Solve
for all real .
Put it in standard form and judge the scale. Reordering:
The leading coefficient is enormous, so the term will reach the size of the constant while is still tiny. That is the opposite situation to a quartic with a very small leading coefficient, where the roots are large.
Get a first estimate by dropping the linear term. Balancing only the two dominant terms:
Here the neglect is justified: at that the linear term is five orders of magnitude below . Because is even, this estimate immediately suggests two roots, one near and one near .
Confirm two sign changes. Sweeping across finds exactly two sign changes, near and . Reporting only the positive root — the easy omission — loses half the answer.
Bisect each bracket to full precision. Halving until :
The two are almost, but not exactly, negatives of each other: the term breaks the symmetry slightly, pushing the negative root marginally further from zero. A quoted value of is not accurate enough — there , not .
Substitute back. At : the quartic term is and the linear term is , together cancelling to within ✓. At the quartic term is and the linear term is , again summing to zero ✓. The remaining two roots of the quartic are a complex conjugate pair.
Need to solve a different problem like this? Open the solver →