Thomas Shey
Legend
As a game designer, my feeling is that if I'm giving you one chance in 216 of success, its almost like a fraud, because people are optimistic and they will expect that they will pay off on that sometimes, but they won't, not in any reasonable finite amount of playing. So no, I don't even want those to exist, and I ESPECIALLY don't want them to be non-obvious (and believe me, dice pool odds are mostly very obscure to people!).
But this is acting like only the extremes are relevant. As an example, if you have two Hero characters who have the same CV, they're hitting vaguely around 50% of the time. But add 3 to either CV, and that drops to 25%/75%. Add one more and it goes to 90%/10%. I think that very much does serve a purpose, and is not beyond people's comprehension.
(Your point about how obscure full dice pools can get is valid, but it still tends to be incremental without being trivial).
I mean, I'm a math guy and I am not going to tell you off the top of my head what the odds are of winning an opposed check in TB2 where I have 8 dice and the other guy has 12 dice (all 4-6 contribute one success to each side). Is it one chance in 3? I bet, without resorting to some online dice odds calculator, that nobody has that answer on a first reading (sure, we can all probably figure it out if we really want to, but at the table?). OTOH everyone knows the odds yielded by a d20, and at least 5% chance of success, while not great, WILL come up now and then.
The question is, do you want it to come up that often? Given the number of rolls made in combat in particular?
For all these reasons I stuck to a d20 based design for my own game. I'm not pooping on dice pools or anything, I just think their virtues are overrated. PbtA's 2d6 always rolled straight up by the player is not bad either. I think people are pretty likely to understand that a 7 is 6 times more likely than a 12, for example.
Meh, again, 95% of the time you want odds in the 25-75% range anyway. Now and then you want something as low as 5% perhaps. I don't really see the growth curve of D&D and such as a bad thing, personally.
Then we just disagree. I don't see too much point any more in having advancement that's mostly illusory unless its gone on a long time. I'd rather have it a bit less frequent but noticable.
I don't know about that. I mean, TSR's FASRIP basically did EXACTLY that (admittedly its a d100 system, but in a practical sense it is the same issues). People love that thing, it works great, they still play it and hack on it, and its been out of print for 30 years! I mean, sure, it isn't a super common technique (though other games certainly have used it) but my point was just that it is very doable, you don't need dice pools. They are OK, but the math is kind of a PITA, actually.
Notice that FASERIP is table-driven; and the table does a bunch of other things at once. Care to point at a modern game that isn't a retroclone of that that goes that route? The current version of RM, maybe? The hobby has stopped being a fan of table-driven resolution a long time ago.