Searching for SGA stats usually ends the same way: a season average, a last-5 line, and no idea what to do with it. Season averages are the worst input for a player prop. They blend blowouts, back-to-backs, minutes changes and matchup swings into one number that describes the past and predicts almost nothing. This guide shows how to turn a Shai Gilgeous-Alexander stat page (or any NBA star’s page) into something you can actually price against a sportsbook line.
Rule one: a stat page tells you what happened. A prop asks what happens next. Those are different questions, and the gap between them is where most bettors lose money.
What SGA stats actually tell you about a prop line
A points average of, say, 30 does not mean a 29.5 line is a coin flip. It means the middle of his outcome distribution sits near 30. The over/under depends on the shape of that distribution, not the center of it.
Three things matter more than the average:
- Minutes floor. A guard who plays 34 minutes every night has a much tighter range than one bouncing between 28 and 38.
- Usage. Shot attempts plus free throw trips plus turnovers, divided by team possessions. High usage stars absorb possessions when teammates sit.
- Game script. Blowouts kill fourth quarters. A 20-point lead is the most common reason a star misses a points line by three.
You can pull the raw pieces from official sources. NBA.com/stats publishes per-game, per-36 and advanced splits for free. That is your ground truth. Then you add the betting layer on top.
We wrote a longer breakdown of the possession side in NBA usage rate explained, and a companion piece on reading a stat page line by line in usage rate vs player stats.
Which splits matter, and which are noise
Not every split earns a place in your model. Some carry signal. Most are trivia dressed up as insight.
Splits worth using
- Minutes with and without the second star. Injuries reshape usage more than any matchup.
- Pace of opponent. More possessions means more chances at every counting stat.
- Defensive rank versus position. A team that gets shredded by point guards is a real edge, not a narrative.
- Rest. Back-to-backs cut minutes for stars on many rosters.
Splits that mislead
- Career numbers versus one opponent. Ten games spread over six seasons with different rosters is not a sample.
- Home/away for scoring. The effect is small and usually swamped by everything else.
- “Last 10 games” with no context. We covered why that framing traps people in this piece on last-10 trends.
A split is only useful if you can explain the mechanism behind it. If you cannot say why the number should repeat, treat it as noise.
Turn a stat page into a probability
Here is the workflow. It takes about five minutes per player once you have done it a few times.
- Count the hits. Over a defined recent window, how often did the player clear this exact line? That is your raw hit rate.
- Check the sample. Six of ten is a small sample. It is a starting point, not a conclusion.
- Adjust for context. Was the rotation the same? Was he playing through a minutes restriction? Strip out games that will not repeat.
- Convert the odds. Turn the price into implied probability so you are comparing apples to apples.
- Compare. If your adjusted estimate beats the implied probability after vig, you have an edge. If not, pass.
Step four is the one people skip. A -230 price implies roughly a 70% break-even rate before you strip the juice. A 60% hit rate against that price is a losing bet, no matter how good the player looks. Run the number through the implied probability calculator and the devig calculator before you decide anything.
What a real hit-rate table looks like
Averages hide the thing you care about: how often a line actually cleared. Below is a real slice of StatsBench cheatsheet data from a September 2026 MLB slate. It is baseball, not basketball, but the structure is exactly what an NBA prop row looks like, and it makes the vig point impossible to miss.
| Player (matchup) | Prop | Hit rate | Edge | Best price |
|---|---|---|---|---|
| Brandon Lowe (PIT vs SF) | Hits+Runs+RBIs o0.5 | 60% (6/10) | 1.7% | -352 |
| Brandon Lowe (PIT vs SF) | Hits o0.5 | 60% (6/10) | 0.6% | -230 |
| Spencer Horwitz (PIT vs SF) | Singles o0.5 | 60% (6/10) | 0.3% | -125 |
| Spencer Horwitz (PIT vs SF) | Total bases o0.5 | 60% (6/10) | 0.3% | -250 |
Source: StatsBench cheatsheet data, September 3, 2026.
Look at the spread. Every row shows the same 60% hit rate. The edges range from 0.3% to 1.7%, and the prices swing from -125 to -352. Same headline number, very different bets.
That is the whole lesson. Hit rate without price is decoration. We unpacked the tradeoff in hit rate vs edge, and what 60% really means over a season in hit rate percentages.
How to handle small samples honestly
Six of ten sounds convincing. It is not, on its own. With ten trials, a true 50% player clears six or more surprisingly often. You need either a bigger window or a mechanism that explains the jump.
Two practical fixes:
- Widen the window to 15 or 20 games, then check whether the rate held or collapsed.
- Look for a role change. A trade, an injury to a teammate, or a new starting lineup gives a small sample real meaning. Nothing else does.
Our full write-up on sample thresholds lives in sample size in betting. The short version: treat anything under 15 games as a hypothesis, not evidence.
Building your own player research routine
You do not need a data science degree. You need a repeatable checklist.
The five-question checklist
- Is the minutes role stable this week?
- Does the opponent play fast or slow?
- How does that defense rank against this position?
- What is the best available price across books?
- Does my estimate beat the de-vigged implied probability?
The fourth question is the cheapest edge in betting. The same prop can be -125 at one book and -140 at another. Over a season, shopping across roughly 60 sportsbooks is real closing line value, and it costs you nothing but a few clicks.
StatsBench does this part for you. The Prop Finder filters props by hit rate, consistency grade, SB Score and defensive matchup, and it shows the best price across books next to each line. Defensive Rankings vs Position answer question three directly.
Where numbers stop and judgment starts
Data narrows the field. It does not make the decision. A stat page cannot tell you a coach is managing minutes ahead of a road trip, or that a player looked off in warmups.
Bet inside a bankroll you would be fine losing. Prop edges are small and slow, and variance is real over any stretch you can actually watch. If a good process loses six in a row, that is normal, not broken. Betting is 21+ in legal US markets and legality varies by state. If it stops being fun, call 1-800-GAMBLER.
Research is a filter, not a crystal ball. Anyone selling certainty is selling something else.
Start with the data, finish with your own read
Pull the raw production from a source you trust. Add minutes, usage, pace and matchup. Convert the odds to a probability. Compare, then decide. That loop works for every NBA scorer, not just the ones with big season averages.
If you want the loop compressed into one screen, open the NBA props preview and see the hit rates, consistency grades and best prices side by side before you place anything.
Frequently Asked Questions
Are season averages good enough for NBA player props?
No. A season average blends blowouts, injuries and rotation changes into one number. For props you need minutes stability, usage and the opponent’s pace, then a hit rate against the exact line you are betting.
What hit rate do I need to beat a -230 prop?
Roughly 70% before you account for the vig. A -230 price implies about a 69.7% break-even rate, so a 60% hit rate loses money at that number even though it sounds strong. Always convert the price to implied probability first.
Is a 6-of-10 hit rate a big enough sample?
Not by itself. Ten games is small enough that random variation produces 6-of-10 often. Widen the window to 15 or 20 games, or find a real role change that explains the jump.
Which NBA stats predict prop outcomes best?
Minutes floor, usage rate and opponent pace do most of the work. Defensive rank versus the player’s position adds matchup context. Home/away splits and career numbers against one team are usually noise.
Does line shopping actually matter for player props?
Yes, and it is the cheapest edge available. The same prop can sit at -125 at one book and -140 at another. Over a full season that price gap is worth more than most handicapping tweaks.