This, while technically a "stat", has zero predictive or analytic value. In hindsight, you could look at that and say "that pen was effective at holding wins during this sample size", but it doesn't tell you why, or give you any valuable information going forward.
To use an easy example, let us say a team had 78 games in which they were leading 20-0 before the 7th inning, and in 77 of those games, the bullpen allowed 19 runs to score, but in one of those games, the bullpen allowed 21 runs to score. By your stat above, that's a great bullpen, but by any realistic measure, that's a pretty f***ing terrible bullpen.