Real Data, Scientific Models, Real Results

Machine Learning Model Won again

Yesterday's man-machine competition was like a roller coaster ride. First up was ML, the original leading competitor, in the first round. Everyone got one or less correct picks, but ML's tips were spot-on with three correct picks. Unfortunately, ML couldn't maintain the momentum. On the other hand, the two computer models, Naive Bayes and SVC, performed exceptionally well. They both achieved a remarkable 52% hit rate and took the crown.

Naive Bayes was Monkey's first developed model and is the simplest one. Its hit rate has just been average, but it seems that with the continues adjustments to the model and accumulating more data, its hit rate has been steadily increasing. This time it reached 52% and made a profit of $17, with a return rate of 6%.

SVC is a unique model with a consistently low hit rate, however because of its unconventional horse selection method. It often picks underdog horses. Surprisingly, this time it achieved high hit rate 52% hit rate and high return horses at the same time. It made a profit of $104, with a return rate of 39%.

A record-breaking 9 members participated this time, and some members performed well. Summer, for example, achieved a 44% hit rate and a 12% return rate. Everyone's performance is improving, and we look forward to the next competition!