How a game total is dissected

Every card on the board is the result of two numbers being compared: the number the sportsbook is offering, and the number we think the game should land on. The board does not show the full math because the math is not the point — the process is. This page explains how a total is dissected from market opener to final grade.

The market total is the starting point

The sportsbook posts a total first. That number is the market's best estimate of how many points will be scored, and it reflects everything the public already knows: recent form, injuries, weather, public betting action, and sharp money. The market total is the anchor. Every projection we make is measured against it, not in isolation.

We pull the median of the major sportsbook lines so one slow-to-adjust book does not skew the board. The market total on each card is the number you would be betting against if you placed a ticket right now.

The projection is our estimate of fair

The projection is the total we think the game would land on if we ran it thousands of times under the same conditions. It is not a prediction of the final score — it is the expected value of all final scores. A projection of 48.5 means that, over a large sample, the average final score would settle near 48.5.

When the projection sits above the market total, the board leans Over. When it sits below, the board leans Under. The direction is simply which side of the market the model lands on.

What moves the projection

A total is built from two broad ingredients: how much volume the game will produce, and how efficiently that volume will turn into points. Change either one and the projection moves.

• Pace: How many plays the two teams are likely to run. Fast offenses facing each other create more opportunities to score. Slow offenses grind clock and shrink the total.
• Offensive efficiency: How good each offense is at turning those plays into points, adjusted for the defense it is facing. A great offense against a weak defense pushes the total up.
• Defensive efficiency: How good each defense is at preventing points. Two strong defenses facing each other pull the total down.
• Weather: Wind, rain, and extreme cold matter. Wind is the biggest lever; heavy rain adds turnovers and shortens drives; a dome removes all of it.
• Situation: Short rest, cross-country travel, rivalry familiarity, and coaching pace all nudge a total in small ways that add up.
• Injuries: Missing starters, especially quarterback, top pass-catchers, and pass rushers, can move a total significantly. The board updates as injury reports do.

From inputs to projection

The model does not add these factors together with a single formula. It weights them by how predictive each one has been historically, then blends them into a projected score for each team. The two projected scores combine into a single projected total. The key idea is that every factor is compared against the opponent's mirror trait — pace against pace, offense against defense, strength against strength.

That is why the same team can have a high total one week and a low total the next. The inputs change constantly. The model does not have a memory of last week's grade; it rebuilds the total from scratch every time the board refreshes.

The edge is the gap

Edge is the distance between the projection and the market total. If the market is 47.5 and our projection is 51.0, the edge is +3.5 points on the Over. If the market is 55.5 and our projection is 50.0, the edge is -5.5 points on the Under.

A bigger edge does not guarantee a win. It means the market and the model disagree by a lot, and the model is siding with one side. The board turns that disagreement into a verdict.

The verdict tiers

Once the edge is calculated, the card gets a grade. The thresholds are simple and fixed:

• HAMMER (4.0+ points): A rare, large disagreement. These are the highest-conviction plays on the board and usually get corrected by the market quickly.
• PLAY (2.5 – 3.9 points): A solid, actionable edge. These are the standard positions the model recommends considering.
• LEAN (1.0 – 2.4 points): A soft edge. Fine for a small position or a parlay leg, but not enough to stand on its own.
• PASS (< 1.0 point): The market and the model agree. There is no ticket to write.

How to read a card

Each card is a mini-version of the whole process. The market total is on the left. The projection is in the middle. The edge is on the right. The gauge below them shows the two numbers visually: the small vertical line is the market, the colored dot is the projection, and the fill between them is the edge. Orange means the model likes the Over; blue means it likes the Under. PASS cards are dimmed so your eye skips them naturally.

What the board leaves out

The board does not show formulas, regression coefficients, or every intermediate step. Those live on the How It Works page for anyone who wants to audit the model. The board is designed for speed: here is the market, here is the projection, here is the edge, here is the grade. Everything else is context.

The bottom line

A total is not a mystery. It is a market number that the model disagrees with, and the board makes that disagreement visible. Start with the market, build a fair projection from pace and efficiency, adjust for weather and situation, compare the two numbers, and size your bet by the size of the gap. That is what every card on the board is doing.