Instead of just saying "No!" because I'm concerned about where it would lead I could say "Yes!" because I know it doesn't have to lead anywhere. It might be precedent, it might just be a one-off set of accidentally perfect circumstances and execution.
* My coach talked me through it in real time, against an opponent who could hear his words as well as I could. Imagine his chagrin when I couldn't pull it off even once properly in practice after that.