AI as support for selecting and following up trotting horses for races

Important results from the project

To get more amateur trainers to register licensed trotting horses for starts, more focus is required in the proposition work. With AI´s capacity, additional selection criteria have been added to propositions, which contributes to an increased registrations, democratization, and audience experience. As a result of the pandemic, measurements of the audience experience have not been carried out. Tests of the ability to contribute to the quality assurance of races, have shown good results.

Expected long term effects

From creating conditions for more female and male trainers to registering horses for the start, to finished AI tools, the implementation rests on equal parts technology and collaboration with trotting players. Parts of the organization’s cautious attitude have proven to be an asset when with surveys identifying factors that have contributed to the AI solution´s functionality and user acceptance. The pre-conditions and the complete solution will soon be presented to trotting community in Sweden.

Approach and implementation

The project has been conducted in collaboration with the sports director, coaches, staff, and management to ensure function and acceptance. A machine learning method was developed which can assess the quality of a proposition and indicate the need for targeted efforts towards underrepresented groups of coaches. As training data for the AI solution, three years of proposition data at Jägersro have been used. The solution is now implemented and will be trained with further data

External links

Webbsida består av landningssida som beskriver funktioner samt filmad demo. Skarp version finns och visas efter ö.k.