As you've already mentioned, cells near the primate macula tend to make one to one connections. Due to this lack of convergence they can be somewhat smaller than normal cells (particularly in the size of their dendritic arbors) earning them the moniker "midget" ganglion and bipolar cells (also P cells). By reducing the magnitude of photoreceptor convergence onto ganglion cells, the retina sacrifices luminal sensitivity for spatial acuity. Or put another way, a downstream neuron in the LGN/V1 can be more "certain" about which photoreceptor is sending information when light is bright, but less certain that any information is being sent when light is dim.
The other side of the question you bring up is how the retinal ganglion cells themselves alter the image formed onto the photoreceptor layer. For those that don't already know, photoreceptors are located at the very back of the retina. Since they're the major light sensitive part of the retina, that means light has to go through the ganglion cells, bipolar cells, amacrine cells, and all of their associated processes before reaching the actual photosensitive part of the eye. So, why not put photoreceptors at the front of the eye? One major reason is metabolic, photoreceptors are tightly coupled with the retinal pigment epithelium which helps to re-process used photopigment as part of the retina's vitamin A cycle.
How can those two factors, a one to one correspondence and a decrease in thickness be present at the same time? For reference, let's locate a cross section of a primate macula
Dark areas are cell nuclei. The innermost layer of cells (bottom of the image) are ganglion cells, the next layer of cells up is bipolar cells (the inner nuclear layer) and the next layer up is photoreceptor cell bodies. you can make out the cilial stalks of the photoreceptors pointing upwards toward the RPE.
From this image we can see a few transformations happening in the macula:
- The photoreceptor layer is thicker, up to 7 or 8 cell bodies deep. This means higher photoreceptor density.
- The outer plexiform layer (photoreceptors to bipolar cell connections) is considerably thicker. A single cone should connect to fewer bipolar cells in the retina, so how does this make sense? We'll find the answer in #3.
- The outer nuclear layer is only one cell thick. This is the first layer that is actually thinner in the foveal region. The cell bodies for photoreceptors from tis region have been displaced laterally, and the outer plexiform layer from #2 had to be larger to accommodate this lateral shift
- The inner plexiform layer, connections between bipolar cells and ganglion cells, is nonexistent. Also absent is the ganglion cell layer itself.
So how do these observations resolve your question? Quite simply, the lateral movement of information from the foveal/macular photoreceptors happens predominantly in the outer plexiform layer. That buys the inner plexiform layer considerably more space to further spread those signals laterally to ganglion cells, and avoids the problematic situation you suggested of having an extra-thick ring of ganglion cells immediately around the fovea. The end result is that ganglion cells that respond to photoreceptors in the foveal region of space are even further out from the retina than the bipolar cells they connect to, which are themselves displaced.
As far as why a cherry red spot looks so large, that honestly depends on two factors: the disease or disorder in question and the zoom used. Bear in mind that the red color comes from, as far as I know, the blood vessels in the choroid itself. It would therefore not necessarily correspond exactly to the size of the macula, but instead to any area which is thinned in that vicinity. I think you'll become more comfortable with that idea if you look back at the thickness of the total retina in the above figure and notice that even outside of the foveal region it is still growing thicker.
I think your question shows good instincts, because I believe that if the eye evolved to have a larger "true" fovea then you would get exactly what you describe: a large area of super high density vision surrounded by a region with non-linearly worse vision designed to support the central region. Instead it appears that the primate macula is just large enough that we have a moderate sized central acute region and a (relatively) linear drop in spatial acuity moving out from it.