Template Tennis

There are several components necessary to become a high performing tennis player. Body composition, athletic skill, mental toughness, neurocognitive (tactical) ability and stroke mechanics (technique) are the main constituents. The reality is that many of these are genetically coded. Others may be improved through instruction and practice.

The most obvious area for controllable intervention is improvement of stroke mechanics and associated shot quality. Surprisingly, many sectors of the tennis player development industry avoid meaningful attention to stroke development. There are several reasons for this. All reasons are fundamentally based in the complexity of stroke development.

Much of the complexity is rooted in the fact that multiple viable options exist for each stroke. There is no empirically based definition of the options. This makes coaching decisions for a player difficult. Further, there is no systematic way to assess progression in stroke mechanics development. Early development competitive results are commonly used as the assessment criteria, but this often leads to long term limitations to success.

The current project will address both primary complexities to meaningful stroke development. The goal is to remove the barriers that have affected both the industry attitude, and coaching implementation of the most controllable element to player development. Template models will be constructed for each stroke to empirically define the options available. Player video may then be compared to a desired model over time to assess player progression.

Decades of research into tennis stoke mechanics has yielded a clear definition of technique options for each stroke. These core options may be further subdivided by player attributes such as stance and level of play. These breakdowns form an infrastructure of model options. Within each option classification, a model will be constructed and continually updated.

Model construction entails utilizing research grade 3D motion capture technology (mocap) to compute a skeletal reproduction of many players while hitting the strokes. These reproductions may be grouped by stroke option classification then combined (averaged) with all other players in the group. This produces a single representative stroke from each group that provides a definition of the stroke characteristics of that group.

The representative stroke (model) from each stroke option group is output in the form of a 3D “stick figure”.  The stick figure provides a powerful visualization of the important characteristics of each stroke option. From a practitioner perspective, this provides an empirically based and easy to understand visual definition for each of the stroke options identified through the research.

The template model stick figures are numerical mappings for each stroke option. They may be used for comparison to any individual’s stroke to assess the current state of the stroke and to plan the next development steps. The comparison tool may be utilized in two primary ways:

1. Overlay Comparison: The model stick figure may be overlaid directly on a video recording of an individual’s stroke. This will allow direct visualization of positional and motion differences over time.
2. Computational Comparison: Quantitative attributes of an individual’s stroke may be compared to the same parameters from any model to produce a simple numerical score indicating how closely the strokes match.

All project components will be made available to practitioners through an “app” designed for mobile devices. The app will provide access to stroke model definitions, the overlay comparison and the computational comparison in conjunction with relevant stroke and instructional information. Combined, the app elements will provide a simple and complete treatment of tennis stroke mechanics – basic and applied.

Advances in computer vision aided by AI algorithms allow numerical body mapping of standard video from a mobile device. This mapping provides the information required for computational comparison of an individual’s stroke video to any template model. All by simply taking a video with a mobile phone.

For example, the trajectory (over time) of any body part or the racquet of an individual’s stroke recorded on video through the app can immediately be scored for goodness-of-fit to any of the desired models based on playing level, stroke classification and stance. Comparison can also be made to biomechanically optimized models determined through our research.

The app is designed to provide a systematic method to assess progress and plan development by providing visual and computational metrics (scores) at regular intervals (i.e. per week). Coaches may use the model of their choosing to guide students to their preferred stroke implementation through score progression over time. Players without access to coaching will be able to develop credible stroke mechanics easily on their own.

The power and ease of use of this app (record stroke video with a phone) should make it desirable to a wide swath of the tennis playing population. It provides users access to research grade biomechanics and results without the laboratory and education. As a result, the app has the potential replace the stigma of stroke mechanics as something to avoid, with the confidence to develop strokes as the most controllable improvement element – as technique improvement is in most other sports.