First, make sure you are logged in to the website. Then, go to the home page or click “predict” in the menu. You will see “Choose a Sport” on the left (top for mobile). Select the sport you’re interested in (like NBA or MLB) from the dropdown menu. A list of upcoming games for that sport will appear. Choose the specific game you want a prediction for by clicking on it. Once a game is selected, the “Predict” button will become active. Click it to start the AI analysis.
Yes, you must be logged in to use the prediction feature. Sign up here for $9.99 all-access pass.
Currently, you can get predictions for NBA and MLB games. Support for NFL, CFB (College Football), and CBB (College Basketball) is planned for the future.
After choosing a sport from the dropdown, a list of available upcoming games will load below it. Click on the row for the game you want. The selected game will be highlighted, and the “Predict” button will become clickable.
Clicking “Predict” tells the AI to start analyzing the game you selected. You’ll see a brief confirmation animation, and a new “pending” entry for that game will appear at the top of your “Your Predictions” table. The AI analysis takes some time to complete, typically a full minute or two. If it’s been a few minutes with no change, try refreshing your page for an update.
To ensure the system runs smoothly for everyone, there’s a limit on how many predictions you can have processing at the same time. Typically, this limit is 3 (we may offer a feature down the road to increase this limit). If you hit this limit, you’ll see a popup message telling you how many are currently running and what your limit is. You’ll need to wait for one of your pending predictions to complete before starting a new one.
Once a prediction is “Complete“, click the “View Analysis” button in the last column of that prediction’s row. A modal window will pop up showing the detailed thoughts from the AI components (often referred to as “Jack” and “Jill”), including factors they considered, potential key matchups, roster impacts, and how they reached the final prediction. (See Screenshot 2)
This output shows a detailed, multi-stage AI analysis for the game (e.g., FLO vs. TOR). Here’s what each section means:
Priority Roster/Injury Data: This lists the players expected to be Active or Injured/Out for each team based on data feeds available before the analysis began. This is the primary roster information the initial AI steps (“Jack” and “Jill”) use. Keep in mind that later injury news might become available, which the final step (“Bob”) tries to incorporate.
Vegas Betting Lines: These odds (Moneyline, Spread/Puck Line, Total) are shown for context only. The AI makes its prediction based on game analysis first and then compares its findings to the market lines. The odds do not dictate the AI’s prediction.
Jack’s Initial Analysis (and Jill’s Analysis): These sections show the detailed reasoning from the early stages of the AI process (“Jack” and a second opinion “Jill”). They break down the game considering roster impacts, key matchups, team strengths/weaknesses based on player availability, goaltending, special teams, 5-on-5 play, and other factors like home ice. They conclude with an initial prediction and probability.
Bob’s Synthesis & Final Decision: This is the crucial final step. “Bob” acts as the concluding agent. It reviews the earlier analyses, explicitly notes any critical last-minute updates (like a confirmed injury mentioned), synthesizes all the available information, weighs the arguments, compares the analysis to the betting lines (again, for context), and provides a final rationale explaining why the final prediction is being made.
Final Prediction: Located at the end of Bob’s section, this gives you the official output:
The system automatically checks for predictions that seem stuck. If a prediction is pending for too long (e.g., over 4 minutes), it will automatically try to restart the processing job. You generally don’t need to do anything, but if it persists, please contact support.
This page displays statistics about the historical accuracy of the AI’s predictions. You can usually filter these stats by sport (All Sports, NBA, MLB). It typically shows:
This table provides a consensus view of predictions for recent or upcoming games, based on all user predictions run for that game. It shows:
You can filter this table by date range (Today, Yesterday, Last 7/30 Days) and often filter to show only games where the average probability was above a certain threshold (like 65%) using a button.
Contender & Champ subscriptions show GLOBAL averages, which is a huge edge and what our system uses to compile parlay of the day picks!
This section automatically highlights a few of the AI’s highest-confidence picks for the current day, presented together as a suggested parlay. It looks for games where the AI has the following:
It’s generated automatically during specific hours (usually 9 AM – 5 PM ET) and requires at least 3 qualifying picks to display.
Just like the Parlay of the Day except we add a condition for “Home Team Favorites” and slightly more strict criteria – that’s the only difference, but we find significant improvements to accuracy given these minor changes.
It’s generated automatically during specific hours (usually 9 AM – 5 PM ET) and requires at least 3 qualifying picks to display.
The Hitter Advantage Score is a proprietary metric used for MLB games to quickly assess which team might have an offensive edge in a specific matchup. It primarily considers factors like each team’s overall batting performance (often looking at stats like On-base Plus Slugging or OPS), the quality of the opposing starting pitchers (factoring in their ERA and WHIP), park factors, and potentially the strength of the bullpens.
The score indicates a favored team and a “point” differential, which is then categorized into a strength level like Low, Medium, or High. For example, “PIT (Med – 4)” would mean Pittsburgh’s offense is favored with a medium strength advantage, calculated to be 4 points better in that matchup according to the HAS model.
It’s designed to be a reference point, especially useful when it aligns with other indicators like Vegas odds or detailed AI analysis.
The “Overall Advantage” is our advanced game analysis metric for MLB games. It provides a sophisticated, data-driven comparison of comprehensive team strength for an upcoming matchup.
This advantage is derived from a specialized machine learning model that has been trained on a wide array of game factors and historical team and player performance statistics. These factors include, but are not limited to:
Our model processes these complex inputs to generate a precise probability of the home team winning. From this, we also determine:
favorite_prob using thresholds identified from rigorous testing. For example:favorite_prob is 0.65 or higher) has historically shown very high accuracy (around 87.5% in our internal tests).favorite_prob between 0.60 and 0.649) has shown good accuracy (around 67-73% historically, with specific narrower bands like 0.63-0.649 sometimes performing closer to 80%).Example Display: LAD (High - 68.2%) means our Overall Advantage model favors the Los Angeles Dodgers, assigns them a 68.2% “favorite probability” for this matchup, and this falls into our “High” confidence tier.
This Overall Advantage metric aims to give a holistic and nuanced view of team strength and game dynamics. We continuously evaluate its performance to provide the most insightful predictions.