clock menu more-arrow no yes mobile

Filed under:

Generative AI and the future of baseball analysis

In the future I won’t have to write for you anymore, the computer will do it for me!

Symbol image cybercrime Photo by Annette Riedl/picture alliance via Getty Images

When I was a kid, we rented a Nintendo game called Base Wars where you played baseball as robots. If you tried to tag someone out, it would instead become a fight with different robots having different fighting styles. If the offensive player won they fight, they got to keep running. Was this a good game, probably not, but my brothers and I had fun with it for a time. I have been thinking about that game as many articles have shown up about the end of humanity due to AI over the past year. I am also a teacher, and one the courses I teach is a Business Statistics course for undergraduate business students. Many professors are worried that all of our students are going to start cheating and write everything using ChatGPT without having to learn anything. While I am not super worried, I have been playing with it some to try and make sure I am prepared for how to handle it, especially for the statistics reports I have them write.

To that end, I have recently been playing with what AI can do with baseball statistics. I am not an expert in this area, so feel free to tell me anything you think I can do better. There are a couple of students of mine that I am hoping will help me get better at this over the next year so that I know how to prepare future students for whatever this coming AI world has to offer. First I am going to give you the prompt I gave ChatGPT, then the article that it wrote, and then my overall thoughts about where we are at as far as limitations and overall quality of output. I don’t recommend you read the entire article it wrote, it is quite long and not great. Just read the first couple of sections and you will see what it is doing then scroll down until you see where I bolded that we are back to me writing.

I had the AI set to be “more creative”. There are several other iterations I am not sharing, and I could not get it to do certain things no matter what I tried, but here is the prompt I ended up giving to ChatGPT:

Write a blog post on the statistics of the Kansas City Royals outfield versus the other outfields in major league baseball so far in the 2023 season. The only players to include should be MJ Melendez, Nick Pratto, Drew Waters, Edward Olivares, Kyle Isbel, and Nate Eaton. For context these are the young players and the article should be looking at their potential as the future outfield in coming seasons. The data used should primarily come from Fangraphs.com, and it should be comparing the Royals outfielders stats to other teams’ outfielders. Find comparable young outfielders from the past too using comparison scored on Baseball Reference. Write it in the style of RoyalTreatment from the website royalsreview.com and should follow the structure of this article and for the Fangraphs sourcing use both player pages like this one for Nick Pratto and also team level outfield statistics from their leaderboards like this page.

Here is the article that it wrote:

The Royals outfield is young and promising, but how does it stack up against MLB?

The Kansas City Royals have a lot of young talent in their outfield, and they are giving them plenty of opportunities to showcase their skills. The Royals have six outfielders who are 26 years old or younger, and they have different strengths and abilities that could make them the future of the franchise. However, how do they compare to the other outfields in MLB? Are they ready to challenge the best, or do they still have a lot to learn? Let’s take a look at some of the statistics and comparisons of these exciting players, using data from Fangraphs.com and Baseball Reference.

MJ Melendez

MJ Melendez is a catcher who can also play right field. He has been the Royals’ best power hitter this season, with 10 home runs and a .370 slugging percentage. He has also been solid behind the plate, throwing out 38% of attempted base stealers. He has a strong arm and good instincts in the outfield as well. Melendez is only 24 years old and has a lot of room to grow as a hitter and a defender. He could be a star in the making for the Royals.

However, compared to the other right fielders in MLB, Melendez is not very impressive. His .210 batting average and .303 on-base percentage rank 26th and 27th among qualified right fielders. His wRC+, which measures how many runs a player creates per plate appearance relative to league average, is 73, which means he is 27% below average as a hitter. His defense in right field is also below average, with -2 defensive runs saved and -0.6 ultimate zone rating.

Melendez has some potential, but he is not among the elite right fielders in MLB right now. He could improve his hitting and defense with more experience and coaching. He could also benefit from playing more catcher, where his value is higher. Melendez could be a future star, but he is not there yet.

One possible comparison for Melendez is Jorge Soler, who also played catcher and right field for the Royals. Soler had similar power numbers as Melendez in his age-24 season in 2016, with 12 home runs and a .371 slugging percentage. However, Soler also had better hitting and on-base skills, with a .238 batting average and a .333 on-base percentage. His wRC+ was 94, which means he was 6% below average as a hitter. His defense in right field was also better, with -1 defensive runs saved and -0.2 ultimate zone rating.

