Yup, I did question myself a lot about putting Gekko in vision, but since team use him a lot like that, I decided to consider him as a vision (Played in 1011 games as solovision, played in 589 games as a solo vision T1 + T2 combined in Split 2 only)
Ascent is also the map with the highest picked comp ever. Any comp that brings something different will have an effect simply because it's different. We've seen multiple times other comps tried (Loud 2024) and deadlock, to name a few, but they never find much value past the first few games, giving them an inflated win rate because once they get figured out they stop playing them.
I would be correct in assuming the non-flash pick rate is from deadlock, right? Teams have tried it out to some success, but what after a few games it begins to get counter-strated easily. The only team that I could say had longevity with this comp was Lev, which doesn't show success as a no flash archetype. These comps have a high winrate initially because of the surprise factor, but once they get figured out they stop using them, leaving them with a high winrate even if the comp didn't last.