Better bisectors?

[kesterlester]

@tpgillam and @kesterlester to think about ways of using the ideas behind Lally ( https://arxiv.org/abs/1509.01831 ) to make for a supralinear/quadratically better bisector