You _can_ use floating point if you are very careful and know what you are doing...

You _can_ use floating point if you are very careful and know what you are doing and know about decimal normalization (see e.g. OpenHFT implementation for high-frequency trading: https://github.com/OpenHFT/Chronicle-Core/blob/ea/src/main/j...)

But if you are not an expert, you better stick to BigDecimal and absorb the performance costs.

Nice share! Interesting sorcery in their code:

> final double residual = df - ldf + Math.ulp(d) * (factor * 0.983);