NBA · Finalupdated 23:52 UTC

Milwaukee Bucks vs Miami Heat

at Miami Heat

No game odds posted yet

MIL 121 – MIA 115

Final

We had no picks in this game; every posted line is still scored below.

Getting a better price than most bookmakers offer is the only thing that made money in our past tests, measured on WNBA bets; NBA has not been measured yet. Player rates are context, not tips.

Explore every player and prop

How to read this page

A pick is a player line, for example Under 13.5 points, that we published before the game together with the price we found for it. We do not take bets.
Odds are shown in the decimal format: 1.93 means you get back 1.93 for every 1 staked, so a higher number pays more. The American format (−108 means risk 108 to win 100; +111 means risk 100 to win 111) appears only in the Data log at the bottom of the page.
'Posted at' is the price we found when we published the pick; 'better' is a higher price for the same line at another bookmaker. Taking the better price is the one thing on this page that made money over many past bets; that was measured on WNBA bets, and NBA has not been measured yet.
'Market puts Under at 48.5%' is the chance implied by all bookmakers' prices together, with their built-in margin removed. Pinnacle is a bookmaker whose prices are widely treated as the most accurate reference, so it is the one we compare against.
A player's rate is how often that player has gone over the line in every game we hold since the 2019-20 season (not only this season), for example over in 41 of 78 games (52.6%), with a likely range such as 32–52%: fewer games means a wider range, and under 10 games we show no rate word at all.
'Past record stronger than the price' means the whole likely range sits above the chance the bookmakers' price implies; 'weaker' means below; 'agrees with the price' means the price sits inside the range, which is the usual case. None of these is a tip.
'Not measurable' is the usual honest answer: too few games to tell a real difference from chance.
'What could go wrong' lists the facts that can change a line: who is out, whether the starting lineup is confirmed, expected playing time (RotoWire's estimate), rest, and how many bookmakers price each line.
After the final, every pick is checked against the official stats and compared with the last price before the game started, the market's most informed price; getting a better price than that is the fairest test of a pick, whether or not it won. A tie means the player landed exactly on the line and the stake is returned; lines ending in .5 cannot tie.
Results are given for every 1 staked: 'would have returned +0.93' means a stake of 1 would have come back with 0.93 in profit. We do not stake anything; the record is what a stake of 1 on each pick would have done. Our record counts a smaller stake when the price advantage was small.
S beside a name marks an expected starter. Times are shown in your device's time zone; our data is recorded in US Eastern Time (ET). Team codes are the league's three-letter abbreviations.
'Your prediction' lets you pick Over or Under on any line before the game starts; we check your predictions the same way and keep your record, and one game of results proves nothing.
Team codes on this page: MIL = Milwaukee Bucks, MIA = Miami Heat.

Saved player research

Picks and prices

Odds of at least 1.70 only · highest odds first · results checked

A line is the number a bookmaker sets, e.g. 13.5 points; Over means more than it, Under fewer. Odds 1.93 pay 1.93 for every 1 staked. 'Posted at' is the price we found when we published the pick; after the game it is compared with the last price before the start.

We had no picks in this game.

Why the better price matters

A higher price for the same line pays more whenever the pick wins; that is arithmetic, not a forecast.

What taking the best price is worth has been measured only on WNBA lines (25 Aug 2026); NBA has not been measured yet.

The players

6 players checked · 4 whose past record points the other way from the bookmakers' price · 0 out · 0 with too few games

Rate: the share of every game we hold for the player since the 2019-20 season (not only this season) in which the player finished above tonight's line. S beside a name marks an expected starter.

Against this opponent, 0 of 4 players have performed clearly differently and 0 show a slight hint · No starting lineup yet; showing every player with a line

Context only A player's rate describes the past; it is not a tip and has not been tested as one. Expected minutes come from RotoWire, a lineup news service.

Sign in to see each player's rate The numbers below (for example 15.5 points) are the lines the bookmakers actually posted.

Milwaukee Bucksplaying MIA

Kel'el Ware · minutes unknown

points 13.5 → 7 actual

Ryan Rollins · minutes unknown

assists 3.5 → 2 actual

Miami Heatplaying MIL

Bam Adebayo · minutes unknown

points 22.5 → 31 actual

Davion Mitchell · minutes unknown

3-pointers 1.5 → 1 actual

Giannis Antetokounmpo · minutes unknown

points + rebounds + assists 48.5 → 61 actual

Pelle Larsson · minutes unknown

points 10.5 → 3 actual

Original forecast · corrected actuals

What did we expect, and what happened?

This reads the saved pre-tip model. Today’s rosters, injuries and model changes do not rewrite it.

Reading the immutable pre-tip record…

What could go wrong

updated 23:52 UTC

Who is out

Out Turner · rest · MIL

May not play Herro · right knee contusion (knee) · MIL

Lineups

Expected starting lineup, not yet confirmed

Playing time

No playing-time estimates yet.

Rest

Both teams had at least two days between their last game and this one.

