The chart is a big average. I fit a set of equations to a sample of data comprised of many different formulas and workflows. So, like hans said, it's a starting point, but also like Jonas indicated, it can be all things to all people. Once you find the little tweaks that make it work for you it's pretty accurate. For me, I find that if it go just a bit more than it tells me - not even a full column to the right - it almost always works perfectly. For NP, I start with cold water, but for anything else, I just use tap water at whatever temperature comes out.