This is just not true. Grabbing and hitting are contingent upon each other only for situations where one leads to the other, but this is not always the case. Very basic example, connecting with Falco's fsmash does not really make you landing a grab more likely afterwards. I'm all about seeing the rule over the exception, but I think your statement has too many exceptions to accept as a general statement. Of course this is where I give my shameless plug that both hitting and grabbing are fully contingent upon the neutral game, but no one wants to hear that kind of content for the Nth iteration.
Side B does have utility on shield but it breaks down into a dichotomy between you and your opponent. For example, if you expect a jump out of shield or you expect the opponent to correctly DI the first hit, you can choose to do the second hit quickly even before you know if the first hit is successful. Of course, if the opponent does neither, the 3rd hit with no direction is good too. The problem with the 3rd hit is that marth/roy dedicate into the move more than the others by moving forward with the swing, which is typically solved by covering it with the 4th hit down for both characters.
I would argue that it is better to try to react to the earlier swings on side B if you intend to use it, and that you should seek superior alternatives before ever using it to begin with. the reasoning is that the prior outlined dichotomy with your opponent does give them margin that can be used to leverage an advantage on you. For a classic example, imagine marth using side B against a crouching falco mashing for the crouch shine. this is not a good place for the marth to be even if he's hitting with the move because the risk for a conversion from falco is so immediately evident.
roy's side b is much better for that character in a relative sense because he does not have any better alternatives. subsequently, roy's entire neutral game when played correctly is movement in and out of grab, dtilt, and side B. marth has better options by default simply because it's so much riskier to jump at marth than it is to jump at roy. next time you play against a roy on FD, replace your ledge game with the laggy full jump from the edge ("the tournament winner") and you'll realize how little he can leverage you against doing it.
The analysis has nothing to do with being dependent on anything, nor does it have to do with a lack of character understanding. you are simply using what you know to be the best option at the time, at least until your knowledge broadens. for marth, that's pretty much just grabbing under the assumption that the opponent doesn't sacrifice themselves to you by jumping around a lot. using grabs correctly and often is good because it leads to a massive increase in win ratio over the long term, and this is a trend that tends to hold pretty well across the better characters in the game. marth's grabs are just particularly devastating compared to even the other top characters because the advantage they give is so strong.