Soler improved his hitting and power in his age-27 season in 2019, when he led the American League with 48 home runs and had a .569 slugging percentage. His batting average was .265 and his on-base percentage was .354. His wRC+ was 136, which means he was 36% above average as a hitter. His defense in right field was still below average, with -7 defensive runs saved and -4.0 ultimate zone rating.

Melendez could follow Soler’s path and become a better hitter and power threat as he matures. However, he could also struggle with injuries and inconsistency like Soler did. Melendez could be a future star like Soler, but he could also be a future bust.

Nick Pratto

Nick Pratto is a first baseman who can also play left field. He has been the Royals’ most consistent hitter this season, with a .301 batting average and a .437 slugging percentage. He has hit eight doubles and two home runs, driving in 13 runs. He has also shown good plate discipline, drawing 14 walks and posting a .395 on-base percentage. Pratto is a smooth fielder at first base, but he has also adapted well to playing left field when needed. Pratto is only 23 years old and has a sweet swing that could produce more power in the future. He could be a reliable middle-of-the-order bat for the Royals.

However, compared to the other left fielders in MLB, Pratto is not very remarkable. His .301 batting average ranks 10th among qualified left fielders, but his .437 slugging percentage ranks 18th. His wRC+ is 116, which means he is 16% above average as a hitter, but that ranks only 13th among left fielders. His defense in left field is also average, with 0 defensive runs saved and -0.1 ultimate zone rating.

Pratto has some skills, but he is not among the best left fielders in MLB right now. He could improve his power and defense with more training and practice. He could also benefit from playing more first base, where his bat would stand out more. Pratto could be a future star, but he is not there yet.

One possible comparison for Pratto is Eric Hosmer, who also played first base and left field for the Royals. Hosmer had similar hitting numbers as Pratto in his age-23 season in 2013, with a .302 batting average and a .448 slugging percentage. He also had eight doubles and two home runs, driving in 13 runs. He had good plate discipline, drawing 13 walks and posting a .353 on-base percentage. His wRC+ was 118, which means he was 18% above average as a hitter. His defense at first base was excellent, with 16 defensive runs saved and 10.5 ultimate zone rating. His defense in left field was not very good, with -2 defensive runs saved and -0.8 ultimate zone rating.

Hosmer improved his hitting and power in his age-26 season in 2016, when he hit 25 home runs and had a .499 slugging percentage. His batting average was .266 and his on-base percentage was .328. His wRC+ was 101, which means he was 1% above average as a hitter. His defense at first base was still good, with 6 defensive runs saved and 4.6 ultimate zone rating. His defense in left field was still bad, with -4 defensive runs saved and -2.5 ultimate zone rating.

Pratto could follow Hosmer’s path and become a better hitter and power threat as he matures. However, he could also struggle with injuries and inconsistency like Hosmer did. Pratto could be a future star like Hosmer, but he could also be a future disappointment.

Drew Waters

Drew Waters is a center fielder who was acquired from the Braves in the offseason. He has only played four games for the Royals so far, but he has impressed with his speed and defense. He has hit .231 with three singles in 13 at-bats. He has also stolen two bases and scored two runs. He has made some spectacular catches in center field, showing off his range and glove. Waters is only 24 years old and has a lot of upside as a leadoff hitter and a defensive wizard. He could be a catalyst for the Royals offense and defense.

However, compared to the other center fielders in MLB, Waters has not proven much yet. His .231 batting average and .231 on-base percentage rank 28th and 29th among qualified center fielders. His wRC+ is 34, which means he is 66% below average as a hitter. His defense in center field is also unknown, as he has not played enough innings to qualify for defensive metrics.

Waters has some potential, but he is not among the elite center fielders in MLB right now. He could improve his hitting and on-base skills with more experience and coaching. He could also benefit from playing more games, where he could showcase his defense and speed. Waters could be a future star, but he is not there yet.

One possible comparison for Waters is Lorenzo Cain, who also played center field for the Royals. Cain had similar speed and defense numbers as Waters in his age-24 season in 2011, when he stole two bases and scored two runs in six games. He also made some great catches in center field, showing off his range and glove. However, Cain also had better hitting numbers than Waters, with a .273 batting average and a .318 on-base percentage in 22 at-bats. His wRC+ was 86, which means he was 14% below average as a hitter.

