How do you solve a problem like Pete Kozma?

So... Pete Kozma's been pretty bad this season.

That's not exactly breaking news, and I'd hate to do more than just pile-on, but Kozma's numbers have been HISTORICALLY bad this season and, even though he might be the 25th man on the roster, it matters.

Jeff Sullivan wrote a great piece a few weeks back on the importance of not having bad players. He added up the positive and negative WAR on each team, and some interesting things emerged. The Astros, while only ranked 15th in positive WAR, were best in terms of negative WAR with only -1.0 WAR from the entire team. The Royals also rated highly in this regard, and it jibes with what we see on the field from those teams: They may not be flush with superstars, but there's nobody lousy out there, either.

The Cardinals graded out very well in terms of not having bad players, ranking fifth. As of the All-Star break, only three pitchers have a negative WAR, Carlos Villanueva, Randy Choate and Mitch Harris. Villanueva and Choate are only 0.1 below replacement level, while Harris is at -0.2. Each of the 17 other pitchers the club has used were at or above replacement level. So that's pretty good.

On the offensive side, you'll find Ed Easley, "Thomas" Pham and Xavier Scruggs at -0.1 WAR, and then Peter Bourjos, who defensive metrics hate this season, at -0.2. Again, nothing much to see here.

But then smack at the bottom of the list are the perennial last men off the bench, Pete Kozma and Tony Cruz, both rated as 0.4 wins worse than a replacement level player.

Cruz's roster spot can at least be justified by the fact that a team needs a backup catcher, and the Cardinals system is barren when it comes to alternatives. Though his career WAR is underwater and you could certainly mount an argument the team should have sought a better option in the offseason, here we are, and it's hard to see better alternatives at the moment.

Kozma, however...

If weighted on-base average is the best overall measure of batting skill, Pete Kozma is the worst hitter in the National League. If wRC+ is the best measure, he's still the worst hitter in the National League. Michael Wacha, Lance Lynn, Tim Cooney and Tyler Lyons all best him by both measures. In fact, were Kozma's 2nd half to mirror his first, his would be among the Top 10 worst seasons in terms of wOBA for any player with that many plate appearances since 1970. He is currently the only player in the league with at least 70 PAs and an isolated power of zero.

In 2013, Ben considered whether Kozma's season was perhaps the worst in team history. At the time, his weighted runs created plus was a mere 61. This season? He has posted a wRC+ of 5.

Of course, Kozma is no longer the starter as he was in 2013. It is a bit of a luxury problem to worry about the production from the last guy on your bench, but as the Sullivan article highlights, not having bad players matters.

The offseason acquisitions of Dean Anna and (though he can't play shortstop) Ty Kelly, coupled with another year of development for Greg Garcia, seemed to suggest the Cardinals were intent on refreshing their bench infielder matrix. But Kozma was the one who found his way onto the opening day roster, and there he has stayed, with Garcia and Anna only getting brief cups of coffee.

Perhaps the club is reluctant to expose Kozma to waivers, which they would have to do were they to demote him. On the other hand, Kozma passed through waivers last season, and it's hard to believe his stock has risen based on 2015's performance. Or maybe they just think very little of Garcia's or Anna's potential, either to improve substantially on Kozma's hitting numbers or to consistently catch and throw the ball at shortstop. As bad as Kozma has been, it surprises me that they haven't at least given Garcia/Anna a shot.

But let's keep in mind the concept of replacement level: This is the level where you ought to be able to pluck any random jackass off the field in Des Moines or Toledo or wherever and have them put up comparable numbers. Sure, the reality is much more complicated than that, but when a player on your major league roster is on-pace to cost you nearly a full win over a season, and if you are for whatever reason convinced your own farm hands will be worse, you ought to be hunting for that replacement player.

Maybe it comes down to the Cardinals making a waiver claim of their own, or a small-time trade of spare parts, or maybe in the mix of a larger trade you can get someone to throw in a piece of their own organizational depth. I'm sure it's easier to advocate for this in the abstract than it is to pull together in reality, but we're not talking about acquiring some kind of impact player here. We are merely talking about finding someone - in the organization or without - who can even achieve replacement level. At this point, it seems exceedingly clear that Pete Kozma cannot.