ChatGPT useful for batting order

Welcome to Discuss Fastpitch

Your FREE Account is waiting to the Best Softball Community on the Web.

Oct 25, 2009
3,360
48
Has anyone tried ChatGPT using player game stats for determining batting order?
I entered our player game stats for the season out of curiosity. Very interesting batting order recommendation and explanation. I entered OBP, OPS, BA, SLG.
If you try this use the microphone. It understands and interprets normal conversation much like a human.
 
Jul 26, 2022
12
3
It actually gives really good recommendations. You can paste in an export from GameChanger and ask it what you should do and it will tell you who should bat where and why.
 
Jun 23, 2024
1
1
I actually went so far as to create a custom GPT for this purpose. It does fine, especially when considering basic hitting stats. I was most interested in how it would do in various complicated situations, including youth league settings where everyone has to bat, and the goal isn't always to maximize run-scoring potential. As with most things GPT, you can get decent results if you're persistent enough at tweaking what you're looking for. But by the time you're factoring in who is likely to bunt, steal, need an extra bat to build confidence, etc., you'd probably do just as well to put together the lineup yourself!

It may be more useful if you are in a situation where you rotate players defensive and want to save time by having it generate an inning-by-inning chart based on suggested positions (especially in a setting where winning isn't the top priority). When I just tried this, it did fine, but most importantly offered the following good advice:

Tips for Managing the Chart:​

  1. Communicate with the Players: Make sure everyone knows their positions ahead of time.
  2. Adapt as Needed: Be flexible and ready to adjust if a player is struggling or excelling at a particular position.
  3. Encourage Teamwork: Remind the players to support each other, especially when trying new positions.
 
May 29, 2015
4,070
113
Going to forgo my opinions on AI, and I will just leave this here to make myself smile. Saw this poster at Baseballism yesterday . . .


BP9CZbdCUAAsBpW.jpg
 
Jun 18, 2023
543
63
I am waiting for the first AI MLB Manager. We are there already with statistics driving the majority of decisions of who to pitch, how long, what batters, where to play, etc.

I am curious to see how MLB handles this.

the data is messy, anecdotal, and incomplete. How do you quantify how fatigued one pitcher feels versus another? I'm sure there's patterns you can derive from some of it, but nowhere near what you'd need to really to make pitching decisions.

Now, pitch sequencing? There's probably some hidden value in there that human eyes haven't picked out. But again, human element is gonna get ya. The batter gets to change his mind. If the data suggests, for example, a fastball up and in on 0-1 to lefties is abnormally successful, it won't take long for the batter to realize pitchers are throwing that, and adjust.
 
May 17, 2012
2,846
113
the data is messy, anecdotal, and incomplete. How do you quantify how fatigued one pitcher feels versus another? I'm sure there's patterns you can derive from some of it, but nowhere near what you'd need to really to make pitching decisions.

They already have sensors that provide detailed insights into a player's fatigue levels, stress, and injury risk.

Once an AI manager makes all of the correct (coaching) decisions (coaching is in the margins) your team will be at a slight disadvantage if you aren't using one.

My point is that we are essentially already there...baseball managers are the lowest paid in all professional sports as they have the least amount of impact. The math has been done and the GM's are imposing their will and will someday cut out the middle man (Manager).
 

Members online

Forum statistics

Threads
43,194
Messages
686,134
Members
22,251
Latest member
travis.goodnight
Top