That's not true at all.
- Japan having 10,000 troops in Alaska kept America having over 100,000 troops troops there. According to your argument, Japan should have sent those 10,000 troops to Guadalcanal. But that would also mean that the US could send 100,000 extra troops to Guadalcanal as well which means Japan loses there more quickly.
- If you are Japan, you want to have numerical equality on fronts/battles that actually matter. Occupying a irrelevant island like Attu kept the US from deploying troops and utilizing it's numerical superiority to places where it matters, which gave Japan a better chance of winning in Southeast Asia. Another example would be the Phillipines in 1944. By occupying a politically important, but strategically irrelevant territory, Macarthur and the US army were forced to fight to retake the Phillipines, which prevents them from being utilized in the Central Pacific Offensives (Which is the only front in the Pacific that truly mattered in the late stages of the war).
- As long as Australia deploys more men to the Darwin campaign than Japan does (Which they would as I have explained in previous posts), that means the Australians can't deploy their troops elsewhere in as large numbers as they did in otl. This means that a overland campaign in New Guinea and the occupation of Port Moresby without naval support could actually succeed, as could other otl campaigns where Japan was barely defeated by Australian forces.
- Also the idea that Japan doesn't have the manpower for this is inaccurate. As someone mentioned up thread, 25,000 Japanese forces sat in Timor during this time. Also, as Ichi-Go shows, Japan has the manpower even in the late stages of the war for offensive operations.