Multiplying by the Velocity Factor for the ¼-wave phasing line has been a topic of endless debate in the radio world! There is also some debate about the actual required length of the 75Ω feeders, and if each dipole needs a Pawsey stub or other balun to match properly.....
I did endless experiments with this configuration (at 1 GHz, to miniaturise the aerials), and with a three-stack, four-stack and even a six-stack! The twin gives you the best return on investment - almost exactly twice the ERP of a single dipole, and a flattened radiation characteristic. The Four-Stack will give a couple of more dB of gain, so the ERP is improved a bit, and the radiation characteristic flattens and even droops a bit, making it a good choice from a high site.
I experimented a lot with the element spacing, with the boom arm dimensions, with the matching lines and with the use of baluns. The best results were with the upper dipole boom coming off the very end of the support pole - the upper dipole is then above the mast, in free space. This changes the match slightly. Pawsey stubs (made from ¼-wave pieces of the 75Ω coax) improve the SWR significantly, and can get you another dB or so actually radiated.
The 75Ω feeders aren't as critical as you might imagine, it's the spacing between the dipoles that really matters. If you build the aerial, then clamp the lower dipole so that the tip of its upper element is ¼-wave from the top dipole, test the SWR (in Free Space - it's no good if you do it at ground level with walls, trees, hedges and so on around you!). Try moving the lower dipole up by about 3cm and test again. If the SWR worsens, move the lower dipole downwards a bit and test again. Eventually, you'll get the thing spot-on. Measure everything!!! Log the lengths of the coax, the length of the stubs, the dipole spacing and size and the boom lengths. You can now reproduce that aerial with minimum hassle!
If you have the time, and want to make the effort, it's useful to build the antenna for four frequencies in the band. I did 88, 93, 98 and 107MHz for my full-size experiments. You can then plot graphs of each dimension, so that you can make the antenna for any frequency in that range with confidence that it'll match and radiate effectively. It's hard work, but worth it in the end!
Is it meant to smoke like that?