A ef 70-200 f2.8 L IS mkII has a min focus distance of 1200mm and offers a max magnification of 0.21 at 200mm. If one focusses at 70mm and then zooms to 200mm the image is very slightly out of focus. Not a parafocus lens, but close.
A ef 70-300L IS has the same min focus distance and offers the same max magnification at 300mm. So it looses 100mm of of focal length when compared to the 70-200. Focus at 70mm and zoom to 300mm and the image is a lot more out of focus than the 70-200 was.
What do these "1200mm" mean? Distance from sensor? Distance from front nodal point? Distance from frontmost point of the lens? You see very quickly that the numbers you state may not say what you think they say, especially if you consider that a 300mm lens must lose more focal length to achieve near focus than a 200mm lens, and that it likely is longer and has a different internal construction.
The Sigma 120-300 f2.8 OS's figures are really curious, MFD of 1500mm (not that far off the 70-200's) but it's max magnification is a lowly 0.12x. So there's a massive focal reduction going on there. Focus at 120mm and zoom and the image is hopelessly out of focus even by 200mm.
If these 1500mm are measured from the sensor plane, and given the total length of this lens, the difference between 1200mm and 1500mm is much more substantial than it would appear to the unwashed masses.
By my conservative estimates, this lens drops to a focal length of around 240mm at 3m or less.
So what? If you look at http://www.photozone.de/canon_eos_ff/510-canon_70200_2is28,
you will see that the 70-200 L IS drops from 200mm to 161mm, that's the same relative drop (161/200 ~ 240/300). And again, nobody is trying to defraud us here, it's simply laws of optics at work here. Unless you give up internal focusing, you will have to live with loss of focal length as you focus close.
A 300mm f4 L IS has a MFD of 1500mm and that's got a Max Mag of 0.24x so I'm sure you can see there is a lot of focal length loss with this Sigma at MFD.
AFAIK the 300 F/4 doesn't have internal focusing, so it likely won't lose much focal length at MFD. If you main goal is having long focal length at MFD, stay away from lenses with internal focusing regardless of vendor.
The whole point of a 300mm f2.8 is to diffuse the background by having the main subject fairly close and the background far away. With this lens, even a 70-200 f2.8 II L will do this better. Pop a 1.4x TC on the 70-200 and it's even easier.
If blurring the far away background is your aim, the lens with the largest entrance pupil is what you want regardless of focal length. As it just so happens, this diameter is more or less the front lens diameter for most tele lenses, so you should pick accordingly. You can also estimate this number by dividing focal length by F number. No teleconverter is going to help you there in any way, shape or form!
In regards to the f stop value of the 85mm f1.2L, no it's an f1.2. F stops are a mathmatical formula and are quite abstract. T stops are a calibrated transmission of light through the optic.
I did not refer to T stops, but to real F stops. If you look at Canon patents, you will see, that stated F stop numbers are not accurate and conveniently rounded down for marketing purposes:http://www.canonrumors.com/2011/12/canon-files-a-patent-for-a-bunch-of-lenses/
It is quite clear that a lens with 105mm filter diameter can't possibly have F/2.8 and 300mm focal length at the same time, so it's obvious that Sigma conveniently rounded the actual numbers in their favour. This discrepancy is not due to laws of physics or the like, it was simply a marketing decision and possible buyers get less than they might think they pay for.
But let's please cut this mindless drivel "I lose focal length at MFD!!!" for good.