All algebra students eventually come into contact with the quadratic formula. It provides a nice way to solve quadratic equations. The obvious question is whether or not similar formulas exist for higher degree polynomials. For degrees 5 and above, it is a non-trivial fact from abstract algebra that no such formula exists. However, for degrees 3 and 4 there are similar solutions.
Consider x3 + Ax2 + Bx + C. Let X = x + A/3. Then

where Q =
and R =
.
Suppose a3 + Qa + R = 0. Write a = b + c where b and c are to be determined. Then a3 = (b + c)3 = b3 + c3 + 3(cb2 + bc2) = b3 + c3 + 3abc. Then b3 + c3 + (3bc + Q)a + R = 0. (*) Also suppose bc = -Q/3. (Don't worry about why, since it works.) Then the middle term in (*) vanishes.
Therefore (*) becomes b3 + c3 = -R and b3c3 = -Q3/27. Thus, from substitution we have
,
or
![]()
By the quadratic formula we have

and, since b3 + c3 = -R,

Thus

where the cube roots are chosen so the product bc = -Q/3. This is a root of the altered cubic equation so x = a - A/3 Once that root is divided out you have a quadratic which can be solved easily.
For the quartic x4 + Ax3 + Bx2 + Cx + D, let x = X - A/4. This gives

where

X4 + QX2 + RX + S = (X2 + KX + L)(X2 - KX + M) where finding K, L and M reduces us to two quadratics.
Expanding the right hand side and equating coefficients of like powers of x gives L + M - K2 = Q, K(M - L) = R and LM = S.
Thus, from the first two of these we get 2M = K2 + Q + R/K and 2L = K2 + Q - R/K. Substituting these into the third equation gives (K3 + QK + R)(K3 + QK - R) = 4SK2 which becomes K6 + 2QK4 + (Q2 - 4S)K2 - R2 = 0 which is a cubic in K2 and can thus be solved using the technique given above. With K known, we can find L and M, giving us the two quadratics which we can solve.
These formulas come from the book An Introduction to the Theory of Groups by Joseph Rotman.