The Program: Implementing Pitch AI
PitchAi is one of the more detailed and High Tech tools that we utilize in our program. It's become part of our assessment process, and allows us to test measurements from the athlete on a level that we can not see with our eyes, or even slo-mo video. Some of the metrics may seem daunting, and the process may seem difficult to implement/use but we have found easy ways to implement, make adjustments, and build the athlete using this data provided. This piece of tech is a great addition to your program, and is part of our overall pitching/throwing assessment.
Implementing this into your program will be one of the fastest ways to increase value and accountability of your programming. This is a ton of data you will be collecting, and as with any system you do not want to be working out the system more than you get to take advantage of the actual data. Do not give yourself another job, instead create a system that makes your job easier. There will be a couple of nuance items that may be mentioned, but this is what we use and I really do feel it creates more value for our program.
Prep & Backdate
PitchAI Coach Subscription is like any program with an individual player profile for every athlete. You enter in all their info, height, weight, handedness, and then you're ready to capture. If you opt for just a player subscription enter in the mean for your team. I believe this influences some metrics such as torque, but you get the most bang for your buck that way.
Since the video can be taken outside of the PitchAI app and uploaded to be analyzed, you will want to start taking old video that you have saved from the face on view of the athlete and uploading to see some of their prior metrics. PitchAI will automatically use the current date & time, so you will have to go back and edit that information on each upload. This is also a great way to familiarize yourself with some of the metrics and what they mean/look like.
Standardize Your Data Collection
You will want to standardize your approach to collecting data. This makes it easier on everyone in your program to be able to participate in the collection of data, and since it's a standard approach, you will never have to worry about who took the video.
Educate Athletes on what you will be using the data for. In our program we use it for testing biomechanics on high intent fastballs. (We also take data from other pitches, but mostly we focus on the mechanics of a fastball. Athletes are told about other pitches, but we program for fastball until they no longer need it, which hasn't happened yet)
We also tell athletes that we will go from whatever form they feel most comfortable with (Stretch or Wind-Up). As long as we are taking velocities they will find out which is more effective for them. Typically after the first throw and velo call out everyone goes from the wind-up. Which is fine, if you feel you need to work individually with someone on the stretch, have them throw from the stretch in one of their bullpens (More on this later)
We will take 5 fastball throws data on PitchAI and a radar gun. For programming and information we look at all 5, and also top 3 in terms of throwing velocity. We want to know what your body is doing when you move fast, but also be able to compare that to every throw you made in the session.
The distance at which you take the video needs to be standardized. This is getting better as PitchAI gets better, but having a set distance makes everything easier when it comes to others taking video as well. Educate your program on this. Same goes for the height of the video. Use a sharpie to mark your tripod and keep it as much as the same environment as possible.
Make sure you remove the "Speed Up" effect that phones do when taking slo-mo video. Many phones will only slow down the movement, and not the entire video. You want clean video or it will mess with the data.
Environment is one of the most important aspects of data collection. We will be building an environment that is easily repeatable, and will be the best way for you to collect as many athlete baselines in a day as you want.
We will be collecting 5 Fastballs with velo data as well. You'll want to set up your tripod to it's distance and height, bring your camera, and set that up before you have athletes show up. They will throw 5 fastballs from either the stretch or the windup, whichever they choose, be consistent for those 5. You will want to make sure that they do not do them 1 right after the other, and take their time in between each pitch. As they throw their fastball, record the video in slow motion, and write down the velo for each video you record. If you miss a video, radar reading, or the video is poor quality make sure you make a note of this. You want the velo readings and the video to be lined up perfectly.
After taking their 5 throws and velos you'll make a report from PitchAI. The "Report" page will take all of the pitches you input into it and create a baseline of their averages for those 5 pitches. I also suggest taking the top 3 of the pitches and creating a report for those as well.
Using a repeatable environment will also help you see what changes have occurred over the last couple weeks. Not everything needs to be repeated every week, but being able to take a baseline, and recreate the environment in 4-6 weeks will help you see which direction their metrics are trending. Creating a test-retest system that can show change objectively is extremely important for showing athletes their results and change.
Now that you've created a baseline, you can then take video every time they throw a bullpen and use that to compare to one of their previous baseline sessions. This will help you see if you are moving in the right direction with your cues and drill-work. Since you need to be actionable with this data it is very important to continue to re-assess. This method of taking video every bullpen session every week after the initial baseline is more of a rolling assessment. The difficult part with this is if you re-assess every week, you may find new problems, and think that the old issues are gone, when they may be gone for 1 bullpen session, they may not be as "sticky" as you'd like for the athlete. This is why taking a rolling assessment every week, and continuing to roll with your initial plan is so important. Your baseline assessment will tell you which direction to go, then you can check in and make sure it's moving in the right direction. I strongly recommend against taking info every week and completely changing up the next week's sessions as some athletes will revert back, and you will end up having not accomplished any goal in the 2 weeks of training.
PitchAI has a compare mode that will let you pick any 2 pitches, I suggest one from the baseline session, and one from the weekly collection, and compare those two pitches with metrics, as well as an overlay. This can be extremely helpful when trying to show an athlete what changes they have made towards their goals in 1 week, and be used to reinforce their training methods.
Once you've exported your data, you'll want to find the most efficient way to attack your individual players throwing mechanics. This can be difficult to try and plan for every single person without some sort of "bucket" system. But with the immense amount of data that PitchAI can be utilized for, we can also see exactly how drills will impact their mechanics. I first saw this from Lennon Richards at BDG, and thought it was an amazing way to implement individualized drill-work for the athlete. This takes a significant amount of time to be done, and likely increases your Assessment into a week long procedure, but it can be extremely valuable to find what drills work best for the athlete.
I don't have a ton of experience with this aspect, but I have started to utilize it with athletes, and it has been a great way to see what will help them progress more quickly over what "I think" is best. Most drills seem to make sense, and can be manipulated with cues, but finding the best way to get the athlete to refine their mechanics without too many cues and adjustments to the drill seems to be best.
Here are 4 drill examples and how PitchAI saw them. This was huge for finding what drills attack what goals for the athlete. We took 2-3 videos of each drill, and provided 0 cues to the athlete other than these drills were to be done at roughly 70% effort level. This was an easy way to get out of the way, and see what the drills actually do for the athlete.
The following results are from one athlete performing these drills, and being able to see some differences in his throwing metrics compared to a regular pitch. Some of them are more useful than others when it comes to changing mechanics. This is where we start to utilize the more effective drills in their programming.
Scap Retraction Throw
After seeing that PitchAI can analyze drill work as well as just pitches we took the 4 drills that we had metrics from, and started to implement them based off of their metrics, and how they correlated to our overall goals. This individual approach took a while. These were collected over 2 separate days, plus a day for the bullpen. Although the plan is to cut it down to only a couple of days, it is still a more lengthy process than just taking PitchAI on pitches and implementing a throwing program based off of those results.
As of right now we have tested only those 4 drills, but the plan is to add in a couple more variations, and even test out cues within those drills.
Building A Plan of Attack
The approach we take in our program is based off of 3 Traits and 4 Characteristics. This works for us, and has been an easy way for athletes to understand what they need to work on, and why they have certain drills and environments in their programs. It's extremely important to inform your athletes on your approach, and why they do what they do. In order to use this system you will pick 1 Throwing Trait and 1 Mechanical Characteristic. These then will be what you emphasize when you create your programs.