It turns out that the previous run hits the flagpole in 4-1 two frames too late. this run fixes that oversight. I was not able to find any other improvements.
I also made a console verification of this run.
I made this is dev. build but it also works in 2.5.2, so it's not showing off any super precise accuracy, but I think it's still a cool run to see console verification of.
Samsara: Judging #233.