For PA to be effective, I assume that many people would need to see the same setup. For example, we see a pinbar on 4H that we anticipate will break down. So we set a short order just below the pin. However, other people would short when the right eye meets the left eye, effectively bringing the price lower until it breaks the pin. This, in turn, triggers a bunch of other shorts, resulting in a successful pinbar setup. Without those many people shorting with you, the pin could have failed.
So, if you were to use certain timeframes to interpret price movements that most people don't see, then that would lower the probability that the signal would be a profitable one.
I am not sure if this is entirely true, so someone please correct me if I am mistaken.