What player prop hit rates say about the next game
DARIMARKET keeps game logs for 1,652 players across six sports. We went back through those logs and asked four questions people argue about: does a 10-for-10 streak keep going, do last-5 trends beat last-10, do players do better at home, and do combo props hold up. Every number below comes from our own data, and every rule is written out so you can check it.
How often does a 10-for-10 streak continue?
On DARIMARKET's fixed line ladder (not sportsbook lines), when a player had cleared their hardest line on a stat in 10 straight games, they cleared it again next game 87.6% of the time (21,603 of 24,665); counting every ladder line, including easier ones nested under it, the rate was 90.1% (36,060 of 40,011, 951 players), vs 81.9% after 9 of 10 and 73.2% after 8 of 10.
Next-game clear rate by last-10 record, every ladder line, all sports
Table
| Last 10 | Cleared next game | Players |
|---|---|---|
| 0 of 10 | 4.2% (10,427/249,639) | 1,412 |
| 1 of 10 | 14.0% (15,796/112,678) | 1,486 |
| 2 of 10 | 22.8% (20,113/88,205) | 1,499 |
| 3 of 10 | 30.8% (23,094/74,927) | 1,500 |
| 4 of 10 | 39.1% (25,627/65,522) | 1,494 |
| 5 of 10 | 47.1% (27,117/57,558) | 1,490 |
| 6 of 10 | 56.2% (29,864/53,180) | 1,480 |
| 7 of 10 | 64.9% (32,283/49,778) | 1,454 |
| 8 of 10 | 73.2% (34,850/47,577) | 1,399 |
| 9 of 10 | 81.9% (36,980/45,180) | 1,247 |
| 10 of 10 | 90.1% (36,060/40,011) | 951 |
- About 1 in 8 hardest-line streaks ended in the very next game (12.4%: 3,062 of 24,665).
- By sport, the every-line 10-for-10 next-game rate ran from 87.5% (NHL, 2,233 line-games) to 91.0% (NBA, 21,656).
- Rates pulled toward the overall average (about 33%): strong records cleared less often than their 10-game rate, weak ones more often.
10 for 10, by sport
| Sport | Cleared next game | Players |
|---|---|---|
| NBA | 91.0% (19,703/21,656) | 247 |
| WNBA | 90.7% (3,934/4,336) | 51 |
| NFL | 88.2% (2,974/3,373) | 136 |
| College football | 89.2% (2,413/2,704) | 156 |
| MLB | 89.0% (5,083/5,709) | 144 |
| NHL | 87.5% (1,953/2,233) | 217 |
How we measured this. For each player, line and game, we counted how many of the player's previous 10 games cleared that line, then checked whether the game itself cleared it. Only earlier games count toward the record. Lines come from a fixed ladder per sport and position (for example NBA points 10+, 12+, 15+ up to 35+), so one game produces several line-games, and a player who cleared 25+ ten straight times also cleared 20+. The "hardest line" number keeps only the top line per stat to strip out that overlap. These are DARIMARKET's ladder lines, not the lines FanDuel or DraftKings posted that day.
Consecutive games and nested lines overlap, so these are not independent trials; the player count is the fairer sense of sample size.
Do last-5 trends beat last-10?
Lines trending up (a better clear rate in the last 5 games than the last 10) cleared the next game 42.1% of the time versus 39.5% for lines trending down (235,124 and 233,938 line-games); matched on the same last-10 record the gap averaged 2.9 points, largest in CFB (+6.7) and NFL (+4.3) and near zero in NHL (+0.7) and MLB (+0.1).
Next-game clear rate, trending up vs down, by sport
Table: every last-10 record
| Last 10 | Trending up | Trending down | Gap |
|---|---|---|---|
| 1 of 10 | 14.9% (8,419/56,498) | 13.1% (7,377/56,180) | +1.8 |
| 2 of 10 | 24.9% (5,250/21,087) | 20.6% (4,220/20,465) | +4.3 |
| 3 of 10 | 32.2% (12,153/37,704) | 29.4% (10,941/37,223) | +2.8 |
| 4 of 10 | 41.9% (7,736/18,452) | 36.7% (6,607/17,999) | +5.2 |
| 5 of 10 | 49.0% (14,158/28,865) | 45.2% (12,959/28,693) | +3.8 |
| 6 of 10 | 58.6% (8,512/14,518) | 53.7% (7,747/14,429) | +4.9 |
| 7 of 10 | 66.0% (16,410/24,878) | 63.7% (15,873/24,900) | +2.3 |
| 8 of 10 | 74.6% (8,136/10,900) | 72.3% (8,017/11,091) | +2.3 |
| 9 of 10 | 82.3% (18,281/22,222) | 81.4% (18,699/22,958) | +0.9 |
- Like-for-like: at 6 of 10, trending-up lines cleared 58.6% (14,518 line-games) and trending-down lines 53.7% (14,429).
