It’s been a couple months since my last post, so I feel I should give a quick update on what I’ve been working on.
Pitch Design tool
Since mid-February, I’ve been focusing on development of the pitch design tool I’ve been working on (with a short trip to Dublin/London mixed in). In that timeframe, I was able to talk to someone familiar with software patents, and the tool is now patent pending.
I’ve vastly improved the tool since the demo version I released in February. There are several new variables including a Pitch Score metric which assesses pitch quality on a scale of 1-10. In the new demo version, the Pitch Scores for pitchers’ current arsenals will be based on in-game results while the Pitch Scores for the to-be-released service for college teams/facilities/etc. will have the Pitch Score predicted by an ensemble of machine learning algorithms based on each pitch’s characteristics (spin axis, velocity, spin rate, etc.). Expect a detailed post regarding the validity of the Pitch Score metric to coincide with the launch of the service.
This service will likely launch early next week by my new company: Howell Baseball Data Services LLC. Early next week, I should have an updated version of the demo I released in mid-February out on my blog, coinciding with the launch of a service by my company where coaches/players/facility owners/etc. can send in their data to have pitch design targets generated and sent back in a report (I had built a version which would allow teams to upload their own data and use the app on their own, but fears of issues with future compatibility has led me to decide to keep the tool on the backend and generate reports for teams on an ad-hoc basis to give me more flexibility to develop workarounds if necessary). Pricing will be on a per player basis (price still TBD… it will be fairly high, but well worth it) and will include an optional follow up report within 3 months of the original report which will include new grades for pitches the pitcher worked on. All questions regarding the paid service should be directed to my business e-mail, email@example.com.
Here’s a sneak peek of the new demo version to be launched next week:
The reports teams can purchase using their own data will look similar to the picture above for each player.
Some other things
Starting in August, I’ll be a Graduate Assistant at LSU- Shreveport where I’ll be helping the Pilots better use data for both recruiting and development purposes. Recruiting and player development are the two main ways a college team can build a winning program, and they go hand in hand. Colleges which consistently demonstrate the ability to develop players will have an easier time recruiting, and teams using advanced metrics to prioritize recruiting will have a serious competitive advantage over schools which confine themselves to their region or use inefficient methods to evaluate players.
So, I’ve also spent some time creating a tool to gain an advantage in recruiting over the last few months, as well as a basic web app which generates very simple scouting reports of NAIA teams so I can put more focus on collecting and analyzing data for player development when I start in August. I’ll probably publicly release these tools eventually, but I’ll keep them under wraps for now. We may not have the resources large NCAA Division I analytics departments have, but I’m optimistic we’ll be directing our efforts better than most.