Inspiration
.we love gambling
What it does
gives us better information on how to gamble
How we built it
xgboost with model stacking and noise reduction as well as heavier weighting on the more recent ticks (using GPT 6.1)
Challenges we ran into
Can't break past 55 and 36 validation scores (probably good because that means probably a lower chance of overfit)
Accomplishments that we're proud of
mid 50s scores and high 30s score
What we learned
That our ideas can be easily prototyped and optimized with AI
What's next for House Edge
Do this irl and hopefully make bank