"Precision does not trump accuracy" is something Michael Blum told me.
Also Charles Wheelan wrote in his recently published "Naked Statistics": “But no amount of precision can make up for inaccuracy.” I myself have been pleasantly surprised at how low multiplicity can be nowadays and still do S-SAD phasing. And If one uses iodide, the quality of diffraction data does not need to be as high as with sulfur-SAD phasing. Does one need to collect about different rotation axes? Not always. I wonder now if it would even hide a signal. Does one need multiplicity of more than 6 to 8? Not always. Be careful about radiation damage with increasing multiplicity and exposure. Does one need to minimize radation damage? I definitely think so. Does one need ot make sure the cryostream is not moving or vibrating your sample? Definitely yes. Does one need enough counting statistics to tease out the signal? Yes, but this depends on the expected signal. And so on. I can provide datasets of images if anyone likes. Jim
