However, if the scales are out of balance, the situation is now more complicated than in micro's procedure, since we only know that the different sac is one out of eight ones (instead of six, namely those on the scales). So we must try to exploit all we know about the sacs and we must exploit all the possibilities of a pair of scales! Let us summarize what all procedures so far have had in common: You divide the sacs into two groups, one group is put onto the scales and one is put aside. We look whether the scales are in equilibrium or not and infer in which of the two groups the different sac must be. But this is not efficient enough even if we can find out whether the different sac is heavier or lighter, as micro did. So reconsider my previous statement:
What will happen with the scales, if the different sac is in this group?peter_at wrote:we may leave some sacs on their scale
(and vice versa): What will happen with the pair of scales if the different sac is in this group?we may transfer some sacs from one scale to the other one
What will happen with the pair of scales if the different sac is in this group?we may put some other sacs aside
You might need them to assure that there is the same number of sacs on both scales.We may use some sacs that we have put aside in the first step
A small remark on micro's assertion (do not consider this as a hint!! it could easily mislead you!!)
This only works if the three Fs you put on the scale in step 2 have been on the same scale in step 1, otherwise you would not know. (of course, you must remember which sacs have been on which scale in (1) and which scale was down and which was up - this assertion may be used as a hint)micro wrote: (3) ... If the scale is out of balance, and the group was Y, and the scale was out of balance in (1), the fake is on the scale and you know on which side it is.
Do not think binary. There are more possibilities for the scales!

