Hi Vaughn, you ask a simple question, but it has a very difficult answer I suppose. I have very little experience in adjusting the difficulty, but I've helped to beta-test the recent DLC's, and I consider myself an 'average' player with regards to skill. So I think I have some idea of what constitutes a 'medium' difficulty.
First of all, I think Razz1 has it basically covered, I would like to add that it helps being very critical at all stages. But I'd also like to give a more concrete answer. I think there are some questions that you can ask yourself while designing a scenario that should help keep an eye out for the level of difficulty encountered by the player.
Such as: What kind of scenario is it? What war year, what factions take part, is it defensive or heavily scripted, does it have special victory conditions that require the player to really tailor their strategy? Is the weather important, what terrain will be encountered, lots of close terrain, rivers, or just open desert everywhere? What size/sort of core would the player have normally?
You can emphasise different aspects and strategies while designing, such as speed and mobility, time constraints (see the stock Moscow scenarios), or being able to take losses without getting a 'Pyrrhic Victory', like many late-year DLC scenarios. Too much of one aspect will degrade the others, and if you try to demand too much from a player (win against the Sovjets in 1944 while having a completely Italian core
) it will turn into a frustrating experience.
I think it would help to have a simple design at first, with a clear focus on the goals you want to achieve. A scenario which has a clear 'mission' will help focus when designing, but it might be too 'rigid' and will require more testing and scripting. But you will also gain experience in scripting while designing, which might help you in improving the design.
If we're talking about a heavily scripted scenario, their main design advantage is that they are predictable, they will require extra attention when scripting but will be able to be fairly balanced easily with some feedback.
A 'loose', unscripted scenario may be more difficult to balance, for example see the multiplayer scenarios. But they can be easier to adjust, and give the player more flexibility. It could make difficulty adjustments a bit more 'mathematical', with adjusting general experience/core size/prestige levels, and being less worried about the exact strength and timing an AI counterattack should have.
Overall unit strength depends on several factors:
What class is it, what exact type, what strength does it have (10 is normal, overstrength will help some types of units to perform much better). For some units the amount of experience matters a lot (fighters, late-war tanks). Some units also offer specific advantages in certain scenario types. For example, in the early years British Matilda or Sovjet KV tanks can be serious opponents when the Allies are defending in bad weather, but if they are attacking over great distances in fair weather they are much less threatening. A pre-placed bridge engineer or strongpoint can be very decisive, see for example some GC '44 West scenarios.
During a scenario that is not part of a campaign, it helps to add up the tally of prestige points available on the map; cities, airfields, Victory hexes. A defensive scenario will give a player far less prestige because taking objectives will not be the goal, while a large, lightly defended map will net the attacking player a lot of prestige simply by being able to take many cities without too many losses.
Take a look at some of the multiplayer scenarios, which often have a prestige-per-turn income. Last MP map I played, 'MP Spoils of War', had 250pr per turn for each side, but depending on playing style the nr. of core slots is more limiting than the amount of prestige. I went for cheap quantity end ended up with a surplus of prestige and a full core prepared to take losses, which were heavy but could almost be replenished at will due to the surplus of prestige. My opponent went for quality but had difficulty in filling up all of his core slots, and in later turns he had to choose between reinforcing or purchasing extra units, and while he could inflict heavy damage on specific targets he could not adequately cover the whole front. It was very close in the end, so a good design I guess.