We normally wouldn’t be posting about a competitor; however, we find it necessary to do so. Back in June we created http://poolstats.ai our approach to utilizing neural networks to analyze the game of pool. We also posted a video on YouTube and on poolstats.ai about our approach. It seems two months later an established company, Orange Loops, has upended us in this approach and adapted our idea with their team of experts.
They have created a dashboard with stats, like us, and are using our model and services as their own. Such is the world of business. However, something even more obscene is occurring. We are being phished for our formulas so they may use them in their stats calculations.
Recently, we received contact from a unsuspecting individual on Instagram, a new user, private account, no posts, no followers. We corresponded with this individual. Although we did not directly give our equations away, recent hits on our white papers and this correspondence leads us to believe that Orange Loops, the creates of drillroom.ai, is seeking the knowledge Pool Stats has worked so hard on.
Yes, we do publish our formulas, but that is for the reason of peer review, not for a company to steal them for their business model without attributing appropriate credit given to the creators of such deep analytics.
This is a cautionary tale of how the little mathematician created a universe only to be pushed aside by Goliath. We want to make our users and the public aware of these atrocities and are hereby formally releasing the matchup prediction model we have derived.
This model can, with high accuracy, predict the probability of the winner between two people, based on previous matches. It’s called the Pythagorean Win/Loss metric. It is derived from Bill James’ formula in Baseball, and has had several academic papers published proving its validity. If you see Orange Loops, or Drillroom.ai utilizing this metric, know that we have adapted it to the game of pool and that they are not the creators of this formula. Also, we didn’t disclose on how to properly fit the formula to a given set of data, and I’m sure they will be unawares of the statistics to do that.
In it’s infancy Pythagorean Win/Loss gives an accurate account of predicting games between a matchup. However, this model can be improved by fitting it the collected data. Something we’ve done and will continue to refine as more data pours in. For now, here is a glimpse at the Pool Stats Pythagorean Win/Loss Matchup Prediction model:
General Pythagorean Win/Loss
In the coming weeks we will release a formal white paper detailing how this formula is calculated. If Orange Loops and Drillroom want to steal this formula from us, we can’t stop them, we only ask that credit be given where credit is due.