I took the liberty of making a small rudimentary simulation in Excel based on the routine example given here (at the end of the page) : https://www.yuque.com/zwopkb/asiair/focus
So, the first curve gives that the best star size is around iteration 7 which is at 9980 eaf steps.
From there, it goes back to 10010 and start the 2nd curve with the iterative step size being 10.
When the 2nd curve gets to iteration 4, the program knows that it's quite good. But it cannot know if the star size at 9970 is worst or better right ?
But writing this, I realize during the 2nd curve, since the backlash is not taken into account, we will have the same number of steps but not the same star size as, when going back to start the 2nd curve, there will be some amount of useless steps. So, yellow cells do not match a known star size... I'm even more confuse on how this routine works '^^.
Don't get me wrong though ! I know this routine gives good results. I just bought mine EAF but a friend of mine has his for months and it gives great results. But I just think that it might be even better ones !