Start for $1

How we grade a capper

We track 899 public capper accounts on X, Reddit, YouTube and Polymarket. So far we have scanned 52,396 public posts and graded 14,213 picks posted from Oct 1, 2025 to Oct 1, 2026.

1. The post and the price

The moment a pick goes public we record it with its timestamp, at the price the capper stated when it agrees with the market at that moment, otherwise at the market median. Where no book quoted the capper's line we keep the price they stated, if any. Edits and deletions after a game starts do not change what was posted. A pick posted after its game started is not graded. Every graded pick is labeled: Live when we recorded it the moment it went up, Reconstructed when we rebuilt it from the public post afterwards.

2. The result

Picks are graded from official results and box scores. Units are flat: one unit risked per pick, and a win pays the price taken. A pick we could not price still counts in the win-loss record and never in units or ROI.

3. Closing-line value, with a 95% interval

When the game starts we take the closing line on both sides and remove the bookmaker margin, leaving the market's fair probability. The price taken against that fair close is the closing-line value (CLV). We report the average CLV with a 95% interval that resamples whole games, so one big slate cannot carry a record.

4. Rankings and the Sharp label

A capper is ranked in a sport after 20 graded picks there, ordered by the lower end of their CLV interval, then ROI. A capper is marked Sharp only when that whole interval sits above zero. Profit alone never makes anyone a Sharp.

5. When sharps agree

Proven rule, walk-forward: >= 2 distinct cappers on the same game + market + side (any line), each of whom, counting only their picks in that sport on games that started at least 6 hours before this one, already had >= 20 graded picks and a CLV 95% lower bound > 0; no proven capper on the other side. 43 plays qualify (skipped: 1 with qualified cappers on both sides, 1 where different lines graded differently). Comparison tier (in profit or proven, 20+ picks): 180 plays (skipped: 35 with qualified cappers on both sides, 0 where different lines graded differently).

6. What is free and what is for members

Counts, records and every graded pick are free, with a link to the original post. A pick on a game that has not started is for members only, and so is the capper behind it: free pages never list a capper's open picks, and the free board shows how many cappers are on a game and market, never the side, the line, the price, or which capper is on it. Capper pages list only bets whose game has started.

7. Notes on the data

  • All records are reconstructed from public posts: graded picks posted from Oct 1, 2025 to Oct 1, 2026.
  • Consensus is walk-forward: a play counts only when two or more cappers were already proven sharps in that sport when it started (20+ graded picks and a CLV 95% lower bound above zero, on games that started at least 6 hours earlier), took the same side at any line, and no other proven sharp took the other side.
  • Parlay legs (props a capper lists as parlay material, such as a trend list) are graded at their own lines and counted in consensus and on the board, but kept out of every capper's straight-bet record: 144 graded, shown apart.
  • No upcoming games have enough matched picks yet, so the board shows the latest slate with at least 8 groups per tab (Oct 1, 2026).

8. Corrections and removal

Spot a grading error? Write to support@hoodedsharp.com with the pick's link. A capper who wants their profile removed can use the removal form.

9. No guarantees

Past results do not guarantee future results. Nothing here is betting or financial advice. 21+ only.