Brad Ausmus should not have pulled Victor Martinez

Victor Martinez has been on a tear since returning from the disabled list. He added three hits and a walk to his offensive resurgence Thursday afternoon, including a two-out single in the ninth with nobody on. With the game tied at 5-5, it appeared extra innings were in order.

Instead, manager Brad Ausmus played for the win in the ninth. He pulled Martinez from the game with two outs and inserted Josh Wilson as a pinch runner. The idea, as he stated in the postgame press conference, was to hope Yoenis Cespedes hits a double and Wilson could score from first -- something Martinez likely would not have done.

This decision was a big mistake on the part of Ausmus. Martinez is one of the best pure hitters in the American League, a player who can change games with his bat. Wilson is not. He's a journeyman player with a career .268 wOBA and just one major league stolen base since 2011. He does not grade out particularly well in the Ultimate Base Running statistic, available at Fangraphs, either, considering he rates out as a below average career baserunner.

The fact of the matter is that the odds that Wilson's running would have affected the outcome of the game are quite low. Run expectancy tables show that the average two-out, runner-on-first scenario scores results in 0.21 runs. That is a very low chance of scoring the winning run. Ausmus should have recognized this and held onto the advantage of Martinez's bat.

To put this another way, let's test Ausmus' notion that it was worth having Wilson in for a potential double from Cespedes. Cespedes has doubled in 22 of 298 plate appearances this year, or about 7.4 percent. Let's say that Wilson would have scored on 50 percent percent of those doubles and Martinez would have scored on 25 percent. This means that the advantage of having Wilson in the game over Martinez was an increase in Detroit's win expectancy by 2 percent in the ninth.

In contrast, the scenario in the 10th inning (bases loaded, two outs, and down by a run) had more of a split between the two players in terms of win expectancy. With two outs and the bases loaded, it's a simple matter: get on base and you've at least tied the game; don't and it's over. Martinez has a career .371 OBP, while Wilson has a career .280 OBP. That's a 9 percent change in win probability -- meaning that leaving Martinez in would have been four times more effective than pinch running.

There was no way for Ausmus to know that the fifth lineup spot would come up with the bases loaded and trailing by a run. However, it is no stretch to imagine a scenario where a bat is needed late in extras. The possible gains by pinch running were not worth taking Martinez's bat from the lineup.