Bullpen problems keep Twins from winning as often as they should

In this age of baseball analytics and advanced stats, there is a clear winner when it comes to the biggest gap between the fanciest name for a specific metric and the level of sophistication of the thing it actually measures.

The winner: Pythagorean won-loss record. If you are searching your brain for long lost facts about geometry and triangles, wondering how this measure relates to the theorem of the same name, you can stop. All it means is what a team's expected won-loss record is based on a formula tied directly to how many runs they have scored and allowed.

You can then compare a Pythagorean won-loss record to an actual record and draw conclusions about whether a team has underachieved, overachieved or properly achieved.

Baseball Reference displays Pythagorean won-loss records for teams prominently, and it sent me down a rabbit hole recently as it pertains to the Twins under manager Paul Molitor and previously under Ron Gardenhire. Here are some relevant numbers and possible takeaways:

• The Twins began Thursday with a Pythagorean won-loss record of 38-45 — three games better than their actual 35-48 record. That suggest they are a little better than their record indicates.

Overall in 3½ seasons under Molitor, they are six games worse than their Pythagorean won-loss record suggests they should be. In their two winning seasons, they were two games better in each. In the 103-loss disaster of 2016, they were seven games worse.

• In 13 seasons under Gardenhire, the Twins were 18 games better than their Pythagorean won-loss record says they should have been.

In fact, every one of Gardenhire's six division-winning teams finished ahead of its expected record — suggesting the Twins overachieved in the regular season to a degree, which is perhaps part of the reason they fell flat in the postseason.

As things turned sour after 2010, the Twins were minus-3 in Gardenhire's final four seasons — all of which were 90-plus loss years.

• A team's record in one-run games correlates strongly to its Pythagorean won-loss record. Molitor's Twins are a combined 55-83 in one-run games — including a putrid 4-16 in one-run games this year entering Thursday. If they were merely 10-10 in one-run games this year, they would be right around .500 overall.

Whereas Molitor's Twins teams are 28 games under .500 in one-run games, Gardenhire's Twins teams were a combined 30 games over .500 in one-run games in his 13 seasons. So clearly Gardy was a better manager, right?

Well, that's debatable. Perhaps this is an explanation: Gardenhire had more to work with, particularly in the bullpen — an area I might argue is the largest predictor of success in one-run games.

Molitor's teams have never finished better than No. 21 in bullpen ERA and rank No. 24 this year, nor has he enjoyed the stability of a lights-out closer like Joe Nathan (who averaged 41 saves and a 1.87 ERA from 2004-09 while making four All-Star teams).

Long story short: If you're looking for a reason the Twins have suffered in recent years, look hard at the bullpen and the stalled progress of all those young hard throwers who were supposed to save the day.