So I have a pretty basic question
what’s the benefit of a dice pool rpg system over just rolling a d100 or something?
Okay so, a single die (but also d% even though this is rolled as two dice) has a flat distribution. There is an equal chance of all results on the die. When you roll a d20 your odds of a 1 are the same as the odds of a 20. The d20 was intentionally used for attacks in D&D because of the swinginess of that probability.
When you use a dice pool, it creates a bell curve instead, broadening and stabilizing your results. This is why PbtA is built around the 7-9 mixed success result - this will happen with much more stable frequency than rolling 10+ on 1d20. Similarly, Exalted uses pool math that allows the player to estimate every 2 dice generating one success, around which all the base difficulties are structured. When you start fucking with things like ones taking away in classic World of Darkness, you destabilize the nice curve and get into bad swinginess again.
The smaller your single die, the tighter (andore difficult) the framework for where you can adjudicate failure. My butthole clenches when I see people using single d6 resolution and have failure on 3 or less.
I don't really think people understand the importance of mathematical stability underpinning your basic resolution, if you're going to do something with randomization at all.
When someone says "We'll just roll a d20" I ask, is it because you want that level of swinginess? Usually no. Usually the answer is because that's what that person is familiar with - which fuels some of my distaste for WotC's stranglehold on the ttrpg zeitgeist.
Anyway, the difference is all in the control you want to have over success, failure, and your system's ability to predict and adjudicate that.