Meet @Berryhorse29, A Compelling New Voice in Sports Betting Modeling And Information

I first became aware of @Berryhorse when a few friends alerted me to his Twitter page. He’s gained quite a following thanks to his (first name Kieran) willingness to share his Major League Baseball betting model in an open Google spreadsheet, plus he provides insight into the model and his picks. Given his success this season, his profile on Twitter and Reddit has steadily grown.

According to his tracking, Kieran is 166-138 (54.5%) on the season , +524.2 units and +22.45% ROI. (Note: We did not contemporaneously track or verify these MLB wagers/positions, but there is no shortage of followers who can attest to his success and are willing to make donations, which he has declined to accept and instead has referred them to charities.)

Berryhorse’s relatively new Twitter account has grown to nearly 8,500 followers. He updates his followers with his daily wagers and new information.  What follows is a Q&A in which Berryhorse explains his background, his thinking and intentions with the model, and what’s to come for football season. (As he notes in this Twitter thread: “If you are new to my page, PLEASE read this thread before taking any action.” Further, here is his FAQ/Guide on the third tab of his open spreadsheet.)

Meet @Berryhorse, Creator of An MLB Betting Model Who’s Taking His Growing Social Media Following Along for a Math Course and the Ride.

Sports Handle (SH): To start, will you tell us a bit about your background?

Berryhorse (BH): I’ve been playing sports and watching sports for essentially my whole life. Basketball, football, golf, baseball and more. At the same time I’ve also always been a huge science and math nerd. I got into computer science my senior year of high school and majored in Computer Engineering and Computer Science in college. While at school I worked for an NBA analytics company, which was my introduction to machine learning and data science. Learning about this in the context of sports was awesome and eventually it became clear to me that it was possible to build predictive models to beat the sports betting markets.

Like most people I enjoy sports betting because of the money I earn, but by far my favorite part about it is that it quenches my thirst for competition. I love competing and winning and really enjoy that element of sports betting more than any other. I also love how it’s such a pure meritocracy. If you’re a winner, you’re a winner.

Over the long run you can put your money where your mouth is and prove your edge over the market. Of course the industry is full of scams and gimmicks but success in the actual betting market is purely based on merit. My favorite sport to bet on is baseball – highest volume of games, and when betting money lines the team’s incentives are always aligned with yours as a bettor, unlike spread betting.

SH: Can you explain how the MLB model works?

BH: First, I have to give a lot of credit to Joe Peta, the author of “Trading Bases.” Much of how the model works is based on his principles. As simply as possible, the model prices each team against one another by projecting how good each team is at run production and run prevention.

There is no simulation of pitches or at -bats or innings, it just says “team A is this good and team B is this good.” Once it has that information it can calculate a probability of one team winning the game and compare it with the Vegas line. When there’s a large disagreement between the two, it’s probably worth betting.

SH: How big a factor is the starting pitcher within the equation for each game?

BH: Starting pitchers matter as much as the specific starting pitcher matters. The run prevention number for the team obviously relies heavily on the starter and this is accounted for. Basically it calculates the RA number for the team if that starter pitched every game for them.

[For more stories like these, follow Sports Handle on Twitter @sports_handle]

SH: What’s the most memorable cover you’ve had this baseball season? Wild ending, walk-off, or otherwise?

BH: The most memorable cover was the White Sox beating Cincinnati in CIN a few weeks ago. Lucas Giolito got rocked for 4 or 5 runs in the first and after spoiling a bunch of chances in extras, CHW eventually won in 12 by a few runs (and covered the run line). Most memorable loss: recently Khris Davis hit a 2-run shot with 2 outs and 2 strikes to win by one. That hurt. The most painful/memorable loss is always the most recent.

SH: Why do you keep the spreadsheet open? Are you interested in the conversation/discussion?

BH: I’m sharing all of this for a few reasons:

#1 and most importantly is I think my process of betting can be helpful for people and encourage them to think more analytically and reject some flawed ideas/myths in sports betting.

#2 I think I can help people win money.

#3 least important, and selfishly, I think it’s a good idea to build a reputation in sports because a lot of my long-term goals are in this space.

SH: What do you have in store for football season? Expect to focus on NFL or CFB or both?

BH: Football: I have models for both college football and NFL but I do not currently plan on making either public. I will share as much as I possibly can though.

(Note: In June, Kieran and his college roommate Jerry launched the podcast “Sports Thoughts” in iTunes . Some words about the football model in Episode 7.)

SH: What’s the biggest lesson or takeaway, mathematical or otherwise, you’ve learned from the MLB season wagering?

BH: Biggest takeaway: Really the biggest takeaway is just simply the legitimacy of this model. I was expecting it to do well, but was not at all confident we’d see any season like I’ve enjoyed thus far.

SH: Can you offer some advice for beginners?

BH: I think one common mistake of beginners is not understanding the uncertainty of sporting events. In general, novice bettors are far too confident in their opinions and need to understand that sports are inherently very random and volatile. Typically this means new bettors skew towards betting a disproportionate amount of favorites.

Beginners should also be calculated with their risk allocation and bankroll management. There should be some element of betting more when they are more confident but still the maximum amount they bet should be a small portion of their overall bankroll. It is crucial to preserve and protect the sanctity of future earnings by never risking too much of their bankroll.