Sharp Consensus Line: How to Build a Fair Line Yourself

Every other tool hands you a “fair value” and stops there. Take it on faith. We won’t do that. This page teaches the actual technique behind a sharp consensus line — a de-vigged, sharp-anchored fair line for two-way markets — worked step by step on a price you already trust, so the number on our board is never magic to you. What we keep is narrow, and we’ll name it out loud: which lines have earned the right to anchor, and how much each one counts. Give a bettor the method. The anchor is the edge.

What the number is

For each two-way market we publish one fair probability per side. The two sum to 100%, vig removed — our best estimate of the true odds. Every +EV flag and every closing-line-value grade on the site measures a posted price against this one number. Sharp consensus fair odds are the yardstick the whole product hangs on. The technique that produces them is below, and you can run it by hand.

Step 1: take the vig out of every book

A posted line isn’t fair. It carries the book’s margin, which is exactly why the two implied probabilities add up to more than 100%. Convert both prices to implied probabilities , and their overround is the vig:

That’s the crude version. The power method corrects favorite–longshot bias, and Shin’s method models insider-shaded prices — the full walkthrough lives in how to de-vig odds. De-vig each book to its fair first, and track how much the three methods disagree while you’re at it. That gap is a data-quality signal: a line whose de-vig is ambiguous earns less trust.

Step 2: one book is not the truth, and neither is a plain average

The classic move is to crown one sharp book (Pinnacle) as the benchmark. That feed is legally fraught to source, and any single book is one point of failure — stale, shaded, or simply gone tomorrow. But the naive fix, averaging every book, is worse than it looks. Soft books systematically shade the popular side, so a plain average gets dragged toward the recreational crowd. You want a pool that leans on the lines pricing honestly and discounts the rest. That’s the whole premise of market consensus odds done right.

Step 3: anchor on a line you trust

The move that decides everything is what you anchor on. Pick the sharpest price in the market — one set where real money trades both sides and the hold is thin for the right reasons — and let it lead. Treat the rest as context around it. Here’s the trap: low vig alone does not mean sharp. Soft books compress their hold to 2%–4% on exactly the high-volume public games they shade hardest, so weighting purely by thin margins hands the shaded crowd the loudest voice. On this page, take the anchor as given — “here is a line I trust” — and build the fair number around it. (Which live lines have earned that trust, and how much each counts, is the part we’ve worked to learn, and the part we keep.)

Step 4: don’t let copycats vote twice

Plenty of “different” sportsbooks run the same odds engine under the hood and post identical prices. Counted naively, ten copies of one line masquerade as ten independent opinions and manufacture false confidence. Collapse near-identical lines into a single vote before you pool, and before you count how many independent opinions you actually hold.

Step 5: average in log-odds, around the anchor

Combine the lines in log-odds (logit) space, not plain probability. Turn each fair probability into a log-odds value, weight-average those, and convert back:

This logarithmic pool is the statistically correct way to merge probability estimates — a plain average is provably underconfident and drifts to the middle. Averaging one number and mapping it back guarantees the two sides sum to 100% with no vig.

Then guard against a shaded majority: clip genuine outliers around the anchor, the line you trust, not around the crowd’s middle. Center the robustness on the crowd and it “corrects” the sharp line as the outlier — backwards. Center it on the anchor and the shaded pack gets discounted instead. That’s the whole ballgame on a public game, and it’s why the choice in Step 3 does the heavy lifting.

Step 6: a confidence number that can’t lie about certainty

The fair line ships with a trust signal built to resist false precision: how tightly the independent lines agree, and how many independent opinions there really are once copycats are collapsed. When no line in the market is worth anchoring on, the honest move is to say so and cap the confidence — not dress up a recreational average as truth. We publish that read alongside every fair line.

The part we keep, and why

Everything above is the technique, and you can run it. What we don’t hand over is the anchor itself: which live lines have earned the right to lead, and how heavily each counts. That isn’t a fixed list — it’s learned. As line-movement and closing history pile up, a line earns weight by consistently beating the closing number (positive closing line value) and predicting outcomes (lower log-loss), under strict no-lookahead discipline. That weighting sharpens quietly as the data grows. The method is open. The anchor is the edge.

The pool, on one page

De-vig every book to a fair probability, then combine the lines in log-odds space rather than plain probability — the statistically correct way to merge probability estimates. Average one number, convert back, and the two sides sum to 100% with no vig. The weights are where the judgment lives; which lines earn them is the part we keep.

Where the technique comes from

None of the machinery here is ours to invent — it’s standard statistics, and in the spirit of showing our work, here’s the reading. Taking the vig out with a model of insider-shaded prices is Shin’s method (Hyun Song Shin, 1992–1993); the favorite–longshot correction is the power adjustment on proportional de-vigging. Merging the de-vigged lines in log-odds is a logarithmic opinion pool — the geometric-mean-of-probabilities rule that is the statistically sound way to combine probability estimates (a plain average is provably underconfident), surveyed by Genest & Zidek (1986).

• Shin, H. S. (1993). “Measuring the Incidence of Insider Trading in a Market for State-Contingent Claims.” The Economic Journal 103(420), 1141–1153. — the de-vig (Shin’s method).
• Shin, H. S. (1992). “Prices of State Contingent Claims with Insider Traders, and the Favourite-Longshot Bias.” The Economic Journal 102(411), 426–435.
• Genest, C. & Zidek, J. V. (1986). “Combining Probability Distributions: A Critique and an Annotated Bibliography.” Statistical Science 1(1), 114–135. — the logarithmic opinion pool.

Those papers give you the technique. What they can’t give you is which live lines have earned the right to anchor the pool on our board, and how much each one counts — the piece we learn from closing-line value over time, and the piece we keep.

Advanced: the pool as a product, and a number for agreement

Two extras for the reader who wants the full depth. Neither changes the recipe above — one shows what the pool really is, and one turns “how much do the books agree?” into a number you can compute.

First, recall the transform from Step 5: each book’s de-vigged probability becomes a log-odds value , and the pool averages those. Exponentiate the average and addition turns into multiplication. The pooled fair line is a weighted geometric mean of odds ratios, one factor per independent book.

Why it matters: odds ratios are multiplicative objects — a coin flip’s is , a heavy favorite’s is — so splitting differences on the plain probability scale is the wrong arithmetic, and this identity is the cleanest proof that the pool isn’t a probability average. It also makes Step 4 visible: ten copycat feeds are literally the same factor repeated ten times — one opinion with an inflated exponent, false confidence written out in the algebra.

Use the product form whenever you want to hand-check a small pool: for two equally weighted lines it’s just the square root of the product of their odds ratios.

Second, agreement. Step 6 promised a confidence read that can’t lie about certainty; its raw ingredient is the weighted spread of the books’ log-odds around the pooled value :

Measuring spread in log-odds, not probability points, is deliberate: on the logit scale a one-point disagreement at 95% counts for far more than one at 55%, exactly as it should. On the worked example above, the five books’ log-odds run from about to , so — a tight market. The 90%/50% pair gives — wide. That’s the calculation to run whenever you want to compare how settled two markets are before trusting either fair line the same amount.

Every formula here lives on the Formula Sheet for quick reference.