- The gap was consistent in this data but small next to the 10-game record itself: going from 5 of 10 to 7 of 10 moves the next-game rate about 18 points; the last-5 trend moves it about 3.
- Football's bigger gap may reflect roles changing mid-season (injuries, depth-chart moves); in baseball and hockey the last-5 trend added almost nothing once the last-10 record was known.
- By sport, the like-for-like difference was CFB +6.7, NFL +4.3, WNBA +3.1, NBA +2.9, NHL +0.7, MLB +0.1 points.
How we measured this. Trending up means the clear rate over the last 5 games was higher than over the last 10 (for example 4 of 5 inside 6 of 10); trending down means lower. Lines where the two rates were equal are left out. Because a hot last 5 usually comes with a better last 10, we compared up and down only at the same last-10 count and averaged those gaps, weighted by how many line-games sit at each count.
Consecutive games and nested lines overlap, so these are not independent trials; the player count is the fairer sense of sample size.
Home vs road: do players clear lines more often at home?
On contested lines (ones a player cleared in 25% to 75% of their logged games), players cleared 48.1% of the time at home and 46.4% on the road (269,381 and 251,724 line-games), a small home-road difference that showed up in all six sports, from +0.7 points (NBA) to +2.7 (CFB).
Clear rate on contested lines, home vs road
Table
| Sport | Home | Road | Players better at home |
|---|---|---|---|
| College football | 49.6% (23,052/46,495) | 46.9% (15,943/34,008) | 175 of 327 |
| MLB | 49.4% (26,356/53,349) | 46.9% (24,278/51,778) | 144 of 258 |
| NFL | 48.1% (16,025/33,347) | 45.8% (14,104/30,764) | 134 of 241 |
| NHL | 44.1% (12,817/29,038) | 42.7% (12,554/29,387) | 175 of 323 |
| WNBA | 46.6% (8,843/18,984) | 45.5% (8,471/18,633) | 48 of 81 |
| NBA | 48.3% (42,563/88,168) | 47.6% (41,476/87,154) | 163 of 309 |
- Player by player it is close to a coin flip: 839 of 1,551 players with 20+ line-games each way cleared more often at home (54.1%), 700 more often on the road.
- College football has the biggest gap, and its schedules put more games at home, often against weaker opponents, so part of that gap is likely the schedule, not the venue.
- The gap is under 3 points in every sport, small next to a player's own record on the line.
How we measured this. Every logged game counts here (no 10-game history needed), split by whether the player's team was at home. Lines a player almost always or almost never cleared are dropped, since they read the same home or away and would bury the difference. Player counts include only players with 20 or more line-games each way.
Do combo props hold up as well as single-stat props?
In NBA and WNBA logs, combo props (points + rebounds + assists, points + rebounds, points + assists) cleared the next game at almost the same rate as points, rebounds or assists lines with the same last-10 record: within 1.3 points at every count from 5 of 10 to 10 of 10 (10 of 10: 92.0% of 12,604 combo line-games vs 90.7% of 10,909 single-stat).
NBA + WNBA next-game clear rate, combo vs single-stat
Table: every last-10 record
| Last 10 | Combo | Single | Gap |
|---|---|---|---|
| 0 of 10 | 2.8% (2,179/78,529) | 2.7% (2,441/90,220) | +0.1 |
| 1 of 10 | 11.2% (2,684/23,905) | 11.7% (3,204/27,374) | -0.5 |
| 2 of 10 | 20.2% (3,127/15,487) | 20.9% (3,893/18,667) | -0.7 |
| 3 of 10 | 28.2% (3,356/11,910) | 28.7% (4,313/15,047) | -0.5 |
| 4 of 10 | 36.6% (3,702/10,114) | 38.4% (4,879/12,719) | -1.8 |
| 5 of 10 | 45.8% (4,064/8,874) | 45.9% (5,232/11,395) | -0.1 |
| 6 of 10 | 54.9% (4,473/8,146) | 54.6% (5,688/10,419) | +0.3 |
| 7 of 10 | 65.2% (5,475/8,399) | 65.0% (6,513/10,016) | +0.2 |
| 8 of 10 | 73.5% (6,408/8,713) | 73.4% (7,505/10,224) | +0.1 |
| 9 of 10 | 83.7% (8,504/10,159) | 82.7% (9,044/10,930) | +1.0 |
| 10 of 10 | 92.0% (11,597/12,604) | 90.7% (9,899/10,909) | +1.3 |
- So a combo line's last-10 record carried about the same weight as a single stat's in this data; the combo didn't fade faster.
