It turns out that if you accept papers based on a fix percentage of submission number, increasing rates of acceptance reduces the pool of unaccepted paper and this larger percentage of the smaller queue ends up giving about the same number of accepted papers overall.

I also have this funnel simulation https://i.postimg.cc/gz88S2hY/funnel2.gif

+ Same number of new produced papers per time unit.

+ Different acceptance rates.

+ But... *same number of accepted papers* on equilibrium! With lower rates you just review more.