Let me save you all some time. There's about 2 minutes of RF serve analysis, the rest is a major Bert plug! ❌️ Should change the title of the video for full transparency
A commercial this is. And an apt demonstration of Mark Kovacs’ strange fixation with the back hip in his arbitrary and flawed serve analysis and model. Do NOT squat on the serve! And focus on useful details in Roger’s serve to better yours!
Might be worth pointing out what devices are going to support this. It would suck to preorder and then not be able to use it because you don't have an iPhone, for example.
iPhone and Android. Prob need at least an iPhone 8 and whatever the Android equivalent is, but I'm not the expert here so take what I'm saying with a grain of salt.
you guys are so b2b sales. I suggest changing to be more b2c so people actually cares about it and not get scared (but don't trust me i'm just a stranger on the internet)
We actually do have enough training data - lots and lots of serves. First pass training the model just wrapped after ~80 hours of training. It's coming along!
@@fuzzyyellowballs but you'd have to go through all those serves and for each one identify the good and the bad components so the model could learn what traits are required. Is Kovaks analyzing all this training data and inputting it into the model?
Step one is to run a bunch of serves through a model to make sure it can accurately detect everything - joints, the racket, the ball. Because out of the box stuff isn't accurate enough. Step two is taking Mark's "ideal model" for a serve and programming it in, which he's working with the developers to do. Then we can compare your serve to the ideal model and tell you what's off. Plus measure everything and track your improvement. Will be pretty cool.
Let me save you all some time. There's about 2 minutes of RF serve analysis, the rest is a major Bert plug! ❌️
Should change the title of the video for full transparency
Thanks man!
If you guys can pull this off it would be insane. Excited to see it when it's 1.0
Thanks! Looking good so far
80% of this video is propaganda, that's some bulls...
It's new, what do you expect?
A commercial this is. And an apt demonstration of Mark Kovacs’ strange fixation with the back hip in his arbitrary and flawed serve analysis and model. Do NOT squat on the serve! And focus on useful details in Roger’s serve to better yours!
what is the analysis software you are using? is it a commercial SW or are you developing it by yourself?
Developing it ourselves
Coach will I ❤ that turning forehands method
Interesting kinesiology concept. Hip speed. Relevant to sex as well?
Might be worth pointing out what devices are going to support this. It would suck to preorder and then not be able to use it because you don't have an iPhone, for example.
iPhone and Android. Prob need at least an iPhone 8 and whatever the Android equivalent is, but I'm not the expert here so take what I'm saying with a grain of salt.
I love you too
This is Kamala Harris level
you guys are so b2b sales. I suggest changing to be more b2c so people actually cares about it and not get scared (but don't trust me i'm just a stranger on the internet)
great waste of time of a video.
I just don't see how there'd be enough training data to make this remotely accurate. Cool concept, but I'd be shocked if it actually works.
We actually do have enough training data - lots and lots of serves. First pass training the model just wrapped after ~80 hours of training. It's coming along!
@@fuzzyyellowballs but you'd have to go through all those serves and for each one identify the good and the bad components so the model could learn what traits are required. Is Kovaks analyzing all this training data and inputting it into the model?
Step one is to run a bunch of serves through a model to make sure it can accurately detect everything - joints, the racket, the ball. Because out of the box stuff isn't accurate enough.
Step two is taking Mark's "ideal model" for a serve and programming it in, which he's working with the developers to do. Then we can compare your serve to the ideal model and tell you what's off. Plus measure everything and track your improvement. Will be pretty cool.
@@fuzzyyellowballs Thanks for explaining. I'm hoping for the best!
Me too lol :-)