Started By
Message
re: For years everyone said their rival school was cheating by buying players...
Posted on 10/27/23 at 10:59 am to scrooster
Posted on 10/27/23 at 10:59 am to scrooster
quote:
If someone is stealing signs ... someone else should be better at hiding signs.
"Regular" sign stealing I'm Ok with and agree with you. What I'm not OK with, and have no real indication is going on any widespread level is using machine learning to decode an opponents signs. With the right coding and programming it's not that difficult a thing to do. Youtuber Mark Rober did it as one of his videos a few years ago. If you've not seen it, it's worth a watch. I'm 100% sure this is what Michigan was doing and what Clemson was doing with a laptop on the sidelines in 2017...
Stealing Baseball Signs with a Phone (Machine Learning)
1
Posted on 10/27/23 at 11:24 am to Lonnie Utah
I really don’t think people understand how easy it is to break these “codes” once you have a games worth of data and a computer. In real-time it is difficult, with a computer it is a piece of cake. These types of solutions are literally things in undergrad stats or analytics courses.
What is even scarier is that you can actually write the code to capture formations and full route trees and use that with sideline image capture then to decide full sets of variable route trees within a play as well as what the hot audibles do, all in real time…
We are getting closer and closer to a world that to keep sports competitive we will need to impose rules about banning the use of real time AI during games as well…
Set up a videographer on the side line during the game and overlay it with your teams video feed recording of the all 22. Take the signs from the first drive and all 22 and let it run for 15 to 20 minutes in real time and see what spits out about play probabilities. While that is running do an additional run with the first 2 drives to refine the model, then 3 and ect. Most likely by halftime you have a large set of confirmed signs….
What is even scarier is that you can actually write the code to capture formations and full route trees and use that with sideline image capture then to decide full sets of variable route trees within a play as well as what the hot audibles do, all in real time…
We are getting closer and closer to a world that to keep sports competitive we will need to impose rules about banning the use of real time AI during games as well…
Set up a videographer on the side line during the game and overlay it with your teams video feed recording of the all 22. Take the signs from the first drive and all 22 and let it run for 15 to 20 minutes in real time and see what spits out about play probabilities. While that is running do an additional run with the first 2 drives to refine the model, then 3 and ect. Most likely by halftime you have a large set of confirmed signs….
Page 1 of 1
Popular
Back to top
Follow SECRant for SEC Football News