NBA Predictions and Picks

What an NBA prediction gives us

An NBA prediction from us is a moneyline argument with a number at the end. We take the price, strip out the book’s margin, and read what is left as the market’s chance for each team. Our chance sits beside it. When the two are at least 5 points apart, that side is a bet. Closer than that, we pass and explain why in a sentence.

The spread and the total are on the card as well. They come from last season’s game logs: how often each team went over that total, and how often it won or lost by more than that spread. We read them as leans. We never stake them.

The games still to play sit at the top, each card listing our call next to the line, the total and the moneyline. Further down, under “Past Game Predictions Result”, is every past prediction with the call it made. Each prediction carries its writer’s name.

The NBA year

The regular season tips off in late October and runs to mid-April, 82 games a team. The NBA Cup is played inside it in November and December, its group games counting in the standings. After it come the play-in, the postseason rounds, and the Finals in June.

In 2025-26 that came to 1,231 regular-season games, the Cup final among them, and 91 in the play-in and the playoffs. Most nights from late October to April have a full slate.

Who beat the number

Covering is a different contest from winning, and the standings do not tell you who won it. In 2025-26 Charlotte covered 50 times in 84, the best rate in the league. San Antonio covered 62 times in a season that ran all the way to the Finals. At the other end, Cleveland covered 40 times in 100.

Should we read anything into that for the next season? Not much. A cover rate describes one team’s prices over one year. What it does say is that a famous name and a good record do not make a team a good bet. Price the game in front of us, and leave last season’s record where it is.

No NBA spread pushed in 2025-26. Every line was posted on a half point.

Rest, and what the line already knows

Every NBA bettor has heard that rest matters. It does, a little, and the line knows it. When one team had a day more rest than the other in 2025-26, it won 232 of 463 games, almost exactly half, and covered 226. With two or more extra days, it won 56 of 98 and covered 53 of 97.

So is rest a bet? On its own, no. A one-day edge in rest did nothing that the spread had not already allowed for. Two or more days showed a lean, but 97 games is a small sample and one season proves nothing. Rest is a reason to look at a line, not a reason to take it.

Betting the NBA total

NBA totals are big numbers, and in 2025-26 they were close to fair at every size. Games posted under 220 went over 94 times in 185; games at 240 or more, 73 times in 149. Across the regular season the split was 608 overs and 622 unders.

Scoring did move with the calendar. Regular-season games averaged 234.9 points in October and 226.5 in January, the low point, before climbing back to 235.8 in April.

And the playoffs were a different game. Play-in and playoff games averaged 216.4 points against 231.2 in the regular season. A playoff total deserves its own reading, not the one from February.

The playoffs, against the season

What changed in the postseason, apart from the scoring? Less than people think. Home teams won 51 of 91 play-in and playoff games, 56.0%, against 55.5% in the regular season. Favorites covered 44 times in 91, against 607 in 1,230.

On these numbers, a playoff game is priced much like any other. A series is not a reason to bet a side; a gap between the line and our number is.

The long shots

Underdogs won 414 NBA games outright in 2025-26 with a moneyline price on them. Most came at short prices. A few did not. Washington beat Detroit as a 16.5-point underdog at +750, and Dallas beat Cleveland getting 14.5, also at +750.

Those are the games people remember. Charlotte beat Oklahoma City at +675, by 27 points. Long shots land, and they are priced to.

How our NBA number works

Our chance for each team comes from a form rating, built in the style of Elo. It moves on:

• this season’s games, and the margin of each one;
• the opponent;
• recent form;
• a fixed home edge for the home team.

Between seasons, each rating moves a third of the way back to the average. Last spring’s form counts, then fades.

Two rules turn the number into a call. The gap between our chance and the market’s has to reach 5 points. The side also has to clear a floor: if the market gives it under 40% to win, there is no bet, however wide the gap. A long shot can look cheap on paper and still lose most nights, so we stay out. The page goes up anyway and explains the pass. Read the gap as a measure of disagreement with the price.

NBA futures and guides

Futures run all season on their own pages. The big one is the NBA championship; then the awards, from the MVP to the best defender and the top rookie; and the in-season NBA Cup. For the mechanics, our guide to betting on NBA games covers spreads, totals and futures.

NBA betting tips, from the ledger

• Wait for the injury report. Who plays is often confirmed late, and the line moves on it. A bet made before the news is a bet on the news.
• Do not bet rest by itself. In 2025-26 a one-day rest edge won half its games.
• Read playoff totals fresh. Postseason games scored 15 points a game fewer.
• Leave last season’s cover rates behind. They describe last season’s prices.
• Keep to a unit, and keep the record. The price we took, not just the side.

Every team, every game

All 30 teams have pages, linked here, each with its schedule and its predictions. Pick either team to reach its next game.

Questions and answers