Consider two primes, P1 and P2, we construct a quadrant with an arc ABC and another arc EFG. These arcs are of length.
arc ABC=P1+P2 and
arc EFG=P1+P2+2
Both arcs are divided by its length to give integer points and provide markings, each of unit length 1,
Since there is symmetry about the bisector of the arcs, only one portion is discuss here. Since, P1+P2 is even, this portion is always odd.
Let m+n=P1+P2
and
j+k=P3+P4=P1+P2+2
Consider,
P2P1+P2=P3P1+P2+2
P3=P1+P2+2P1+P2P2=nn+m(P1+P2+2)
kj+k=P3(P1+P2+2)=nn+m
So,
1j+k=(nk)1n+m
since n<k
the unit division along (j+k) is a fraction of the unit division along (n+m). P1 is not a factor of P4 and P2 is not a factor of P3.
In fact, given that the length of arc EFG is just one plus the length of arc ABC. This gain in a unit division is spread over the length of EFG There is no line through the origin that will meet both markings on ABC and EFG.
If a line OF is drawn from an integer mark on EFG, this line will not intersect with any integer mark on ABC. This is true for any marking on EFG, and so for all values of P4. A missed integer marking means the obtained value is not an integer; it is not a valid value.
Also since, arc EFG has length,
arc EFG=P1+P2+2
and the lower arcs have length P1+P2, P1+P2−2... ...6,4,2
There is a possible a>1, such that
P1+P2+2=a⌢L
where ⌢L is the length of a lower arcs.
1⌢L=a.1P1+P2+2
The unit markings on arc L is of multiple unit width on arc EFG.
A line OF through an integer point on this lower arc ⌢L will cross another integer point along EFG. This would mean P4 has a factor (like P1) on ⌢L, given by this intersection.
Consecutive marking on this lower arc L will intersect EFG at a spacing apart. That means there are a−2 choices of P4 for an a size aperture on EFG that do not have this factor.
a decreases with with longer arc length L. The choice of P4 to avoid an integer marking on L also decreases. With a lower arc length of 2 that accommodates for the case P4=P3, there are at most (P1+P2−2)−2+1,P1+P2−2 (odd numbers discarded) possible values of a (every lower arc provides a factor and is avoided), excluded are L=(P1+P2) that provides no factors and L=1. So, there is at least one choice on arc EFG that allows line OF to avoid all integer intersections on the lower arcs. (The one choice on arc of length P1+P2−2 when all lower arcs have factors).
That means P4 can have no factors. P4 can be prime. The primes P1 and P2 guarantee that there is at least one choice of P4 to be prime. The arc ABC allows line OF to sweep for all integer on EFG, to obtain a value for P4 that is prime.
When P4 is prime, and by symmetry P3 is also prime (swap P3 with P4). For P3, start at the horizontal instead, line OF missing all markings on lower arcs. It is the same line when P4 was discussed.
So, given two primes, P1 and P2 the sum of which is P1+P2, there are then P3 and P4, such that the next integer,
P1+P2+2=P3+P4
and both P3 and P4 are prime.
So by induction, given that (1+1=2), (2+1=3), (3,1=4) and (3+2=5) all even numbers can be represented as the sum of two primes.
Goldbach's Conjecture is proven.
Note:
Factors found using this graphical method divide P3, P4 and their sum P3+P4=P1+P2+2.
If OF crosses with integer markings on EFG and any of the lower arcs L, then a factor is found on L. That cannot happen. P1+P2+2 cannot be divided. This derivation does not choose P4 and P3.
Starting from an integer marking on arc L, but the line OF misses integer markings on arc EFG, means P4 and P3 are not integers. Starting from an integer marking on EFG, but the line OF misses integer markings on L, means the lower arc does not provide a factor to both P4 and P3 and P1+P2+2. Not just P4. This graphical method find factors for P4, P3 and their sum P3+P4=P1+P2+2.