I've owned the game quite awhile and have played off and on over the last couple of years. I always found the system rather opaque but repeated playing give greater insights and eventually I start learning and gain greater understanding. I welcome rebuttals, it will improve my game. I'm pretty sure I'm far from mastering the game yet.
Observations
New units, appear semi randomly where ordered, a better method should exist to show ACTIVE arrivals. currently I see best way as purchasing regionally to find them easier after they arrive.
AI, very powerful it beats me like a drum.
Combat: at times it seems skewered. In current game Longstreet (def 6) was defeated in battle of Richmond Fortifications by Banks (off 0) The CSA lost worse than 2:1 in casulties despite entrenchments to an incompetent they outnumbered 28000-16000. Can not imagine this happening often, if ever in real life. Yet it is all too common of my experiences in the East. Union incompetents beat and beat consistently Lee, Jackson and Longstreet. Yet historically they lost one battle Gettysburg (Antietam a tac win, strategic loss) and won or drew the remainder, though later victories over Grant usually Phyrric. It feels more like Franco-Prussian War than ACW at times.
East
AI small Union forces have no problem wandering all over the deep south from the earliest days. Capturing their supply bases seems to have no effect. The effect is that the CSA in east seem pinned early, Once these forces reach a certain level the CS is forced to counter them to protect rail lines, no possibility to start any Northen invasions ever seem to open up. Feels more like '64 even in late '62 (a common problem in ACW games).
West
Though usually similar to East my CSA did better in current game, April '64 and CS controls Vicksburg, Memphis, NO, Bowling Green and Nashville. In late-62 most were Union (this is '62-2-theater) in '63 and overmatched by Union units that had just taken Meridian I sent AoT to Nashville, took it then BG, cut rails to Miss. and amazingly US forces retreated, CS then took Memphis. Forcing the US back felt "right" yet this doesn't show in East where every general eventually retreated to D.C. after setbacks. AI does not unless US is fighting over Manassas.
US should have greater impetus to retreat to D.C. US should have greater need to protect supply bases, at least until later in war. Player controlled forces wandering around enemy territory tend to vanish, AI should too, unless I'm missing something.
Defending units should have tremendous advantage particularly with good leaders something I'm not seeing.
CS takes higher losses in every campaign I've played, unsure if from combat flaw or over aggression.
Leader inactivity, units seem to move too easily, particularly into combat. Occasional random event pins a particular unit but by and large most leaders (particularly low strat values) should be difficult to get into battle, with no leader in force even harder. Reforming this would probably fix my percieved problem with massive AI forces wandering about the map.
Union Manpower
Rumors to the contrary the US never had a great manpower advantage over the south until late in the war, The reason is inefficient enlistment lengths, causing constant churn in troops. Unclear if this is modeled.
Best example was Hooker went south in 5/43 with 110,000, Meade got the army a month later now sized 90,000 despite a number of new units added. But troops disgusted with the conduct of the war mustered out in droves after Chancellorsville, a constant union problem.