Because trade is not "between countries" but between firms, relationships matter. If tariffs divert trade, it means that relationships between firms are broken, leading to seeking out new supply chains. Those relationships become "sticky" after tariffs are removed, meaning that things do not return to "normal." In other words, country-specific import elasticities of demand are often altered. In addition, prolonged trade disputes can also lead to changes in relative currency values, that further alter those relationships.
I think in a short time US is winner in the trade war, but in long time the strategy will unsuccessful. The Trump's administrations not able to manage world crises, The trade policy brought a better economic condition for Americans, but the American First Slang will bring a few crises in the international relations for the superpower.
First question: are you defining winners and losers on a national basis or are you defining them in terms of interest groups within nations?
If the former, as Derek Pyne pointed out a trade war, since it reduces total trade, reduces the gains from trade for all nations involved. In the longer term if a trade war ends in a peace that allows for greater expansion of trade it is at least possible that all gain. The longer the war and the smaller the eventual drops in trade barriers due to the peace treaty the less likely that the present value of the war is positive.
If the latter, of course some interest groups (classified by industry for example) in any country putting up barriers can gain--that is usually the point of starting a war. Others will lose, as their export goods are priced out of some markets. To make it worse, unless it is a very short war investment is directed to industries that, when peace is restored, will not be competitive. This is an often neglected longer term problem and one which means it is harder to make peace than to keep the peace, since those who invested in the "wrong" place have more to lose than before if peace is made.
The idea that countries set tariffs in response to their market power in international markets is an old result. I would trace it to the work of H. Johnson in the 1950s and 1960s on optimal tariffs. Basically, it relates to the price elasticity of demand for traded products. So, elasticities, market power (the size of your economy) are the basic ingredients you are looking at. It is a controversial result in international economics linked to the so-called « terms of trade » effects in trade theory. I guess that a Google search would return a lot of papers on this topic.
This is theory and yes, a large country can win a trade war if the proper conditions are set. But this is theory.
Previous answers provide important insights on the dynamics (something not treated in standard terms of trade literature).
How to measure welfare losses when trade is not only in final products?( @Darren Hudson reminds us that countries do not trade, firms do. Up to now, there is no good theory on trade when trade is in intermediate goods. And global value chains statistics indicate that the large firms are quite integrated, using lots of imported inputs).
There is a RG debate opened on this theoretical aspect, with many insights offered: https://www.researchgate.net/post/Inter-industry_trade_in_the_21st_century_Do_we_still_need_a_trade_theory?
What do you do once you have won a trade war? Keep the high barriers? But this has a welfare cost. Return to low tariffs ? But this might endanger the protected industries... @Neil H Garston ‘ s insight illuminates the point. By the way, the issue of time consistency extends also to all kind of industrial policy using protectionism (e.g, of the Prebish-Singer type).