Think of a prop screener the way a stock investor thinks of a stock screener. A stock screener doesn't tell you which stocks to buy — it filters the universe of thousands of available securities down to the handful that meet your specific criteria, so you can focus your analysis on candidates most likely to reward further research. A prop screener does the same for NBA player props.
The NBA slate on a given night might include 150 to 300 individual prop markets across 10 to 15 games. Most of those markets are efficiently priced and offer no meaningful edge. A screener filters that universe down to the props that have been hitting consistently — those where a player has cleared a given threshold at a rate meaningfully higher than the book's implied probability. Those are the props worth researching. The rest you ignore.
This guide explains how to use a prop screener effectively: what hit rates mean, which game windows to trust, how to set your filters, and a concrete 10-step daily workflow for turning screener output into actionable decisions. To see the screener in action right now, head to our NBA Prop Screener.
What is a prop screener?
A prop screener is a tool that calculates, for each player on tonight's slate, the percentage of recent games in which their actual stat cleared a given threshold. That percentage is the hit rate. The screener surfaces players where that hit rate is high enough to be interesting — typically above 60% or 70% depending on the game window — and where tonight's sportsbook line sits at or below the threshold.
The core logic is simple: if a player has gone over 20 points in 8 of his last 10 games, and tonight's line is set at 19.5, the market may be underestimating his probability of clearing that threshold tonight. That's the raw signal. Whether it represents a genuine edge depends on several additional factors — the cause of the hit rate, the matchup, the line's position relative to the player's true distribution — but the screener is where you start.
Three properties of a useful hit rate
Not all hit rates are equally meaningful. A useful hit rate has three properties.
It is measured against a realistic threshold. A hit rate measured against a threshold the player clears nearly every game (e.g., points at 8+ for a player averaging 22) tells you nothing. A hit rate measured at a threshold close to the player's median output — where there is genuine uncertainty about which side of the line the game will fall — is informative. The tighter the threshold sits to the player's average, the more meaningful the hit rate.
It has an adequate sample size. Five games is not a meaningful sample for a hit rate. Ten games is borderline. The closer your hit rate window is to the sample size needed for statistical significance, the more you need to discount it. We cover the sample size math in detail in our hit rates and sample size guide. In practice, treat 5-game hit rates as directional signals and 20+ game rates as substantially more robust.
It is calculated against tonight's line, not a fixed threshold. A hit rate that measures "how often has this player gone over 20 points" is directionally useful but not immediately actionable. A hit rate that measures "how often has this player gone over tonight's specific sportsbook line" is directly comparable to the implied probability in the price. When a player's historical hit rate at the current line exceeds the book's implied probability, you have a candidate edge.
Game windows: last 5, 10, and season
The question of which game window to use — last 5 games, last 10 games, or the full season — is one of the most consequential choices you make when using a prop screener. Each window captures a different signal.
Last 5 games is the most recency-sensitive window. It captures the most current form but is highly volatile — a player can go from a 20% hit rate to an 80% hit rate in the last 5 games purely through variance. Use last-5 data as an alert, not a conclusion. When a player shows a very high last-5 rate, investigate why: has something structurally changed, or is he just running hot?
Last 10 games strikes a reasonable balance between recency and sample. With 10 games, the sample is still too small for firm statistical conclusions, but it's large enough to filter out pure one-game flukes. For detecting role changes and emerging trends, last-10 is often the most useful window.
Season to date is the most statistically stable but least sensitive to recent changes. A player who had a role change six weeks ago has a season hit rate that blends his old and new situations. Season-level hit rates are most useful for identifying stable, structural tendencies — a player who consistently underperforms his line on back-to-backs, or consistently outperforms against a specific defensive scheme.
Cross-window consistency is the strongest signal. When a player shows a high hit rate in all three windows — last 5, last 10, and season — the signal is robust. When last-5 is high but season is average, treat it as variance. When season is high and last-5 is low, investigate whether something has changed recently that might explain the recent underperformance.
Threshold selection
The threshold you screen against determines which hit rates you surface. Two approaches are common.
Fixed threshold screening asks: for a given stat (say, 20+ points), which players have the highest hit rate? This is easy to compute and intuitive to read, but it has the flaw that high hit rates on low thresholds are trivially easy to achieve. Filter for thresholds that sit within 20% of the player's recent average to ensure the threshold is meaningful.
Line-relative screening asks: for each player, how often has their actual performance cleared tonight's specific sportsbook line? This is harder to compute but directly relevant to bet sizing. When the hit rate against tonight's line significantly exceeds the implied probability of the price, you have a quantifiable edge to work with.
