They don't. It's that old trust thing again.
More the point was that "showing your work", as it were,
exclusively after "your work" no longer communicates anything, won't really affect the trust-o-meter. It won't hurt, or at least it shouldn't if the players are reasonable. But nothing is really gained or sustained in the trust department by only revealing mechanics after they've become irrelevant.
Conversely, there would be a significant potential for
gain if you show the work while that work is still relevant, because then the players can
see you playing by the rules you've set. It isn't perfect, paranoia can always find an excuse of course, but at least for me I would very much appreciate that gesture, whereas I really wouldn't feel much of anything (except mild curiosity) if shown a mechanic when it's been reduced to pure novelty information.
So if you were DMing that scenario with the floating gem generating flashing lights each colour of which had different effects ranging from useful to deadly, assuming the PCs had no prior experience with or knowledge of such a thing, what if anything would you have revealed about its mechanics, and when, and why?
Well, I don't genrally make that kind of assumption unless there's a reason to. But even if we grant that? Probably would tell them at least the general outline after the end of their first (or second, if it seems like there will be more than two) encounter. Doesn't have to be specific numbers, just descriptive stuff. Generally, I'd try to choose effects that would leave reasonable evidence behind, e.g. something that burns leaving scorch marks where it fell, something that induces sleep having the sound of a lullaby, etc. Something beneficial might have the sound of a battle theme or feel like the gentle warmth of a campfire or the like, depending on what the effect is.
The way I see it, while learning about an obstacle is interesting up to a point, making smart decisions about information you do in fact have is almost always interesting. By informing the players while it still matters (even if it's not 100% guaranteed to be of significant importance), I'm giving them that chance to make informed decisions and even try to manipulate the situation in their favor (e.g. perhaps they come back with mirrors in the hope that they can
reflect a color-flash at an enemy!) Further, that way, choices good and bad are squarely and obviously theirs.
When you-as-PC encounter something unusual, one would expect there to be some mystery involved as to what makes it tick.
Sure, but it doesn't then
mean very much if you wait only until after the unusual thing is unequivocally never going to show up again. Further, plenty of unusual things can still be understood pretty well even without needing to be an expert on it first. A watering fountain would probably be mystifying to a
neanderthal at first, but a little experimentation would have probably made him comfortable with it quickly. I find many PCs are in a similar boat, even if they don't know the precise theory they pick up the core essentials easily enough.
As it happened, they had a PC in the party who had encountered a similar gem before, though that one only flashed one colour and only had one possible effect, which meant that through that PC they had some clues going in as to what might be involved. Had they not, though, they'd have been flying blind until-unless they did some investigation and-or experimenting.
Ah. See, I thought this was "the players DID investigate and experiment, didn't learn anything, then the whole area was erased/destroyed and
only then did I tell them the mechanics." Also it's now sounding like this was a singular event? Like they only interacted with that gem the one single time and then nothing else? I was under the impression that they had seen it and even limitedly/distantly interacted with it (e.g. not in the room itself), but despite multiple encounters nothing was really learned until after no more encounters could ever happen.