Ok, so I notice that several threads are now being derailed by this concept. It's not my theory, so I don't know if what I am hearing is a bastardization of it or not. But, what I am hearing is either deeply mistaken or based on some deeply mistaken hidden premises. I want to take a few minutes and lay out its flaws.
What's IP-proofing?
Note that this isn't my theory, so I will be doing my best to characterize it. If I am mistaken, please correct me. The term is "Iterative Probability proofing." And, the idea is that in D&D (really, any RPG, I'd suppose) your character is going to be subject to a number of threats, which can effectively kill you. So, with some probability a "catastrophic event" will occur -- you will fail the crucial save, get crit on by the crucial attack, and so on. Over the course of your adventuring career, this probability will occur: if there's only a 5% chance of suffering a catastrophic event in each encounter, over enough encounters that's going to converge to 100%.
Ok, so as a basic idea goes, that's totally fine. It simply means that, if you're interested in making it over the long haul, you need some sort of defense against such things. That defense could be rerolls, immunities, or the ability to recover from the failure.
Example 1: suppose you will fail a save or die 10% of the time. That means eventually you will fail one. If you have a reroll, though, from one of any various magic items, so long as you're not subject to too many saving throws over the course of the day or what have you then you're fine. It essentially resets your "IP counter" till you fail again. Even the humble Fighter can take advantage of this. At higher levels you might need more investment, as the frequency of such effects will increase, so you'll have to get more in the way of defenses.
Example 2: in the course of the fight, bad rolls, etc. have claimed the lives on one of the party. One characters casts revivify on him, or delay death, or some similar effect so he can continue adventuring.
As these examples indicate, I think IP-proofing is pretty straightforward, and not hard to attain. It is a pretty subtle charopp concept, though, and one that I spend a little bit of time teaching new players. It may be the case that some classes have more of it built in than others, but it's readily available in troupe play. Arguably, a correctly optimized healbot cleric's job is something along those lines. I mention this b/c it seems one vehement complaint about weaker classes is that they cannot do so.
Basket Burner's Death Spiral
I don't like to call out people by name, but he's (assumed pronoun, sorry if it's mistaken) obsessed with this idea and it colors all his posts. The idea, I take it, is that every time one of the members of the party fails against a catastrophic event it does one or both of the following. First, it weakens the party for the rest of that encounter, increasing their probability of failure. Second, it weakens that character in the long run due to level loss for raise dead or other wastes of resources.
However, I also think this is not true, and missing some crucial premises. Two obvious responses, and then I'll get to the more subtle one. First, it assumes that every encounter is on the knife-edge of life and death. That's probably not the case, even in other people's campaigns. Second, this isn't Gygaxian AD&D -- revivify, delay death, and so on allow for people to come back to life without level loss. Also, it's unclear to me how big a deal level loss is. If catastrophic events are relatively rare (e.g., 10%), then the increased XP gains will lead the lower level guy to catch right back up before another catastrophic event occurs. Even if he's facing an increased probability of doing so (e.g., 12%).
Now, I anticipate that BB will say something to the effect of: "all the resources you just spent raising that jackass just cost the party something, making it weaker." So, that revivify the cleric cast could, in the absence of a jackass (a basket weaver?) could have used that slot for something more meaningful. The problem I have with that logic is that it artificially divides out resources spent to ex ante IP proof and those spent ex post to IP proof.
Here's what I mean, I think no one is going to say that a wizard casting Greater Mirror Image is wasting his resources. It's a great spell that keeps you alive. What, then, is the difference between that and Delay Death? The only relevant difference, depending on the ability we're talking about, is whether that character is still in the fight or not. That, in turn, depends on (1) how much fight there is left, and (2) whether you're on that knife-edge described above.
Now, it may be the case that ex ante IP proofing is generally speaking more effective than ex post. I'm not sure. It might be true when it comes to character death, less so with other conditions.
My real point, though, is that I don't see how there is any clear mapping from a character not being able to IP-proof themselves and relying on someone else in the party to do it, or on magic items to do it, and the supposed IP-proofed idea.
P.S.: it occurs to me that there is also some ellision in the knife-edge scenario. I've never played the most powerful character I can in a game. That's not what I'm out for, I usually want to play a character of some sort. The knife-edge cases would be charitably read as an encounter where for that given party it's nearly a matter of life and death. But, I think it often colors character creation discussions as well. If that's the case, though, then the implication is that everyone should be playing Pun-Pun or equivalents.
