Can weapons end the war?

Suppose you're the world's policeman watching two states fight. Engaging directly is too expensive, but you can send weapons to one party which will improve its chances of winning. Should you do so? If yes, whom to help?

If you only wish to end the war, you should help the stronger state, i.e. the one with better chances of winning. Helping the stronger state de-incentivizes both parties from fighting: the weaker state becomes even weaker and exerts little or no effort to win the war; the stronger state can win by little effort, so there's no need for more. The result follows directly from the unfair Tullock's all-pay contest that economists use to model wars, lobbying, sports and other forms of competition.

Yet there's something wrong about this advice. In fact, we don't just want to stop the war; we want the 'right' state to win. For instance, under most circumstances we prefer a successful defence against the aggressor. If the aggressor is stronger - which is usually the case - the two goals dictate opposing measures. To end the war and save lives, we'd need to help the aggressor; to restore justice, we should help the defender.

The straightforward solution is to make the defender very strong. Then we'd have a quick war wherein the 'right' state wins. Yet, the solution is often too costly or too time-consuming. That being said, justice usually comes at a price of human lives. But isn't it the goal of justice to save human lives, in the first place?