Glad to see this thread is still going; I just picked up playing again a week ago and had a few thoughts to share, in no particular order. (Some of these may have been mentioned before, apologies if I missed them)
I. Restrictions on Naval Repairs
In Jan. '42, Saratoga took a torpedo which forced her back to the mainland for repairs. This took her out of action for several months and risked another submarine attack on her transit to the states could sink her. If she'd been repaired at Pearl, she might have been available for the battle of the Coral Sea or the battle of Midway, significantly changing the balance of power in those engagements.
Currently, if I have a ship whose strength is reduced to 1, I can sail her into any port (say, Wake with supply 5), and repair her to strength 8 in one turn. Given the facilities actually present at Wake (i.e., no drydock, lack of material or experienced shipwrights), this seems unrealistic. I see two possible modifications which could provide an interesting solution. One is the addition of a new rule that would only allow the repair of a ship if the port supply was greater than 10 minus the ship strength. Now my badly damaged ship of strength one needs to make it back to Pearl or the west coast for repairs, and could be intercepted by the enemy or encounter bad weather and sink on the way back. In addition, if a port my ship flees to is damaged by bombing, I'll be unable to repair my ship in that port until it recovers. Another possible change is to split ports into "regular" ports and drydocks, and requiring ships damaged beyond a extent to be repaired in a drydock (perhaps even over several turns) instead of a port.
II. Convoy Raiding
One of the principal damages of convoy raiding was not simply the loss of material, but the loss of shipping capacity. Next time the attacked convoy of ships moved goods, it would move fewer goods unless the sunk ships could be replaced. I think it would add an interesting dimension to the game if the throughput of a convoy route was reduced for several turns as the result of a raid, rather than just for the turn it was raided. Raiding these routes would be similar to strategic bombing of a production center, reducing a city's MPP production from 10 to 5 causes the city to underproduce by 15 MPPs over 5 turns. Likewise, a raid on a convoy route which destroys 5 MPP of shipping could reduce the throughput of that route for 5 MPP for several turns, until new ships are built to handle the cargo.
III. Economic vs. Military Manpower
Nations were often faced with problems in allocation of manpower: larger armed forces lead to a lower economic output to supply those armed forces. I've never come across a game which models this well, but I think it would make for an interesting dimension to the game if it could be implemented well. I'm not sure I have a good solution in mind, but I do think it's work putting some thought into. Options might include reducing MPP output for each manpower intensive unit created beyond N units, with a small 'restoration' rate to mimic young children 'coming of age' and becoming economically productive. This could be simulated using events, too; i.e. the Japanese could be given the option to mobilize their population for the defense of the homeland in the event of an American force landing on the home islands, giving them a number of free units, but at the cost of permanently reducing the MPP production of their cities. I could also see connecting MPP production to national morale, or to the number of units lost and rebuilt. It's complicated to do right in a simple, intuitive manner, but manpower allocation was one of the most critical strategic decisions a nation could make and would fit nicely into the game.
IV. Amphibious Limits
I've seen mentioned before the desire to limit the number of units which can be amphibiously transported each turn; I do think this is a good idea, as sealift capacity was a critical bottleneck in amphibious operations. I think the most efficient way to do this is to create a new unit, the 'amphibious landing ship'. Any land unit can be loaded onto this unit for an amphibious attack. This naturally limits amphibious attacks to the size of the sealift capacity you've decided to build, and having these units sunk means you have to rebuild them, reducing your ability to conduct seaborne operations for a time.
V. Barbarossa - Alternate Scenarios
The timing of Barbarossa ('41 vs. a later year) and the near capture of Moscow provide some interesting bifurcation points in WWII, and I think are some of the most important strategic decisions not modeled by the game. A scenario which can completely capture these turning points would be exciting to play, but a beast to make. More specifically, giving the player an option to start Barbarossa in either '41, '42, or '43 would add some interesting variety to the game play (I usually modify any scenario I'm playing so I can start Barbarossa in '42, if I so choose). This is arguably one of the biggest and most important strategic decisions of WWII, and giving a player the flexibility to make that decision would, I think, be a step forward. Another potentially 'game-changer' is the Russian reaction to the capture of Moscow. It would add an interesting dimension if the Russians could either surrender, offer a negotiated settlement, or continue fighting (at present I believe they always continue fighting). The probabilities of offering one of these options could be dependent upon the date at which is Moscow is captured; the sooner Moscow is captured, the larger the chance they surrender or offer a negotiated settlement. This provides a strategic opportunity to gamble on taking Moscow while ignoring other, more economically rich areas, with the hope that Russia will surrender; it may or may not pay off. It also provides the Russian player more incentive to hold on to Moscow (depending on how this is implemented with humans playing the Allies), as the fall of Moscow could lead to the surrender of Russia.