Well, that's the thing about those one variable formulas. If the variables are X, Y and Z, it is evaluating just X, and assumes Y and Z correlate with X based on averages. Unfortunately not that many people are average. There seems to be too much variation in human proportions to make that assumption for this purpose. For example, someone with a 32.5" inseam could very reasonably be 5'10" tall. Meanwhile, I have a 34.5" inseam and am 6'0 tall. ALL of the height difference is in the legs and torso length could be identical. Then, with longer than average legs you'd assume longer than average arms, but no, I have average arms. So in this case I could need the same reach as the 5'10 guy but a roughly 2" higher saddle height and 2" higher stack to achieve the same fit.

That's why the competitive cyclist calculator is popular, it seems to take a lot of variables into account. I'd be very interested to see the calculation going on behind the scenes. There's a quirk with the top tube results that suggests it could use some tweaks. In fact that is the reason I started this thread.

The site says, "The Eddy Fit typically features a saddle/bar drop of only a few centimeters." However, when my Eddy Fit results are actually plotted out (using AutoCAD for ease and accuracy), it resulted in a 16+cm saddle to bar drop. To fix that error and achieve the prescribed Eddy Fit with "a few cm" drop you would bring the head tube straight up, making the top tube longer than is prescribed due to seat tube angle. Long story short i think the saddle-bar reach and saddle position results are sort-of accurate, but the top tube result is useless. I thought about making a separate thread with the details. Since that is the calculator I figured there might be interest in a pretty glaring flaw. But that would take some time.