Sorry - you're saying that we're dismissing the complexity of the game, yet you're boiling down small sample sizes to a single number.
WAR is extremely useful in the aggregate / long run, but it's next to useless if you're taking it in one or two game chunks.
On an individual game level, momentum matters. That tends to even out in the long run IMO (thus my affinity toward aggregated statistics like WAR), but in a game like last night when your leader is "out of it" and makes mental mistakes, it's contagious. No, he didn't give up the two home runs to Naylor, but you can't tell me that pitchers don't feel the extra heat.