Using machine learning and historical event data to leverage predicted odds for future events

A system comprising:

• one or more processors;
• a non-transitory computer-readable medium storing a plurality of instructions, which when executed, cause the one or more processors to:
• obtain historical information for past events associated with one or more participants including predictive variables associated with the past events, and historical odds for outcomes of the past events;
• train a predicted odds machine learning model, wherein the predicted odds machine learning model is a neural network, to determine predicted odds, comprising:
• create, using the predictive variables as independent variable input and the historical odds as dependent variable input, the predicted odds machine learning model including:
• training a set of linear regression models by implementing forward stepwise variable selection using the predictive variables;
• and
• determining an average model of the set of linear regression models;
• obtain current information for at least a first upcoming event associated with the one or more participants including a set of predictive variables associated with the first upcoming event;
• and
• determine predicted odds for the first upcoming event by inputting the set of predictive variables associated with the first upcoming event into the predicted odds machine learning model;
• and
• a plurality of gait sensors pairs, each gait sensor pair comprising:
• a water-proof housing, a unique identifier, an accelerometer, a processor, an onboard clock and a wireless communication subsystem, and wherein each sensor pair unique identifier is associated with a unique identifier of a particular participant of the one or more participants, and wherein the plurality of gait sensor pairs produce accelerometer data and location data for a particular participant, wherein a plurality of onboard clocks are synchronized to a same time;
• wherein the one or more processors are configured to:
• associate, a first participant identifier of a first participant to a first gait sensor identifier and a second gait sensor identifier;
• associate, a second participant identifier of a second participant to a third gait sensor identifier and a fourth gait sensor identifier;
• display, via a user interface, the first participant identifier being associated with the first gait sensor identifier and the second gait sensor identifier;
• display, via the user interface, the second participant identifier being associated with the third gait sensor identifier and the fourth gait sensor identifier;
• send an activation command to the first, second, third and fourth gait sensors, wherein the activation command includes pairing information of associated gait sensors;
• receive real-time movement data comprising a first data stream of accelerometer data and location data from the first gait sensor, the second gait sensor, the third gait sensor and the fourth gait sensor, wherein the first gait sensor and the second gait sensor are attached to the first participant, and the second gait sensor and third gait sensor are attached to the second participant;
• determine based on the real-time movement data whether the first gait sensor or the second gait sensor has been attached to a participant other than the first participant;
• and
• input the real-time movement data as an additional set of predictive variables into the predicted odds machine learning model and determine current odds for the first upcoming event.