Post
by Lurch » Tue Nov 07, 2017 7:21 pm
Perhaps I should've been clearer with the values I outlined, the numbers 20, 19+, 18+, etc don't mean your base will save value, but the roll required to beat a given check. Your will save could be 5 or 27 or 42. All that those numbers are meant to show is that rerolling gives the greatest benefit, in the absolute sense (as if having +5 will), when the roll by itself is 50/50. Say, when your will save is 8 and you need to pass a DC of 19, or it's 26 and the DC is 37. Now you might ask what difference does this "absolute sense" make; Mostly it was just there to compare and contrast the strength of slippery mind with iron will, a feat often thought to be lackluster.
With overwhelmingly good odds, the reroll is a rather nice insurance policy on top of everything else. However, failing a will save is never lethal by itself. A random 5% fail every now and then is usually inconsequential, especially if whatever was causing it is so under powered that it needs you to roll a natural 1 to have any effect to begin with. Now, if we were talking about fortitude saves and you had to weather a storm of save-or-die spells, then the 1-in-400 filter as opposed to 1-in-20 is very sensible. It should be noted that this whole situation is not likely for you if, as you say, your saves are poor to begin with.
In the hard case of overwhelmingly slim odds, the reroll almost never succeeds, just like the first roll and thus shouldn't be relied upon, is what I'm saying (this situation is more likely, with tough opponents and your poor saves).
SEGWAY INTO SYSTEMS DESIGN: Relative percentages do have an important place, best example being attack rolls vs. armor class, where D&D has for the longest time held a system that some game designers would call perverse. Instead of linear or diminishing returns (like most MMOs), raising armor class instead provides increasing returns against attacks, where every additional increase is more and more impactful than the last, at least until you reach the 95% misses mark at which point all increases are meaningless (concering that specific type of enemy). At this point only concealment helps any further.
This type of system favors ultra specialization where you have tanks that grab every point of AC (and concealment, assuming it's available) at the expense of most other things, or by contrast just making characters forget about it completely, as certain builds simply have fewer avenues to raise AC and struggling to get a middling score is usually a fool's errand, when there's stuff like summons, kiting, stealth, crowd-control or damage reduction to be had instead.
Thankfully there are ways to combat this phenomenon though, by having wildly variable attack bonuses amongst different players and monsters (in the same level range), where the percentage of encounters that one is impervious to (or vulnerable) shifts in a controlled fashion as a function of defense values.
The other obvious way is to have several additional "defense stats", like discipline, spell resistance, 3 different saves and even HP itself. Having different attacks target and/or ignore some defenses requires most characters to be at least mindful of these values, and very few builds can meaningfully excel at them all. And if they can, they should be gimped offense-wise as a consequence.