The Statz Prop Screener uses line-relative screening by default: every hit rate displayed is measured against the actual posted line for tonight's game, not a fixed threshold, making the output immediately comparable to sportsbook implied probabilities.
Ten-step daily workflow
Here is a structured daily routine for using a prop screener effectively.
Step 1: Open the screener 2–3 hours before first tip. Lines are still moving and news is still emerging. Opening too early means the lines may not yet reflect the day's injury news; opening at tip means you have no time to research candidates.
Step 2: Filter for a minimum hit rate of 65%+ across the last 10 games. This is your initial filter. The goal is to reduce the full slate to a manageable list of 10–20 candidates.
Step 3: Sort by hit rate descending within the filtered list. You want the players with the strongest recent tendency to clear their respective lines at the top.
Step 4: For each candidate, check all three windows (last 5, last 10, season). Flag the candidates where all three windows show elevated hit rates — these are your highest-confidence targets.
Step 5: Check the cause. For each high-confidence candidate, identify why the hit rate is elevated. Role change? Strong matchup? Consistent usage? If you can explain it, the hit rate is more trustworthy. If you can't explain it, it may be variance.
Step 6: Cross-reference tonight's matchup. How does tonight's opponent defend this player's primary stat? A high hit rate against weak defensive schedules may not hold against tonight's defence. A defensive matchup filter significantly improves screener output quality.
Step 7: Check injury news. Is the player listed as questionable? Is a key teammate returning that might change usage? Is tonight a back-to-back? Any of these can invalidate a hit rate signal regardless of its strength.
Step 8: Verify the live line. Hit rates can be calculated against a line that existed hours ago. Before betting, confirm the current line and price. If the line has moved significantly against you — say, from 19.5 to 21.5 — the edge may have already been taken by the market.
Step 9: Check implied probability vs hit rate. Convert the current price to an implied probability (1/decimal odds). If the player's 10-game hit rate at tonight's threshold is 70% and the implied probability is 52%, you have a potential edge of 18 percentage points. If they're within 5 points of each other, the edge is marginal.
Step 10: Size the bet. Use fractional Kelly or a fixed unit system based on the size of the edge. Larger, more confident edges get larger stakes. Never stake flat amounts regardless of edge size — that ignores the information content of the hit rate differential.
Filters: team, odds, and hit rate
Three filters materially improve screener output quality beyond the basic hit rate threshold.
Team filter: Screen within tonight's games only. Hit rates across games the player isn't playing in are irrelevant. Some screeners show all players; filter to tonight's slate before drawing conclusions.
Odds filter: Set a maximum odds filter to avoid extremely short-priced overs where the edge, even if the hit rate is high, doesn't cover the vig. Props priced at 1.30 or shorter require an implausibly high hit rate to generate meaningful EV. Focus on markets priced between 1.70 and 2.20 (roughly –125 to +120 American) where both sides are genuinely contested.
Hit rate filter: Set different minimum thresholds by window. For last 5 games, 75%+ to filter out variance. For last 10 games, 65%+. For season, 60%+. These are not hard rules but calibrated starting points that focus attention on the most interesting candidates.
Common mistakes
Treating the screener output as a picks list. A screener surfaces candidates. It doesn't make picks. Every candidate requires the additional research steps described above before you should consider betting. The players at the top of a hit rate sorted list include many variance-driven signals that the research steps will filter out.
Ignoring line movement. A high hit rate calculated at an old line may be irrelevant if the line has moved. Always verify the live line before acting on any screener output.
Over-filtering on sample. If you set your minimum sample requirement too high, you'll filter out every player on a streak that started with a role change — precisely the players where the hit rate is most likely genuine. Last-10 data with 65%+ hit rate and a clear cause is worth investigating even if the last-30 rate is unremarkable.
Not tracking results. A screener's value is only provable over a large sample. Track every bet triggered by the screener, noting which filter settings and which window you used. After 50–100 bets, you'll have data on which combination of filters is most predictive. Without tracking, you're flying blind.
Screener vs other tools
The prop screener is best used as the first step in a multi-tool workflow. It identifies candidates efficiently across a large slate. What it doesn't do: explain why a hit rate exists, incorporate matchup data automatically, or account for line movement between when the hit rate was calculated and when you bet.
The Player Streaks tool gives you streak context — the consecutive nature of recent over-performance, which is a different signal from the hit rate aggregate. The Projections page gives you model-generated point estimates and the implied probability of clearing the line based on those estimates. When a player's high hit rate, an active streak, and a model projection all point the same direction, that convergence is the strongest possible signal the Statz platform can generate.