All articles
September 29, 2026 · LineBearing Team

How to Read College Basketball Projections

Understand LineBearing Hoops projections: expected possessions, adjusted efficiency, market blending, model edges, the 30-minute lock, and accuracy tracking.

The Hoops Projections page turns team strength, expected pace, venue, and market prices into an estimated score for each Division I matchup. It is a comparison tool—not a promise about the final score.

This guide explains what each number means, how it changes, and how to combine it with Live Signals and Pace & Style.

What each game shows

A projection can include:

  • Projected score for each team.
  • Projected margin, calculated from the difference between those scores.
  • Projected total, calculated from their sum.
  • Moneyline, spread, or total lean when the model’s estimate differs enough from the consensus market.
  • Edge, the size of that difference—not a guaranteed return.
  • Lock state, which marks the version preserved for grading.

A projected 76–70 score means the model’s center estimate is a six-point margin and 146 total points. It does not mean 76 and 70 are the only likely scores.

The model begins with tempo and efficiency

College basketball teams do not play the same number of possessions. The model first estimates a possession range from both teams’ pace, then estimates how efficiently each offense may score against the opposing defense.

In simplified form:

projected team points = expected possessions × matchup points per possession

The real process also adjusts for opponent quality and context. Raw points per game are not enough because they mix pace, schedule strength, and efficiency.

Rating sources and fallback

LineBearing’s primary college basketball team ratings come from T-Rank through CBBData. Those ratings provide adjusted offense, adjusted defense, tempo, and overall team strength.

If the primary ratings are unavailable, the pipeline falls back rather than breaking the slate. Its final fallback builds in-house ratings from completed ESPN game results with opponent and venue adjustment.

A fallback keeps the page available, but source quality and early-season sample size still affect confidence. The model should be read with wider uncertainty when a new season has little current data.

How the market enters the projection

The model does not operate in isolation. It blends its independent estimate with the consensus market from LineBearing’s four priority books:

  • DraftKings
  • FanDuel
  • BetMGM
  • Caesars

The market receives a 40% blend in the current Hoops model. This keeps one noisy rating or matchup estimate from pulling the projection too far from a price informed by many participants.

That does not make the market “right.” It makes the projection appropriately skeptical of extreme disagreement.

Why a ranking is not a projected spread

A ranking orders teams. A game projection estimates one matchup at one venue under current conditions.

The fifth-ranked team might be only a small road favorite over the fifteenth-ranked team. A highly ranked slow team may also project differently against a pressure defense than against a half-court opponent.

Never convert rank difference directly into points. Venue, pace, matchup, and price still matter.

The 30-minute lock

Each matchup locks 30 minutes before its scheduled tip. The locked version preserves the projection and picks used for later grading.

Before lock, the page can update as games, ratings, and market prices refresh. After lock, an in-game score does not rewrite the pregame forecast.

This matters when reviewing accuracy. Compare the result with the locked projection—not with an earlier screenshot or a price that moved later.

Reading spread and total edges

Suppose LineBearing projects Team A by 5.5 while the consensus spread is Team A −3. The raw model difference is 2.5 points.

That difference is useful, but context still matters:

  • Is a key player questionable?
  • Is the game truly neutral?
  • Are the books broadly aligned?
  • Did the market move after the model update?
  • Is the current season sample stable?

For totals, a projected 148 against a market 144.5 is not automatically an Over. Check whether the possession estimate is robust and whether lineup or venue information could alter efficiency.

Projections, Live Signals, and Pace & Style

Use the three views for different questions:

  1. Projections: What does the model estimate?
  2. Live Signals: How is the priority-book consensus moving?
  3. Pace & Style: What type of game creates that estimate?

Agreement can strengthen a thesis. Disagreement is not a command to abandon it, but it is a reason to investigate or reduce confidence.

Accuracy is more than picking winners

Winner accuracy rewards favorites and can look impressive without showing betting value. A useful review separates:

  • straight-up winner accuracy;
  • spread results at the locked line;
  • total results at the locked line;
  • average score and margin error;
  • calibration of estimated win probabilities;
  • closing-line value.

A model can select many winners and still fail against the spread. It can also lose a short run of picks while producing competitive prices. Evaluate a meaningful sample and use the appropriate metric for the market.

Common mistakes

  • Treating the score as exact. It is the center of a range.
  • Reading rank as spread. Rankings and prices answer different questions.
  • Ignoring the market blend. Part of the consensus is already inside the final projection.
  • Comparing to the wrong line. Grade against the preserved pregame number.
  • Overreacting early. New rosters and small samples increase uncertainty.
  • Double-counting tempo. Pace already affects the projected score.

A simple workflow

Open Projections, identify meaningful differences from consensus, inspect Pace & Style, check Live Signals, confirm venue and player status, and define the price at which the idea stops being attractive.

Then record the decision. The projection helps create a disciplined starting point; price and context determine whether there is a position.

What to read next

For entertainment purposes only. Not betting advice. Markets carry risk — only stake what you can afford to lose.