I suspect this has more to do with reducing the level of the floor bounce (crowd head-bounce?) which combines with direct sound path from the PA at the microphones, somewhat reducing the combing effect of that multipath arrival, than it does with pointing the mics at a specific portion of the PA.
It is true that large hanging line-array PAs may have various sections of the array EQ'd differently and more SPL coming out from the upper sections, such that the sound projected from the topmost part of the array is heard as well balanced from the far back of the venue or the upper balcony, versus the sound from the bottom most part directed to the front of the audience. However, from typical audience distances even highly directional microphones will not offer sufficient angular pattern differentiation across such a narrow front area to be able to pick up more of the top section when located in the radiation zone of the mid or bottom section of the PA. The reason EQing a line array that way works is that the PA has sufficient vertical pattern differentiation to not have the radiation from the top PA section leak downward to a listener (or microphone) located in a section closer to the front. Pointing the mics up won't favor the top of the PA array versus the bottom, as both sections of the PA remain relatively close together in angular sense as seen from the recording position. But changing the height of the microphones can do so, if it raises the mics up into a different PA radiation zone.
As heathen mentions, with an ambisonic recording you can point the virtual microphones however you like afterwards, so an important step for me in dialing in the optimal configuration is choosing the most appropriate elevation angle. For AUD recordings, the main effect of adjusting the elevation angle is usually a change in clarity and high frequency response, and I listen for for best clarity when making that decision. That adjustment is more significant than I first thought it would be and those observations have lead to me intentionally pointing my non-ambisonic PA facing mics higher than I otherwise would, and I now generally point at the top or just over the top of the PA stacks as standard practice. I do that to sort of split the difference between having the direct on-axis point of the microphone still pointed at the PA while maximizing the angle above the audience and "floor bounce vector" from the ground in front.