He has this chance because over the last 30 years, teams with a 1 run lead with two runners on and two outs in the top of the 7th win 78% of the time. Now let’s say the pitcher allows a 3 run bomb. They’re now down by 2, 2 outs, bottom of the 7th. So let’s just say historically in this situation, the home team wins 38% of the time. So that pitcher contributed -40% to his team’s WPA. It doesn’t matter who the pitcher is, it just looks at situations and how often they’ve happened in the past 30 years and uses the that huge sample to create a sort of probability for each situation. Has nothing to do with the players on the field, just uses historical situations to create a probability for current situations.