CBB + March Madness — angle stacking, pace, line value

College basketball is the largest priced market in American sports and the least carefully priced. More than 350 Division I programs play roughly 30 games each — north of 5,000 games a season, often 100 or more on a single Tuesday in January. No sportsbook models that slate with the care it gives a 16-game NFL Sunday. This session is about where that structural gap lives, and how to stack angles into a read without fooling yourself.

What this session covers

Why CBB lines are soft and which corners of the slate are softest. Pace, which drives totals here far harder than in the NBA. And angle stacking — combining reads without counting the same signal twice, which is the most common way members turn a real 2% edge into an imaginary 10% one.

The structural edge

Pricing capacity is finite and gets allocated to handle, which concentrates at the top of the market. A Duke-North Carolina line in February gets real attention. A Tuesday-night mid-major game with a $500 maximum gets a model with thin inputs and almost no human review. The gap between the sharpest-priced game on a January slate and the softest is far wider than in any pro league, so your edge is mostly a function of where you spend your time.

Two facts keep it that way. Sample size is tiny — a 30-game season rests every team's profile on few observations against uneven competition, and strength-of-schedule adjustment does enormous work that reasonable methods disagree about. And roster turnover is extreme — with the transfer portal and early departures, preseason lines routinely price a team that no longer exists.

Pace, and why it matters more here

Possessions drive every counting stat and the game total directly. NBA tempo is compressed: fastest and slowest are typically separated by less than 10% in possessions per game — call it 104 against 96, about 8%. In Division I, the fastest teams run in the mid-70s and the slowest in the low 60s, roughly a 23% spread, nearly three times the NBA's range.

Two implications. Pace mismatches are common and large, and the game usually lands closer to the faster team's tempo than to the average of the two, because the fast team controls more possession decisions. And because the range is so wide, a modest pace error produces a much larger error in the projected total than the same mistake would in the NBA. Totals are where CBB pricing is loosest, and pace is why.

Efficiency and tempo are also independent — a team can be elite per possession and slow, which is a good team with unremarkable totals. Public perception conflates them constantly.

Tournament specifics

March is a different market from January, and not in the direction most people assume.

Neutral courts. Home court is worth roughly three points in college basketball and disappears entirely in the tournament. A team that looked six points better at home is closer to three better on a neutral floor. Ratings that are not venue-adjusted overstate teams that piled up home wins.

Rest, travel, and unfamiliarity. The bracket compresses games into Thursday-Saturday and Friday-Sunday pairs, so a team can be playing its second game in 40 hours after a cross-country flight. Neither side has film on the other the way conference opponents do, so style-mismatch effects run hotter in the first weekend than anywhere else in the season.

Public money floods in. Tournament handle is the highest of the college season and overwhelmingly recreational. That makes marquee lines sharper, because volume attracts serious money to the other side, and more biased, because recreational weight persists on brand names and popular overs. The useful spots are the games nobody is discussing.

Angle stacking without double-counting

Here is the failure mode, with the arithmetic.

Say two angles each make you 55% on a market priced at -110. Angle A: the opponent is on the second night of a road back-to-back. Angle B: their defensive efficiency has fallen off over five games. Treat them as independent and multiply the odds — 55/45 is 1.222, squared is 1.494 — implying 59.9%, a claimed EV of +14.4%.

Now suppose the two are one signal wearing different clothes: the defensive decline is the travel fatigue. Truth is still 55% and real EV is +5.0%. You found a 5% edge and sized it as if it were three times bigger, which is how a good angle produces a bad bankroll outcome.

The discipline:

Write down the mechanism, not the metric. Two angles are one angle if they trace to the same cause. Fatigue, injury, and motivation each get counted once.
Rank, then discount. Strongest angle at full weight; each additional one gets a fraction of its standalone value. Stacking should move an estimate a point or two, not ten.
Test it. Backtest each angle alone and then the stack together against past seasons. If the stack does not beat the better single angle, they were never independent.
Score on CLV, not results. If the stack is real, your lines beat the close. If it is double-counting, CLV goes flat while confidence stays high — the most expensive combination there is.

What the live session adds

A live pass through a January slate, sorted by how much attention the market has paid each game
A worked pace projection on a tempo mismatch, taken through to a total and a price check
A worked backtest separating two angles that look independent and are not
Open Q&A, heaviest in March on rest, travel, and second-weekend spots

Session status: Replays post to this page. The next live date is announced in Discord and in the Sharp Report.