This thread is arguing that for each tier in difficulty the game should draw from different lists of items, with increasingly useful loot to compensate the player for skill point investment, a system which actually exists within the game's architecture but was grievously ignored. In providing an example to contrast this point, you cited only two of the many locked containers in Fallout 3, both of which are entirely outside of the framework being discussed.
I have to agree as well with this. The premise of the argument is that the risk vs. reward formula applies equally well here as it does in real life. In fact many would argue that getting the risk vs reward formula right is of paramount importance to making the player feel like they are accomplishing something, that they are moving forward with their character development. You can look at chest loot from many angles; one being that if I the player put in X amount of time solving a mini-game, the reward should be equivalent to that effort In General. The other argument is that all rewards should be surprises and what you get for an effort doesn't apply to the loot your rewarded with. I am strongly in the former camp, as are I think most video game developers, in that the risk vs. reward has to be there and it has to be right or the player feels over- or under- rewarded for their efforts. When done badly, it can quickly lead to frustration.
The problem I have with random loot drops Overall is that they can result in great frustration for the player, especially if the loot comes out low for something the player invested a bunch of time in. In these cases the game ceases being Fun, the burden factor goes up, frustration goes up and this leads to bad things. However when the loot is at least Somewhat predictable versus the effort you have to apply, then the player gets more choice in how they spend their time and on what they spend their time on. If I'm up for a big adventure and want big loot, I like knowing that (for example) Evergreen Mills is loaded with loot and waiting for me. If however I just want some light action and don't care about big loot, I'll scavange the wasteland. In that predictability, I can better plan my gaming time and know that I will be amply rewarded for it.
One can argue this both ways, but the gaming companies tend to stick to the more predictable path, as that leads to happier players and more popularity for the game overall, which ultimately means more money. We puny human beings like predctability in our lives, and that trends right into our video games that we spend a million hours playing. So it's little wonder why most of us want to know that if we spend tons of time on a very hard lock, that we're reasonably sure that something good is inside. It's expected that randomness is involved, but in general the risk has to balance with the reward. Todd Howard has spoken of this difficulty in more than one podcast now, so while we GECKulate over it here you can rest assured that Bethesda and Obsidian contemplated the same question themselves. How successful Obsidian will be? Oct 19th.