I am happy to see that the war in Asia is finally getting a little love but I am a little concerned on how it will play out. The Sino-Japanese war as modeled in the HoI series has never been done well. The Japanese AI over focuses on China and is never adequately prepared for a war in the Pacific in 1941. I had hoped that by splitting factory production into Naval docks and military production would have helped this but sadly it didn't. The Japanese navy is always behind the power curve when it comes to technological advancement of it's naval forces. The US faces no serious opposition in 1941-42 like it did historically. I have a feeling that by putting some meat into the Chinese forces will put the Japanese even further behind the eight ball because China will no longer be a cake walk.
Sino-Japanese War was not modeled correctly in any of the HOI series because you need to model it from supply and politics point of view. During WW2, China averaged around 100 *bullets* per active soldier per year, not to mention other equipments. That's not enough to train a soldier properly. As a comparison, Japanese issued 200 bullets per soldier per *month* for *training*. Other Allied troops are much more lavish.
The problem is China had too much troops to supply them all properly. Chiang was NOT stupid. Before the war, there were around 180 divisions, and he tried to reduce that in 1930 -- and he was forced to back down or face a civil war. Japanese invasion of Manchuria gave him political capital to try again, but Communist uprising forced him to delay force reorganization to 1935-1936. He was planning to reorganized the army to around 60 modern divisions, but war with Japan broke out before that was complete. After WW-2, he tried again, and the result was he was forced to retreat to Taiwan.
In game term, if Chinese division count drop pass a certain point, then China suffer a massive hit to national unity -- it may cause China to capitulate immediately due to territories taken by Japanese. If during peace, any of the warlords or communists exist, national unity drops, civil war happens, and the warlords/PRC get extra manpower. So the only time it is safe to reduce the number of divisions is when at peace and China is unified. But for this to work, China needs to have the ability to build divisions *without* all the equipment so it won't be force to capitulate -- and this has to be changed from Paradox side. So this means China is forced to build a lot of divisions that has no combat power, and that will screw up equipment and supply situation. This is much closer to the reality than all the workarounds. And this will allow the use of real historical OB for Chinese.
As for Japan, it will almost works better if Japan is treated as two nations -- Imperial Japanese Army and Imperial Japanese Navy. However, in generalized term, production and *research* should be allocated in fix percentage to army/navy/air force/infrastructure, and any change to the allocation has to be paid from political power (or focus/events). AI tend to fixate on one thing at a time, so this will place a limit on them. This will also require change from Paradox's side.