Bookmakers

The reference bookmaker, Pinnacle, offers 0 of the 135 lines; the rest rely on other bookmakers.

Every pick is priced by at least 2 bookmakers.

When a player is out, that player's lines are removed: in our past results this helped. Guessing which teammates benefit was also tested and lost money, so nothing else on this page is adjusted for it.

What happened

In 2 of 4 cases where a player's past record disagreed with the bookmakers' price, the past record was right in this game. One game proves nothing.

FinalWe had no picks in this game · every posted line is still scored below

6 other players with a posted line

Pelle Larsson · points 10.5 → 3 · 7.5 below the line · played 39.1 min · no final price

Kel'el Ware · points 13.5 → 7 · 6.5 below the line · played 18.8 min · no final price

Ryan Rollins · assists 3.5 → 2 · 1.5 below the line · played 14.3 min · no final price

Bam Adebayo · points 22.5 → 31 · 8.5 above the line · played 42.6 min · no final price

Davion Mitchell · 3-pointers 1.5 → 1 · 0.5 below the line · played 39.4 min · no final price

Giannis Antetokounmpo · points + rebounds + assists 48.5 → 61 · 12.5 above the line · played 43 min · no final price

Each prediction is recorded as if you had staked 1 at the line you saw, and your record is shown here. One game tells you nothing yet.

More context

Head to head

MIA vs MIL · last 38 games · MIL leads · MIA record 18–20 · average score 112.6–114.9 · MIA has lost by 2.3 on average

Team results are background only; they say nothing about any single player's line.

12 Mar · MIL 105 at MIA 112MIA won

24 Feb · MIA 117 at MIL 128MIA lost

What each defense gives up per game

MIL gives up121.2 points · 48.9 rebounds · 27.8 assists · 35.7% from three-point range

MIA gives up120.0 points · 45.7 rebounds · 29.7 assists · 39.5% from three-point range

Defense by position

How much this is worth

Defense-by-position numbers are mostly random from season to season; shown for context only. We tested it: a team's defense-by-position numbers from one half of the season do not predict the other half, so treat them as noise.

MIL gives up 121.2 points a game · 28th of 30 · league average 113.3+7.9 vs average

MIA gives up 120.0 points a game · 26th of 30 · league average 113.3+6.7 vs average

How the game odds moved

No history of the game odds on file.

Data log (technical)

Technical detail for auditing; the plain version is above. Internal ids and formula names print here only.

Line-shopping measurement: project_conditional_minutes_refuted

Defense-by-position measurement: split-half reliability r = -0.033 across 12 cells (7 of 12 negative) against a shuffle null of 0.000; a placebo using the wrong position family reproduced the effect to within 0.12pp · project_dvp_blowout_gate_refuted · 2026-08-26

Our picks, raw · 0

No raw pick rows for this game.

Checked results, raw (basketball-stats shorthand) · 0

No graded rows for this game.

Data fetches · 37

sds games/{id}ok 200 · 23 ms

sds rest/team (home)ok 200 · 37 ms

sds lineups/{id}ok 200 · 38 ms

sds rest/team (away)ok 200 · 39 ms

sds injuries/game/{id}ok 200 · 60 ms

dh lineups/{id}/deltaok 200 · 65 ms

sds games/{id}/boxscoreok 200 · 65 ms

sds analytics opponent_defense (home)ok 200 · 74 ms

sds defense/team (away)ok 200 · 74 ms

sds history (h2h)ok · 75 ms

sds defense/team (home)ok 200 · 94 ms

sds analytics opponent_defense (away)ok 200 · 97 ms

grading picks/nightlyok 200 · 98 ms

sds players/team (away)ok 200 · 61 ms

sds players/team (home)ok 200 · 62 ms

sds schedule/team (meetings)ok · 101 ms

grading clv/picksok 200 · 17 ms

grading verdict/boardok 200 · 125 ms

sds analytics player_vs_teamok 200 · 72 ms

sds analytics player_vs_teamok 200 · 112 ms

sds analytics player_vs_teamok 200 · 113 ms

odds game timelinefailed 503 · 257 ms

sds analytics player_vs_teamok 200 · 132 ms

sds analytics player_vs_teamok 200 · 133 ms

sds analytics player_vs_teamok 200 · 176 ms

sds analytics player_vs_line (points)ok 200 · 218 ms

sds analytics player_vs_line (points)ok 200 · 220 ms

sds analytics player_vs_line (assists)ok 200 · 221 ms

sds analytics player_vs_line (three_pointers_made)ok 200 · 231 ms

sds analytics player_vs_line (points)ok 200 · 235 ms

sds analytics player_vs_line (pra)ok 200 · 238 ms

odds player timeline (pra)failed 503 · 270 ms

odds player timeline (points)failed 503 · 274 ms

odds player timeline (threes)failed 503 · 288 ms

odds player timeline (points)failed 503 · 289 ms

odds player timeline (assists)failed 503 · 294 ms

odds player timeline (points)failed 503 · 394 ms

Classic view →

The older, longer version of this page is still available as the classic view. Nothing there is tracked in our results; this page is.