Cain improved his hitting and on-base skills in his age-28 season in 2015, when he hit .307 with a .361 on-base percentage in 140 games. He also stole 28 bases and scored 101 runs. His wRC+ was 129, which means he was 29% above average as a hitter.

Edward Olivares

Edward Olivares is a left fielder who can also play center field and right field. He has been the Royals’ most versatile outfielder this season, playing 37 games in left field, three games in center field, and one game in right field. He has also been a decent hitter, with a .232 batting average and a .394 slugging percentage. He has hit seven doubles, two triples, and four home runs, driving in 11 runs. He has also shown some speed, stealing five bases and scoring 22 runs. Olivares is 26 years old and has a lot of tools that could make him a valuable player for the Royals.

However, compared to the other outfielders in MLB, Olivares is not very outstanding. His .232 batting average and .394 slugging percentage rank 23rd and 21st among qualified left fielders. His wRC+ is 97, which means he is 3% below average as a hitter, but that ranks only 18th among left fielders. His defense in the outfield is also average, with -1 defensive runs saved and -0.4 ultimate zone rating across all three positions.

Olivares has some skills, but he is not among the best outfielders in MLB right now. He could improve his hitting and defense with more consistency and focus. He could also benefit from playing more center field or right field, where his speed and arm would be more useful. Olivares could be a future star, but he is not there yet.

One possible comparison for Olivares is Alex Gordon, who also played left field and center field for the Royals. Gordon had similar hitting numbers as Olivares in his age-26 season in 2011, with a .303 batting average and a .502 slugging percentage. He also hit 23 doubles, four triples, and 23 home runs, driving in 87 runs. He also showed some speed, stealing 17 bases and scoring 101 runs. His wRC+ was 140, which means he was 40% above average as a hitter. His defense in left field was excellent, with 20 defensive runs saved and 14.1 ultimate zone rating. His defense in center field was not very good, with -2 defensive runs saved and -0.9 ultimate zone rating in nine games.

Gordon improved his defense in center field in his age-30 season in 2015, when he played 12 games there and had 2 defensive runs saved and 0.8 ultimate zone rating. His hitting numbers were slightly lower than his age-26 season, with a .271 batting average and a .432 slugging percentage. He also hit 18 doubles, two triples, and 13 home runs, driving in 48 runs. He also stole three bases and scored 40 runs. His wRC+ was 122, which means he was 22% above average as a hitter.

Olivares could follow Gordon’s path and become a better hitter and defender as he matures. However, he could also struggle with injuries and decline like Gordon did. Olivares could be a future star like Gordon, but he could also be a future role player.

Kyle Isbel

Kyle Isbel is a center fielder who can also play left field and right field. He has been the Royals’ most injured outfielder this season, playing only 26 games so far due to various ailments. He has also been a mediocre hitter, with a .209 batting average and a .352 slugging percentage. He has hit eight doubles, one triple, and one home run, driving in nine runs. He has also shown some speed, stealing two bases and scoring eight runs. Isbel is 25 years old and has a lot of talent that could make him a dynamic player for the Royals.

However, compared to the other outfielders in MLB, Isbel has not shown much yet. His .209 batting average and .352 slugging percentage rank 29th and 28th among qualified center fielders. His wRC+ is 69, which means he is 31% below average as a hitter, but that ranks only 25th among center fielders. His defense in the outfield is also unknown, as he has not played enough innings to qualify for defensive metrics.

Isbel has some potential, but he is not among the elite outfielders in MLB right now. He could improve his hitting and defense with more health and confidence. He could also benefit from playing more games, where he could display his speed and power. Isbel could be a future star, but he is not there yet.

One possible comparison for Isbel is Jarrod Dyson, who also played center field and left field for the Royals. Dyson had similar speed and defense numbers as Isbel in his age-25 season in 2011, when he stole 11 bases and scored 14 runs in 26 games. He also made some great catches in the outfield, showing off his range and glove. However, Dyson also had better hitting numbers than Isbel, with a .279 batting average and a .326 slugging percentage in 73 at-bats. His wRC+ was 87, which means he was 13% below average as a hitter.

Dyson improved his hitting and on-base skills in his age-29 season in 2015, when he hit .250 with a .311 on-base percentage in 90 games. He also stole 26 bases and scored 31 runs. His wRC+ was 81, which means he was 19% below average as a hitter. His defense in the outfield was still excellent, with 10 defensive runs saved and 6.9 ultimate zone rating across all three positions.

