Solution
Solution 1
p(0,0)p(1,0)p(−1,0)=a0=0=a0+a1+a3+a6=a1+a3+a6=0=−a1+a3−a6=0
Adding the above two equations gives a3=0, and so we can deduce that a6=−a1.
Similarly, plugging in (0,1) and (0,−1) gives a5=0 and a9=−a2. Now,
p(1,1)p(1,−1)=a0+a1+a2+a3+a4+a5+a6+a7+a8+a9=0+a1+a2+0+a4+0−a1+a7+a8−a2=a4+a7+a8=0=a0+a1−a2+0−a4+0−a1−a7+a8+a2=−a4−a7+a8=0
Therefore a8=0 and a7=−a4. Finally,
p(2,2)=0+2a1+2a2+0+4a4+0−8a1−8a4+0−8a2=−6a1−6a2−4a4=0
So, 3a1+3a2+2a4=0, or equivalently a4=−23(a1+a2).
Substituting these equations into the original polynomial p, we find that at (ca,cb),
a1x+a2y+a4xy−a1x3−a4x2y−a2y3=0⟺
a1x+a2y−23(a1+a2)xy−a1x3+23(a1+a2)x2y−a2y3=0⟺
a1x(x−1)(x+1−23y)+a2y(y2−1−23x(x−1))=0
. The remaining coefficients a1 and a2 are now completely arbitrary because the original equations impose no more restrictions on them. Hence, for the final equation to hold for all possible p, we must have x(x−1)(x+1−23y)=y(y2−1−23x(x−1))=0.
As the answer format implies that the x-coordinate of the root is non-integral, x(x−1)(x+1−23y)=0⟺x+1−23y=0⟺y=32(x+1) (1). The format also implies that y is positive, so y(y2−1−23x(x−1))=0⟺y2−1−23x(x−1)=0 (2). Substituting (1) into (2) and reducing to a quadratic yields (19x−5)(x−2)=0, in which the only non-integral root is x=195, so y=1916.
The answer is 5+16+19=040.

Solution 2
Consider the cross section of z=p(x,y) on the plane z=0. We realize that we could construct the lines/curves in the cross section such that their equations multiply to match the form of p(x,y) (same degree of x and y in terms) and they include the eight given points. One simple way to do this would be to use the equations x=0, x=1, and y=32x+32, giving us
p1(x,y)=x(x−1)(32x−y+32)=32x+xy+32x3−x2y.
Another way to do this would to use the line y=x and the ellipse, x2+xy+y2=1. This would give
p2(x,y)=(x−y)(x2+xy+y2−1)=−x+y+x3−y3. (But
(Another way would be to use the hyperbola from Solution 1. Interesting that different curves both work.)
At this point, we consider that p1 and p2 both must have (ca,cb) as a zero. A quick graph of the 4 lines and the ellipse used to create p1 and p2 gives nine intersection points. Eight of them are the given ones, and the ninth is (195,1916). The last intersection point can be found by finding the intersection points of y=32x+32 and x2+xy+y2=1. Finally, just add the values of a, b, and c to get 5+16+19=040

Solution 3 (Ansatz)
We can plug in the values to obtain
p(0,0)=0⟹a0=0
p(1,0)=0⟹a1+a3+a6=0
p(0,1)=0⟹a2+a5+a9=0
p(−1,0)=0⟹a1−a3+a6=0
p(0,−1)=0⟹a2−a−5+a9=0
p(1,1)=0⟹a4+a7+a8=0
p(1,−1)=0⟹a4+a7−a8=0
p(2,2)=0⟹2a4=3a6+3a9⇔2a4+3a1+3a2=0.
Now, this means that
p(x,y)=a1x+a2y+a4xy−a1x3+a7x2y+a8xy2−a2y3.
After some simplifying, we obtain
p(x,y)=a1(x−x3)+a2(y−y3)+a4xy(1−x).
Since p(x,y)=0, 3a1+3a2+2a4=0, and we suspect that:
x−x3=y−y3
and
xy(1−x)x−x3=23⇔y1+x=23⇔32+32x=y
.
Plugging this into the first equation, and factoring, and cancelling (x+1), and simplifying, we get 19x2−43x+10=0, so we find that (x,y)=(195,1916)⟹a+b+c=5+16+19=40.
~~pinkpig
Video Solution
2008 AIME I #13
MathProblemSolvingSkills.com