Solve for :
Try the rational root theorem first, and let it fail. Any rational root of a monic integer cubic must be an integer divisor of the constant term . Testing the plausible candidates: , , , , , . None of them vanish, and the same holds for the remaining divisors, so there is no rational root and no factorisation over the integers - the cubic has to be solved numerically.
Locate the roots by sign changes. Those same evaluations of already bracket every root. Since and , a root lies in . Since and , a second lies in . And since and , the third lies in . Three sign changes for a cubic means all three roots are real and simple.
Refine each bracket with Newton's method. With , iterating from the midpoint of each bracket converges in a handful of steps to The derivative is comfortably away from zero at each root (, and ), so Newton converges cleanly.
Verify with all three Vieta relations, not just the sum. For the roots satisfy , and . Our values give matching , and exactly. Checking only the sum is not enough - a wrong pair of roots can still add up correctly while the product is badly off.
State the answer. The huge spread between them is what the coefficients predicted: a leading behaviour of pushes one root out past , while the large constant keeps the other two comparatively small.
Need to solve a different problem like this? Open the solver →