⚾ Orb MLB Model - Moneyline Machine 💸🤖
🚨New Orb Model Alert🚨
That’s right. After expanding our product offering into the NBA back in January, we are now diving into the world of baseball and the MLB. While the page itself has been quiet since regular-season basketball ended, the work behind the scenes has not been.
This MLB model was a long time coming for us, and I cannot go any further without publicly recognizing and celebrating the amazing work of Orb Apprentice. What started with a fun and natural next step of wanting to predict baseball games has turned into our most complex modeling system yet. He worked tirelessly for months, building a powerful training dataset with over 150 predictors dating back to the 2009 season, as well as an intricate daily API feed to match. Andy and I can’t thank you enough for your continued hard work on the Orb Project!
Sign Up For Free
I know some readers care less about the nitty-gritty details of our modeling process, backtesting, and current live results than others and just want to sign up for new free models.
If that is the case, no worries at all! You can sign up for free daily MLB model picks delivered straight to your inbox here:
About The MLB Model ⚾ 🤖
Every sport and associated gambling market is unique. Rather than try to fit one strategy to all our models across different sports, we believe in fitting our strategy to the sport and its data. If you have followed us from the start of this project and thought the NFL and NBA were unpredictable, get ready to learn the true meaning of chaos with the MLB. There is a reason this sport needs 162 regular-season games to separate contenders from pretenders. Baseball is inherently random game to game, with starting pitching rotations and a single swing sometimes deciding the outcome. But bats going hot and cold wasn’t enough to deter us from trying to predict the chaos, and potentially could help give us an edge over the daily betting markets.
Baseball markets are fundamentally different from what we are used to in the NFL and NBA. For those sports, the spread is a dynamic number that moves while prices typically stay consistent (within a given range). Here in the MLB, the spread stays consistent, always at 1.5, while the price can vary dramatically. This often leads to the inverse of what we have grown used to, where the moneyline prices in baseball more closely resemble the spread prices in the other two sports, like in the example below:
For this reason, and in line with our general modeling philosophy, we designed this model to be a moneyline predictor and focus on who is going to win the game, rather than by how many runs. But we haven’t forgotten the game we’re playing, and of course incorporate market data into the equation.
Using our most complete and powerful dataset yet, we were able to fit our most promising model to date. In fact, the model showed so much promise of predictive power throughout backtesting that every attempt to blend it with additional model types only hurt performance. So as far as the accuracy metrics and controlled testing were concerned, we somehow struck gold on our first attempt, a record for the Orb Project that I can’t imagine we pull off again.
I always believe in being transparent about our strategy so our readers know exactly what goes into every pick they see on this page. Rather than type out line by line how we get from a raw classification win probability to the pick that is delivered to your inbox, I had our good friend Claude break down our MLB modeling strategy and backtesting results:
Live Results To-Date
As always, you should take backtesting results with a grain of salt. The real question is how the model performs live, so that is exactly what we asked. On 6/22/26, we turned this thing on and started emailing out daily picks to a small group for QA assurance. Rather than accuracy, the main thing we focused on was the system’s tracking record, as this was the biggest issue with our automated NBA model.
Since going live, the MLB model has been nothing short of “on fire.” As of the time I am writing this, it has made 53 moneyline predictions, going 31-22, 58.5%, +12.65 units for an ROI of +23.9%:
Of course, this is over a relatively small and statistically insignificant sample size. Still, it is quite encouraging and within the ballpark of the ROI seen in backtesting results. As always, I should remind you that past performance is not a promise of future results when it comes to predictive models. The model’s current 84.6% accuracy when picking away teams, as well as its record above 50% when picking +money underdogs, are not exactly guaranteed to continue. In fact, and in line with every previous model we’ve released, the only thing that I can guarantee is that this one also will miss picks, will not go 100%, and may even lose units from here on out! But early results resemble backtesting and model accuracy numbers, which is as promising as it can be at this stage.
*The live results above were taken on 6/30/2026 and do not include the results of tonight’s model picks*
MLB model picks are brought to you by Novig!
🚀 Sign up today & use code ‘ORB’ for $50 in bonuses when you spend $5
🔥 Play Smarter with Novig – America’s #1 Sports Prediction Market 🔥
✅ Better Odds – Play against real users, with no house cut (VIG)
📲 Click the button below to download Novig from the app store - Must be 21+
Post Go-Live
This is our second time building a fully autonomous modeling system. Results so far have been great, but of course are not in our control. What is in our control is continuing to improve the system around the machine that makes the actual predictions. The tracking of results by pick type has matched up perfectly since going live on the 22nd. However, improvement does not mean perfection. We still expect bugs to pop up and would love feedback from you when they do! Again, a missed pick, a losing stretch, or even losing all the units it has won so far is unfortunately not a bug to report.
We will monitor the system internally and try to respond to any issues we notice or ones that are brought to our attention as quickly as possible.
Thank You!
None of the Orb Project would be possible without you, our readers. Thank you for your support throughout this whole journey, whether you were with us from the start or this is the first Orb model you are following.
We love building new products and continue to try and achieve our original goal of using the power of modeling to predict the future, and maybe even win some money along the way. Most importantly, we will continue to have fun whether our model picks 100% or 0%.
Follow along our newest, most exciting, and fully automated MLB model by signing up here, completely for free:
Let's have a great MLB season, Orb Faithful!
-Team Orb Analytics
.
.
.
.
.
.
DISCLAIMER:
The information provided on this website is for informational purposes only. It is not intended to be gambling or financial advice, and should not be relied upon as such. We are not responsible for any actions or decisions taken by readers based on the information provided on this website.
The picks and predictions provided on this website are based on our own research and analysis, and are intended to be used for entertainment and informational purposes only. We do not guarantee the accuracy or completeness of the information provided, and we are not responsible for any losses or damages incurred as a result of using this information for gambling or other purposes.
By accessing and using this website, you acknowledge and agree to the terms of this disclaimer, and you assume all risks and liabilities associated with your use of the information provided on this website.