Monday, October 7, 2019

A New Baseball Stat Leads to a New Baseball Record

If you've watched baseball for more than a decade, you are aware that the game has changed, and not for the better. Games take longer, there are more pitching changes, fewer complete games, plenty of home runs and strikeouts, and not nearly as many balls in play. In fact, 2018 saw more foul balls than balls in play. Great if you are a kid, but not very interesting if you like to see a bit of action on the field.

Another way to tell that the game has changed is from the statistics. As a simple example, homers have increased, hitting a ridiculous 6,676 this year, up 62% from 4,186 in 2014 (that was the lowest total since 1995, before steroids). Hits, however, were up a mere 1% in that time. The difference is that singles are down 8.7% and triples down 7.5%. I like to see rallies rather than a bunch of solo shots, but I understand that most fans prefer home runs. Still, devaluing them by changing the composition of the baseball is not the way to go, despite journalistic shills like this. Note that there is no mention of the juiced baseball that led to all these homers. When Justin Verlander had the temerity to mention the changes to the baseball, he was reprimanded by Joe Torre and others. The lack of integrity in the game is appalling.

Anyway, these are the obvious ways to figure out that the game has changed, but there are other stats that take a bit of digging, but reveal just as much. When I was in Pittsburgh last week, I saw Cincinnati reliever Raisel Iglesias give up a walk-off homer to Pittsburgh's Kevin Newman. It was the 12th loss of the season for Iglesias, which in itself is nothing unusual. But then I checked his stats, and he had only pitched 66 innings. So he lost a game for every 5.5 innings he pitched. That seemed like a very low number (the Reds lost 87 games with 1,438 IP, or 1 loss every 16.5 IP). Of course, relievers will have a lower number, since they pitch fewer innings, often in close games. Still, 5.5?

I decided to do a bit of investigation on the stat Innings Pitched per Loss (IP/L). Using Python, I grabbed all pitching data from 2000-2019, limiting the result set to those pitchers who had lost 11 or more games. I wanted to see if any other pitchers came close to Iglesias (who threw one more inning on Saturday to finish with a 5.58 IP/L). The lowest number before this year was in 2010, when none other than Charlie Morton, then with Pittsburgh, went 2-12 while pitching 79.2 innings (6.64 IP/L). But Morton started all 17 games in which he appeared; he just sucked that year. There are many starters who had terrible seasons with a lot of losses, but they usually averaged at least 5 innings per game, so their IP/L should be at least 7 when you include no-decisions. An example of this is Homer Bailey in 2018 - he started 20 games, went 1-14 with 106.1 IP for an IP/L of 7.6.

To make the comparison meaningful, I had to focus on relievers. In 2004, Luis Ayala of the Expos appeared in relief 81 times, pitching 90.1 innings and finishing with a 6-12 record. That gave him an IP/L of 7.53, the lowest for a reliever in the 2000s (again, with a minimum of 11 losses). I then used Baseball Reference's play index to get a quick idea of what had happened before and found that Gene Garber went 6-16 in 1979 while pitching just 106 innings in relief, for an IP/L of 6.63. That seemed to be the lowest I could find. So it looked like Iglesias would be setting a record. But then I noticed that another pitcher had done even worse this year. Adam Conley of Miami lost 11 games while pitching only 60.2 innings, for an IP/L of 5.52. Take a bow Adam, you are in the MLB record books!

Now, I chose 11 losses arbitrarily, because Iglesias lost 12 and I wanted to see if anyone had lost that many while pitching so few innings. But by looking at hurlers with 10 or fewer losses, you do find even lower IP/L results. In 2017, Sam Dyson lost 10 games between Texas and San Francisco, tossing just 54.2 IP for an IP/L of 5.42. Look at 9 losses and you have Edwin Gonzalez in 2004, who started 10 games and pitched just 46.1 innings for an IP/L of 5.15. As you reduce the number of losses, you get crazier stats, such as Montreal's Woodie Fryman in 1983, who pitched just 3 innings, losing 3 games before retiring at age 43.

So let's use the 11 loss minimum to keep this stat meaningful. Which means the 2019 season saw the lowest two IP/L numbers in baseball history. What does that mean? Not much. Teams change pitchers more often than they used to, so naturally relievers will see fewer innings thrown. And with the ball leaping out of the ballpark, more late inning lead changes are happening, especially to crap pitchers. Combine those two factors and you have Conley and Iglesias setting ignominious records that were completely ignored by everyone else.

You are probably wondering if the sister statistic IP/W is affected. I'll look at that in the next post.



No comments:

Post a Comment