Originally posted by iamatiger
oops, no, because I haven't proved that abcd is as large as it can be 🙁
Yes, that seems to be the most difficult part of the proof. Rigorous, formal proof is not my strong suit, but here is my best attempt.
We want the smallest possible fraction for 1/E; such that:
1)).........1/A+1/B+1/C+1/D+1/E=1 ;
with A<B<C<D<E ..all integers.
To make 1/E as small as possible, we need the sum of the first four to be as large as possible. Suppose: that 1/P is the smallest possible fraction such that :
2))........1/M+1/N+1/O+1/P=1
If 1/P is as small as it can be, then the sum of the first three is as large as it can be. Clearly, then, the largest sum we can get (less than 1) with only four terms is :
1/M+1/N+1/O+1/(P+1) So:
3))........1/M+1/N+1/O+1/(P+1)+1/E=1
Subtract 3)) from 2))
4))........1/P-1/(P+1)=1/E
Because we have assumed that 1/P is as small as possible, the sum of the first three fractions must be as large as possible.
Now suppose 1/Z is the smallest value such that :
5)).......1/X+1/Y+1/Z=1
The largest sum(less than 1)that we can reach with only three terms is :
1/X+1/Y+1/(Z+1) . So:
6))........1/X+1/Y+1/(Z+1)+1/P=1
Subtract 6)) from 5))
7))........1/Z-1/(Z+1)=1/P
Now, all of the above is intended to be a proof that if we can find the smallest value for 1/Z, we can deduce the smallest value of 1/P and then of 1/E by substitution.
It seems obvious that if 1/X+1/Y+1/Z=1; the only solution within the constraints set in the given puzzle is: 1/2+1/3+1/6=1; and therefore the smallest possible value for 1/Z is 1/6. I hope this is obvious enough to be accepted as self evident, so I don't have to bother trying to prove it. So if Z=6, then substituting back through equations
7))....and 4)), we find P=42 and E=1806