- This compares lines by their own 10-game record, not by sportsbook price. FanDuel and DraftKings set combo and single lines in different places, so it says nothing about which side of a posted line is better.
- 383 basketball players; football and baseball combos (rush + rec yards, hits + runs + RBIs) are left out because their samples are smaller.
How we measured this. Same rolling 10-game method as finding 1, limited to NBA and WNBA points + rebounds + assists, points + rebounds and points + assists lines (combo) against points, rebounds and assists lines (single). Comparing at the same last-10 count is what makes the two groups like-for-like.
How we measured all of this
Source. DARIMARKET hit-rate game logs (public.hr_players): ESPN box-score stats synced twice a day, up to the last 40 games per player across this season and last. The logs hold 48,723 player games from Mar 19, 2025 to Oct 28, 2026; findings 1, 2 and 4 score 32,709 of them (games from May 28, 2025 to Oct 28, 2026 with 10 earlier games to look back on).
- Lines: the fixed line ladder DARIMARKET uses for each sport and position group (366 lines in all, for example NBA points 10+, 12+, 15+ ... 35+; NHL shots 1+ to 5+; MLB pitcher earned runs 3 or fewer). The same ladder applies to every player in the group.
- Cleared: the player's stat in that game was at or above the line (or at or below it for 'or fewer' lines like walks allowed).
- Line observation: one player, one line, one game, where the player had at least 10 earlier logged games with that stat. L10 counts clears in the 10 games before it, L5 in the 5 games before it; the game itself is the 'next game'. Only earlier games are used, never the game being scored or anything after it.
- The 10 earlier games can cross into last season. Games where a stat wasn't recorded are skipped for that stat.
- Trending up means the L5 rate is above the L10 rate (for example 4 of 5 inside 6 of 10); trending down means below. Comparisons are made at the same L10 count so a better 10-game record can't explain the gap.
- Home vs road uses every logged game and only contested lines: lines the player cleared in 25% to 75% of their logged games.
- Combo vs single compares NBA and WNBA points + rebounds + assists, points + rebounds and points + assists lines against points, rebounds and assists lines at the same L10 count.
- Observations overlap (rolling windows and nested lines on the same stat), so they are not independent trials. Read the counts as how much data sits behind each rate, and small gaps (under about 1 point) as noise.
- Survivorship: hr_players holds the players DARIMARKET tracks today, so players who dropped out of rotations or the league earlier are missing. Their games would likely pull the streak numbers down a little.
- Game dates are as ESPN stores them; a late game on one night can carry the next day's date.
Sample by sport (findings 1, 2 and 4):
| Sport | Players | Player games | Line-games | First | Last |
|---|---|---|---|---|---|
| College football | 279 | 2,103 | 55,939 | 2025-11-12 | 2026-10-27 |
| MLB | 258 | 7,477 | 131,517 | 2025-05-28 | 2026-10-28 |
| NBA | 300 | 8,214 | 402,486 | 2025-11-12 | 2026-10-28 |
| NFL | 239 | 2,677 | 78,147 | 2025-11-14 | 2026-10-27 |
| NHL | 346 | 10,092 | 111,012 | 2025-10-30 | 2026-10-28 |
| WNBA | 83 | 2,146 | 105,154 | 2025-06-15 | 2026-10-27 |
Reproduce it. The read-only SQL and the script that rounds the counts are in our repo (scripts/research/hit-rate-research.sql (read-only SQL) then scripts/research/build.mjs). The full counts behind every number on this page are at /research/data.json (generated 2026-10-29).
What this is not. These are rates from past games, not a forecast. They don't include sportsbook prices, so they can't tell you whether a FanDuel or DraftKings line is a good bet. Injuries, minutes and matchups change from game to game.
Cite this research
DARIMARKET (Oct 28, 2026). "Original research: what player prop hit rates say about the next game." Data through Oct 27, 2026. https://darimarket.us/researchLink a single finding with its anchor, for example https://darimarket.us/research#streaks. Data: https://darimarket.us/research/data.jsonDARIMARKET is research, not a sportsbook, and nothing here is a guarantee. 21+ and only where legal. If betting stops being fun, call or text 1-800-MY-RESET (1-800-697-3738).