I've been thinking about this.
1. Gentlemen's agreement
A gentlemen's agreement, provided it concerns all WMDs (i.e. all NBCs), might be slightly less unrealistic than some here think. The reason is that by now, even a direct war between major powers would probably not be a war for final supremacy like it would have been in the 1970s. It might well be more like the Seven Years' War, a situation in which the combatants try to gain some limited advantage over the enemy. Conquering border territory and border fortresses, colonies, spheres of interest - but without going all-out to finish off the enemy. In such a contest, both sides might eventually accept a defeat in conventional warfare, preferring not to up the ante to WMD use.
2. Casus belli
I think the casus belli needs some thought that I don't much see in the thread. It's less unlikely than one would think, if we combine dangerous mix of internal political necessities, miscalculations, and a plurality of parties involved. This might make such a war more likely, because it's a case of blindely blundering into war, as opposed to really throwing the dice against a peer power. Again, if the casus belli is something like that (as opposed to a cold-eyed and madly decision to go and invade everywhere), the unspoken agreement above and the scenarios below might be more likely.
3. Descent into war: cyberwarfare and non-conventional warfare
If that is the case, it's entirely possible that you don't go from Defcon5 to Defcon1 in one night or one week. The enemy might well start by "supporting" "local" "liberation movements" (non-conventional warfare) and by cyber attacks (accompanied by propaganda). It might come down to a situation where one civilian population doesn't see the need to have another economy slump (caused by serious cyberwarfare) for such a small disturbance (that border territory or country). One side would be more vulnerable than the other here - the side with more democracy. But once cyberwarfare is out of the box, this second side will also use it with no qualms.
4. Limited operations
Assuming this second side still decides to go to conventional war, I really don't see those Risiko-style strategic movements into the interior of the Asian continent. Who's got the bucks and the guts for those? All the more so if the casus belli is in a Danzig-like place. Both sides will initially try to limit the operations. They won't be very successful, i.e. the theater will get involved as a whole; but I don't see new theaters in very inconvenient places opened.
5. Air strikes
Even though the land front will be hot on that one most convenient theater, a sizable part of the war will depend on strategic air strikes. The combatants will try to cause damage to the other side's infrastructure, in particular power production, fuel production, key industries. The side with the longest reach, and secondarily the side with the best tech, will be favored here. It is entirely possible that this is the part of the war that decides the overall outcome, regardless of how things go on land. The side that is seeing its industries demolished will throw the towel in first.