I think we still need to work on scale and redraw maps. I never really envisioned Garnia for instance being much larger than, say, France or maybe Germany, but merging the old maps together makes the various nations huge. The steppes should be immense, I have no problems with that. Khazarak should be like the Himalayas and Hindu Kush mountains or more. Tirnakaur should be the size of the Iberian peninsula roughly. Frodia is slightly smaller than Garnia and Wodanslund is 1/2 the size of Frodia.
Based on the scale of the original maps, Garnia has a size of roughly 1000 miles by 700 miles (700,000 sq. mi.). France is 260,558. Germany only 137, 847. The Iberian Peninsula is roughly the size of France, 224, 507. So if Garnia is the size of France, it needs to shrink to about 36% of its current size.
On the other hand, Alaska is 663,268 sq.mi., roughly the size of Garnia on the ancient maps. My point is that leaving the countries with the scale that already exists in the old maps is not an unworldly concept. Even if the land area of Alaska is only 571,951, surely some of Garnia is water as well... Texas is #2 for US states at 268,580, about 40% of Alaska, so roughly in line with your ratio of Garnia:Frodia. (Frodia on the maps is actually about 800 by 500 or 400,000 or about 57% of Garnia). I envision Frodia with more water area than Garnia though, so land mass area (on the old maps where inland and coastal waterways not shown very well) of Frodia:Garnia could very well mimic that of Texas:Alaska.
The old map is drawn as if the rather boxy landmass now called GW is largely surrounded by water, but this is likely not true, and easily attributable to the vagaries of early cartography, perhaps coupled with relatively little long distance shipping. Assuming that the old water boundaries are, in fact, largely continuations of various land features (for instance the steppes continue to tundra to the north, with a possible polar ocean largely covered with icecap) and extension of the western landmass (which you have effected with the placement of the Romans in the far west, essentially "off-map" for the ancient maps) which would logically extend south some. Quite simply after sailing along to the south and east of the great desert a spell, our ancient cartographers extrapolated, mostly incorrectly.
I extrapolated the size of the Himalayas at about 255,000 sq.mi., so if we use the smaller sizes and use the Himalayas (although in a less spread out kind of more bunchy way) as the guidelines for Khazarak, that works. If we go bigger, then the mountains of Khazarak are either roughly 3 times bigger in area than the Himalayas, or there is some non-mountainous terrain over in there somewhere that our ancient cartographers misrepresented, or Khazarak is simply not as big as the old timers believed, and some of the western territory is in fact the Romans.
I am in favor of the original scale. This makes the "known world" roughly the size of the United States & Canada (together about 7,000,000-ish) and makes it easier to expect that in 1000-2000 years these far flung cultures mostly have not encountered each other. If we shrink it, using your still-not-perfect-but-more-to-scale map the whole thing (including the Romans) comes in about the size of Alaska - about 798,000 sq.mi.
My point is, I am going to draw it at the large scale for a new map and a globe, and I guess this was my reasoning why...