Загрузка видео...
Не удалось загрузить видео
What actual HITTING looks like: training swing decisions within practice. The game is random; we don't know what pitch is coming or where it's going to be. However, very rarely do you see any training to mirror this. We want a pitch right down the middle every time.
305,331 просмотров • 1 год назад •via X (Twitter)
Комментарии: 10
Nothing like the actual game. I love doing this because the hitters can't cheat to a spot. Ball is moving all over, forcing them to lock in every rep. If you want to make good swing decisions, you have to train it.
This kid is all over the place.
Kids way off balance. Work on staying balanced after the swing
But no pitch comes from that angle.
My kid is a goof hitter, I give partial credit to his success to my terrible BP pitching, it’s definitely improved over the years and I can get in a groove, but he’s always been forced to be selective, also find ways to hit balls randomly all around the zone,
It’s called live batting practice 😂
Why does the ball seem to come from where the second basemen stands?
problem is, it’s still not replicating a release. Live pitching when not in the game still gonna be 100x better than any machine….but it is a good drill that’s my only reason for not liking off a machine, but you do what you can with what you have so no hate keep grinding
Can you worry about how unbalanced he is and how disengaged his lower half is before you worry about pitch selection
Try to set up to where the ball is coming from a realistic angle. That ball is released 2’ off the ground and another 2-3’ off center of where a mound would be.
!["One of the very confusing things about the models right now: how to reconcile the fact that they are doing so well on evals. And you look at the evals and you go, 'Those are pretty hard evals.' But the economic impact seems to be dramatically behind. There is [a possible] explanation. Back when people were doing pre-training, the question of what data to train on was answered, because that answer was everything. So you don't have to think if it's going to be this data or that data. When people do RL training, they say, 'Okay, we want to have this kind of RL training for this thing and that kind of RL training for that thing.' You say, 'Hey, I would love our model to do really well when we release it. I want the evals to look great. What would be RL training that could help on this task?' If you combine this with generalization of the models actually being inadequate, that has the potential to explain a lot of what we are seeing, this disconnect between eval performance and actual real-world performance"](https://image.24vids.com/tw-1993450075616690474/amplify_video_thumb/1993449024045932544/img/W-1TjOXcXeiUbHaH.jpg)
