Tangotiger Blog

The poster suggests that for “top pitchers” and closers (who are top relievers), leadoff walks are worse than leadoff singles. Any thoughts on why that would be the case?

Walks may be more indicative of them not having their stuff/command that day, and not as influenced by BABIP? And since e.g. Mariano regularly walked 2 batters/9 while allowing 7 hits/9, a single walk might provide more signal than a single hit. Just a thought

Ben:

This is not necessarily a point of fact.  While the numbers are accurate, notice two things:

1. The poster shows since 2008, and including Troy Percival in there is going to be basically just two seasons or something.  Rivera will be missing most of his career, etc

2. Random Variation is not accounted for

Certainly, I can show you even MORE relievers with that .60 / .34 kind of split.  Even pitchers like Kershaw.

But I can also show you the reverse.  When you look at it overall, the main conclusion shown (that .94 / .93 number) is the bigger takeaway.

You can’t draw any other conclusion without accounting for random variation and sampling bias.

And I commend Bailey for not going over the top with their conclusions, really focusing on that .94/.93 number and leaving the other splits really more as a talking point.

The thing that occurs to me is that walks and single might be distributed differently by context.  It seems likely to me that leadoff walks are going to be slightly more likely from hitters at the top of the batting order, and leadoff singles proportionally more likely from hitters at the bottom of the order.  If this is true, what we see here might actually be evidence that leadoff singles are more valuable than leadoff walks, since the expected runs from [runner on 1B, 0 outs] is going to higher for top-of-order hitters than bottom, and yet, the actual scoring is equal.

The data from the top 5 pitchers kind of seems like corroboration of this theory.  Those pitchers are not nibbling around anyone, and walks do produce more runs than singles - because those walks I feel certain are disproportionally drawn from the top of the order.

In other words, you need the expected wOBA of the following hitters to adjust for.  Otherwise bias is influencing the result due to high-walk hitters being disproportionately placed at the top of the order.

https://blogs.fangraphs.com/effectively-wild-episode-2204-is-a-leadoff-walk-worse-than-a-hit/

At the 58 minute mark, Bailey does give a reasonable point regarding the leadoff walk given up by a good pitcher.  That said, it’s not a conclusion, just a conjecture.  So, with a reasonable question, we can go about finding the answer.