Track record
Updated automatically after each game kicks off.
Two records, kept apart. The process tracker asks whether our picks beat closing line value (CLV) on the pick'em sites. The results tracker asks whether our Prime Overlays actually won. Every week is shown in both, good or bad.
Process tracker
Did our picks beat closing line value (CLV)? Every time Overlay shows a pick with an estimated edge of 1 point or more, we note the line and payout the pick'em site offered. When the game kicks off, we check what that site was offering by then: its closing line. If it had moved the line or cut the payout in the direction the sportsbooks pointed, the pick we showed was a better deal than the close: it beat CLV. Every flagged pick counts, good or bad, and every week is shown.
Loading…
How to read it
- Picks: lines Overlay showed with an estimated edge of 1 percentage point or more, counted once each (a player, stat, line and side), in the week of kickoff.
- Edge when shown: the average estimated edge, in percentage points, the first time each pick appeared.
- Had a close: picks we last saw within 2 hours of kickoff. A pick taken down earlier (a player ruled out, a line pulled) has no closing line, so it counts as a pick and nowhere else.
- Beating closing line value (CLV) on the pick'em site: of the picks with a close, the share where the line and payout we showed were better than the pick'em site's own at kickoff: by then it had raised the line on a Higher pick, lowered it on a Lower one, or cut the payout. Worse means the site's closing line was the better deal; the rest didn't change. PrizePicks and DK Pick6 pay by entry, so for them only a line move counts.
- Prime Overlays beating CLV: the same, counting only Prime Overlays, the picks every sportsbook pricing them agreed on. They are the same picks the results tracker below grades, counted here once they have a close.
This measures our estimates against the market, not wins and losses: picks that hold their edge still lose often, and none of this predicts future results. Picks under a 20% estimated hit rate (hidden on the board by default) aren't counted. PrizePicks and DK Pick6 edges are measured against the default entries: PrizePicks 6-pick Flex; DK Pick6 3-pick until October 29, 2026, and 6-pick since. Tracking runs on every board Overlay publishes. Football's record starts at 3 PM Central on October 1, 2026, when our list of sportsbooks last changed: picks first shown before then, about a day's worth, were priced from a different set of books and aren't counted. Basketball's starts with each league's first board: the WNBA's on October 5, 2026, the NBA's on October 18, 2026.
Results tracker
Did Prime Overlays win? Each morning we check the previous day's Prime Overlays against the players' final stats in the published box scores. Every Prime Overlay counts, won or lost, and every week is shown.
What is a Prime Overlay?
A Prime Overlay is a pick that every sportsbook pricing it agrees on. On the board it is marked with an amber bar over the player's name. A pick earns the mark only when all seven of these are true:
- Two or more sportsbooks price the line, and three or more on a touchdown line, a double-double, a hockey goals line or a sack Over (sack Overs marked before November 1, 2026 needed two). Each has to post the site's number or one within 50% of it; a book further away doesn't count. A pick resting on a single book never qualifies.
- On a touchdown line, one of those books is FanDuel, Hard Rock, Pinnacle or Novig, the books whose touchdown prices have come closest to how often players scored. This rule started on November 2, 2026.
- Each of those books, on its own, puts the pick above what its payout needs. Not just their average: if even one book's price says the pick loses money, there is no mark.
- The estimated edge is 2 points or more, and 3 points or more on lines one play can decide (touchdowns, sacks, interceptions and longest plays) and on double-doubles and hockey goals.
- It isn't a long shot. A pick we estimate hits less than 20% of the time never qualifies.
- It qualified at least 2 hours before its game. Books take prices down in the last hours before kickoff, so a line can't become a Prime Overlay then. This rule started on October 10, 2026.
- Sportsbooks haven't just pulled the player. When several books take a player's props down at once, news usually follows (a lineup change, say) and the books still posted may not have caught up, so the line can't be marked until they return. This rule started on November 2, 2026.
They are the picks with the broadest agreement behind them, not sure things: they are still estimates and they lose often, which is why their results are shown below.
Loading…
Latest graded Prime Overlays
The five most recent Prime Overlays to finish as a win or a loss, newest first: the player, the side and line that carried the mark, the site's price and the chance we estimated when it was first marked, and what the player finished with. It is updated each morning with the previous day's games.
How to read it
- Prime Overlays: lines that carried the mark on the board, counted once each (a player, stat, line and side on one pick'em site), in the week of kickoff. A day's games are added the next morning.
- Won and Lost: the player finished over the line on a Higher pick, or under it on a Lower one, or didn't.
- Skipped: ties (the player landed exactly on the line), and picks the box score can't settle: the player isn't listed, usually because they didn't play, or the stat isn't in it. Skipped picks are in neither Won nor Lost.
- Hit rate: Won as a share of Won plus Lost.
- Hit rate needed: the average break-even of those picks' prices when they were marked. Comparing the two is a check on our estimates, not a measure of profit: picks pay different amounts, and what an entry returns depends on its picks, payouts and rules.
- Hit rate we estimated: the average chance Overlay gave those picks when they were marked. Over many picks, the hit rate should land near it if our estimates are right.
The results record starts with the process tracker: for football at 3 PM Central on October 1, 2026, and for basketball with each league's first board (the WNBA on October 5, the NBA on October 18). On October 21, 2026 we changed how basketball estimates are made, with wider spreads fitted to the sportsbooks' own lines; basketball picks marked before then stay in the record, estimated the earlier way. For football: on October 5, 2026 we raised the bar for a Prime Overlay (it was an edge of 1 point or more for every line) and changed how touchdown lines are estimated. Picks marked before that change count only if they clear the bar as it is now, and touchdown picks marked before it are left out, because their estimates were made the old way. Since October 8, 2026 a touchdown line needs three sportsbooks for the mark; touchdown picks marked with two before then stay in the record.
On October 30, 2026 we added more caution to Overs on some stats and a small boost to Unders on a few, where sportsbook prices have tended to run high on Overs. Picks marked before then stay in the record as they were, estimated the earlier way.
A few weeks of picks is a small sample, and one bad Saturday moves these numbers a long way. Prime Overlays are estimates and lose often; none of this predicts future results. Picks are graded one at a time: an entry combines several picks and pays only on its own rules, so an entry's results differ from these. The same player and stat can count more than once, when two sites post the line or a site moves it. PrizePicks and DK Pick6 picks are measured against the default entries: PrizePicks 6-pick Flex; DK Pick6 3-pick until October 29, 2026, and 6-pick since. NFL, college football, NBA and WNBA; college basketball will be added with its season.