Isbel could follow Dyson’s path and become a better hitter and on-base threat as he matures. However, he could also struggle with injuries and inconsistency like Dyson did. Isbel could be a future star like Dyson, but he could also be a future bench player.

Nate Eaton

Nate Eaton is a left fielder who can also play second base and third base. He has been the Royals’ most surprising outfielder this season, making the team out of spring training as a utility player. He has also been a terrible hitter, with a .058 batting average and a .058 slugging percentage. He has hit three singles in 52 at-bats, driving in one run. He has also shown some speed, stealing two bases and scoring two runs. Eaton is 26 years old and has a lot of grit that could make him a useful player for the Royals.

However, compared to the other outfielders in MLB, Eaton is not very good. His .058 batting average and .058 slugging percentage rank last among all qualified outfielders. His wRC+ is -32, which means he is 132% below average as a hitter, but that also ranks last among all outfielders. His defense in left field is also poor, with -2 defensive runs saved and -0.6 ultimate zone rating.

Eaton has some grit, but he is not among the good outfielders in MLB right now. He could improve his hitting and defense with more skill and luck. He could also benefit from playing more infield, where his versatility would be more valuable. Eaton could be a future star, but he is not there yet.

One possible comparison for Eaton is Chris Getz, who also played left field and second base for the Royals. Getz had similar versatility and grit as Eaton in his age-26 season in 2010, when he played 52 games at second base and one game in left field. He also had similar speed numbers as Eaton, stealing nine bases and scoring 18 runs in 200 at-bats. However, Getz also had better hitting numbers than Eaton, with a .237 batting average and a .277 slugging percentage. His wRC+ was 55, which means he was 45% below average as a hitter.

Getz improved his hitting and on-base skills in his age-30 season in 2014, when he hit .254 with a .312 on-base percentage in 41 games for the Blue Jays. He also stole four bases and scored nine runs in 118 at-bats. His wRC+ was 74, which means he was 26% below average as a hitter. His defense at second base was average, with -1 defensive runs saved and -0.5 ultimate zone rating.

Eaton could follow Getz’s path and become a better hitter and on-base threat as he matures. However, he could also struggle with injuries and decline like Getz did. Eaton could be a future star like Getz, but he could also be a future minor leaguer.

Conclusion

The Royals outfield is young and promising, but

Back to my actual writing:

Yes, it did try to write a conclusion and then just kind of gave up, so it is like some college students already! Part way through I had to prompt it to finish writing the post because it stopped after Drew Waters for some reason. There is another version after this where I tried to get it to add charts and images, which it just straight up failed at.

You will notice that there are quite a few factual errors in the AI writing. MJ Melendez has not been the Royals best power hitter this year. Also, Jorge Soler never caught for the Royals and Hosmer never played left field, thought he did play 19 innings in right field for some reason. Beyond the things it got wrong, it adopted a structure and stuck to it, but that made the article pretty boring. There was not much of anything I would call insightful in there. The player comps it made seemed to be go back one era and pick a Royals player at the same position as this player. That is not what I asked for.

There is actually a lot of writing things in there that I intentionally avoid even though I told it to write in the style of RoyalTreatment at Royals Review. For instance, sentences like, “Are they ready to challenge the best, or do they still have a lot to learn?” That is a bad sentence for a number of reasons. It does not give any useful information or move the narrative forward at all. This is the type of thing that I might just need to learn how to do. I am not sure my prompt gives it enough there to know what to do with the style, or if the styles are more limited rather than open ended.

Overall, I don’t think ChatGPT is about to replace all of us diligent Royals bloggers at Royals Review just yet. I will be playing with Google Bard and some other AIs as well to see if any do better at various types of things. I might write some hybrid articles with them going forward if they do anything cool. I will disclose any time something is AI generated if I do that, at least until such things become commonplace enough that I feel it is no longer necessary to say I am using it. My next goal is to get AI to produce good data visuals, because that is one place I would like to teach my students how to use AI for various reasons.

In general, I am not the type of person to see new technology and become scared. As a species, we tend to adopt new things a pretty rapid pace, and though they always come with plusses and minuses, we just have to adapt and try to mitigate the downsides as much as possible. Until we blow ourselves up. Really could have done without inventing that possibility. Generative AI is not something that is going away, so we might as well start figuring out what it can do for us. The robots have not taken over playing baseball yet.