I wouldn't call a lack of competent Russian officers in 1904 surprising, given the system they came from. But anyway, the Russo-Japanese War was a big gamble for the Japanese and I don't think it could have gone significantly better without major changes in its course.
Smaller improvements would include:
- a more successful Mukden. The Japanese achieved complete tactical victory, but the Russian army remained in being because they could not sustain the pursuit. A successful encirclement, probably realistically beyond the capacities of the battered Japanese forces, but initially envisioned, would effectively remove the Russian army in Manchuria from the equation and would allow Oyama full control of any land he cared to march into.
- no Second Pacific Squadron. If the Russians hadn't gone with that crazy idea, the Japanese fleet would have been available to screw with Kamchatka or closely blockade and shell Vladivostok. Both might make bargaining chips come the negotiations.
I still suspect the best POD is still that the British government, for any reason, decides to stand by its ally and does not allow them to be squeezed out of much of their gains at the peace conference.
If you want to think really big, have a war result from the Dogger Bank incident. The Royal Navy will sink the Russian fleet, and with no real front to open elsewhere, London will lavish loans and equipment on the Japanese so they can keep gnawing away at the Russians. With the 1905 revolution on the horizon, a collapse of the Russian warfighting capabilities on the scale of 1917 is possible, and that could well mean all of Manchuria, Sakhalin, and potentially even the Maritime province, Kamchatka and the Russian Aleutians (though I doubt the Japanese could hold on to those gains in the long run). But that's effectiely WWI ten years early, so I don't think it qualifies for what you want.