I collected some data on how delaying a battle affects initial battle timing.
13452 13934 482 35
13456 13969 513 66
13460 13942 482 35
13464 13969 505 58
13468 14022 554 107
13472 14033 561 114
13476 13954 478 31
13480 14048 568 121
13484 14057 573 126
13488 13960 472 25
13492 14061 569 122
13496 14018 522 75
13500 14047 547 100
13504 13951 447 0
13508 14062 554 107
13512 13994 482 35
13516 14028 512 65
13520 14002 482 35
13524 14028 504 57
13528 14083 555 108
13532 14092 560 113
13536 14015 479 32
13540 14108 568 121
13544 14116 572 125
13548 14026 478 31
13552 14120 568 121
13556 14078 522 75
13560 14106 546 99
13564 14011 447 0
13568 14128 560 113
13572 14054 482 35
13576 14088 512 65
13580 14062 482 35
13584 14088 504 57
First Column : Frame where I pressed A to cancel the "????? got a level" box from the 2nd fight of Narshe
Second column : Frame where Terra command box showed
Third column : Difference between second and first column
Fouth column : Third column minus minimum value
Results change only each 4 frames.
Let's see a graph of these result :
It gets interesting. There is a periodicity of 16. I will make further test to see if it generalize to all battles.
Could it means I would have to test only 16 possibilities to have the best battle startup ? No, because of preemptive battles. I still havn't figured out how it works. There is 1/8 change I get a preemptive battle and on 34 try I got none.