500 Affinity is not negligible, especially after percentage boosters. It's simple, really. Attack stat has two functions. The first is that it's used in damage calculation as as flat number. This much is obvious. The second? It also functions as armor penetration. Your attack stat is compared against their defense stat, and then another percentage multiplier is calculated based off of this comparison. Note that this percentage has non-linear scaling, and I'm too lazy to find the formula. However, do note that this means that 1% attack is actually quite the booster. For example, losing 1500 attack dropped by asura spin damage on oushuu tier monsters by about 13% per spin. (Reasons why I say Baal Buster is a piece of junk)
Let's pretend that we all have 105 magatama and well ornamented final slotted magatama just for the sake of discussing things at their maximum potentials. We will also use Twin Swords as the weapon to argue because they have access to one of the best affinity titles in the game.
Let's assume the player has 5k affinity to begin with, and an infinite affinity cap. Add in the 14k attack from D3 Crosses, and the 10% bonus from +100, you get about 21k total attack stat.
Since I'm assuming hax ornas, let's say they have double ele + double affinity on their non 105 maga.
Adding in 100 total power grants the player 500 affinity
Miyamoto Musashi title + 2 ornas = 100% affinity increase
That leads to the conclusion that the player gets a 1k affinity increase, with an additional 10% from the +100 bonus.
So let's say the player now has about 22k attack.
Looking at that nominally, we only see a 5% or so increase in the player's total attack stat. But realistically, noting my 1.5k change (which is about 5.9% of my attack stat) is a 13% drop, I'd say it's fairly reasonable to conclude that this is about a 10% damage increase on Oushuu tier monsters. You'll sadly just have to take my word on this as I'm far too lazy find the exact numbers. Taking this into consideration, one can indeed conclude that it is not nearly as efficient as adding 30 of an element... assuming that the player has zero to begin with. But, let's, for example, use that lightning title, with our twin sword build here.
Before the player acquires the title, they would have 80 lightning. How am I arriving at this number? 40 from crosses, and 20 from each heavenly shield ornament. (Note that average is 19 on each roll, so this is fairly reasonable).
This means that the player already has an elemental modifier of 1.56 on their 70% lightning element skills. Adding in another 30 element pulls the player up to a 1.77 multiplier on said skill(s), which is a approximate 13.5% damage increase over the previous multiplier.
Make sure you don't just judge this as a 3.5% damage increase over the build that utilizes attack stat though - it's a little bit less if you divide 1.135 by 1.1.
Hence, if you are only spamming one element, it is indeed more efficient... by about 3%, which, to be fair, is quite a bit of damage. However, the point that I wanted to get to here is 500 attack is not negligible with end game tier builds.
Edit: I messed up. I forgot to include the extra affinity increase from the player's 5k base affinity that would be gained when taking the Miyamoto Musashi title. (This is significant as the lightning build would not have it.) So yeah, even more attack, meaning that this build would be superior by a tonne. Too lazy to edit the rest of my shit to reflect this.