Friday, November 17, 2023

stathead.com produced mistakes and caused needless, tedious, extra work.

This post will be sent to stathead.com so that the problems can be fixed.  support@stathead.com

Work on the post below from my Radical Baseball blog revealed problems and issues:

Saturday, June 22, 2019

528 pitchers OPS+: 525 below 100.

Updated 11/15/2023.

Careers, From 1903 to 2023, Played at least 50% of games at P.

For combined seasons, since 1903, Playing in the AL or NL, Played at P, in the regular season, requiring Plate Appearances >= 500, sorted by descending Adjusted OPS+.

This link should show the full list as of the date clicked: https://stathead.com/tiny/Ws6T9
________________________

Notice the creation year is 2019 but I updated it in 2023 because I noticed that the baseball-reference.com tool for ad hoc queries stathead.com had produced mistakes and caused needless, tedious, extra work.

Before I get into the primary issues: one of my criteria was to select baseball players who had played at least half of their games as pitchers. So I wrote "Played at least 50% of games at P". But when I did copy/paste of the criteria description of my query generated by stathead.com 50% or anything like it is missing. I still wonder if it was applied. Maybe stathead.com can address that and fix it.

A preceding post indicates that I would not necessarily know which pitchers might be on which list where criteria might be OPS+, Home Runs, Plate Appearances, Batting Average, ... 

Saturday, June 22, 2019
Pitchers batting is like place kickers playing quarterback.

OPS+ is an overall assessment of batting that gives a pretty good rating and is readily available online. 100 is league average. 

The pitcher with the best OPS+ was Doc Crandall who played from 1908 to 1918. Who? See what I mean.

Crandall was one of only three pitchers who had OPS+ greater than or equal to 100. Crandall had 122. Next two were barely at 100. So I checked on Crandall. He had played two seasons in the Federal League (FL) and I had not specified any leagues. When I ran it again in 2023 Crandall was still number one but down from 122 to 107. Reb Russell was second with 105.

But I noticed anomalies and ran the query again, this time with players among the top who had not been there before and some who had been there but now were not.

I clicked on different headings and noticed more anomalies.

As the title of my baseball post indicates there were 528 players found: 528 records. stathead.com limits them to 200 per page. Why can't we specify that number as a parameter? Clicking on the column heading of one of the pages sorts only the 200 records on that page and ignores the rest. Really, I'm not kidding.

There is other sloppiness:

1. You have to page down to the bottom to learn that there are more. It doesn't state the number at the top of page one.

2. It doesn't state the number of 200 record pages. You know, something like 1 of 3, 2 of 3, 3 of 3. So that you have some idea of where you are.

3. There's no way to go anywhere other than page by page. For instance, you cannot skip to the end.

4. The column headings disappear once you go down once, never to appear again, not even at the bottom. This is a real pain. In a spreadsheet you can freeze rows and/or columns.

In order to sort all 528 records, I had to go back and click on "Show Criteria" and alter the sort field to something else. I forgot to also keep OPS+ as a Criteria field, which then causes it to be included as among the left columns along with the new sort field.

When the data was found and on the screen I clicked Export Data (even though I've learned that the best you can do is copy/paste), then select "Modify, Export, & Share Table".

I picked X for some fields I thought filled too much of the screen to remove those columns. I had to do this needless, tedious, extra step for each additional run of a sort field, which was necessary because stathead.com sorts only on the records it presents on a page, remember. Argh! Why can't stathead.com have an option to save that view of the data, which includes moving fields, not just for one query but for others?

OK, that's enough. stathead.com is a really good tool and baseball-reference.com is a great source. But they could be better.