Article 01 · Win probability · Hot hand · Player impact

Momentum, myths and who really matters

Three of basketball's oldest arguments: is a lead ever safe, do shooters get hot, and who actually moves the scoreboard? We tested all three on every NBL play since 2015-16.

I

Win probability & clutch

How safe is a lead?

How often the team in front went on to win, by size of lead and time left (every game since 2015-16)

Game flow

Is it honest?

Held-out seasons: when the model says X%, how often does the home team win?

How early can you call it?

Share of held-out games where the leader at that point went on to win, per the model

The biggest blown leads

Who closes games out?

The biggest shots since 2015

Made shots ranked by how much win probability they swung

Win probability added

Clutch shooting

Last 5 minutes, within 5 points (40+ such shots)

II

Is the hot hand real?

After a streak of makes

Change in make rate after k straight makes vs k straight misses, in percentage points

Why the raw numbers mislead

Bias-corrected change in make rate after k straight makes, split into the part explained by the type of shot taken and what is left over

Individual shooters

What changes after a make

Change in the share of each shot type after 1 and 2 straight makes vs misses (bias-corrected), percentage points

The free-throw test

Two-shot trips, where nothing about the shot changes: does making the first affect the second?

III

Who really moves the needle?

Player impact (RAPM)

Points per 100 possessions added to the team's margin vs an average player, after adjusting for teammates and opponents. Bars show the 95% range from refitting on resampled games.

Impact vs scoring

More than their points

Highest impact rank relative to scoring rank (last 3 seasons)

Points without the impact

Highest scoring rank relative to impact rank

Every rated player

Tap a column to sort

Does it predict?

Does it carry over?

Year-to-year correlation for the same players (800+ possessions in both seasons)

Lineup data quality

Share of possessions where we could confirm exactly who was on court

← All research

Article 02 · Home advantage · 1979 to today

The home edge is disappearing

For forty years, playing at home was worth four to five points a game in the NBL. Since 2020 it has been worth about two. We tested the usual suspects (travel, crowds and the whistle) to find out what changed.

I

Forty-seven seasons of home cooking

Home edge by season

Points a game, after adjusting for the strength of both teams. Shaded: 95% range.

By era

II

Is it the travel?

Extra home edge by visitor's trip

Points, compared with a typical interstate trip (1,000–2,500 km)

Which clubs get the most from home?

Extra points on top of the league's home edge, since 2010-11

III

Is it the crowd?

Crowd size vs a usual crowd

Points, compared with a normal-sized crowd for that home team that season

"Home" games away from home

IV

Follow the whistle

Home minus away, by period

Box-score gaps per game, regular season, home games in the home team's own state

Checking the result

← All research

Article 03 · Shot quality · 2015-16 to today

Good shots are table stakes

The NBL has quietly had its analytics revolution: the midrange jumper has halved and every team now hunts the same efficient shots. So what separates the best teams? Whether the shots go in. And that is the part that doesn't last.

I

What a good shot looks like

Make rate by distance

Actual vs the location model, half-metre bins

Points per shot by zone

All regular-season and finals shots with a location

II

The quiet analytics revolution

Where NBL shots come from

Share of all shots by zone, each season

Threes: NBL vs NBA

Three-point attempts as a share of all field-goal attempts, regular season. NBA figures from Basketball-Reference league averages.

III

Taking good shots vs making them

Every team-season

What carries over?

How consistent each part is: odd vs even games of the same season, and the same club the next season (1 = perfectly consistent, 0 = pure chance)

IV

Is shot-making a skill?

Shot-making, last three seasons

Points per 100 shots above what the same shots would earn for an average player. Estimates pulled toward average by how noisy each player's sample is; bars are 95% ranges.

Every player

Tap a column to sort. Shot quality = expected points per shot from where they shoot.

Checking the model

← All research

Article 04 · Runs, momentum and timeouts

Runs end on their own

Two in three NBL games feature a 10–0 run, and the bench reaches for a timeout to stop it. But runs fade at exactly the same rate as any other stretch of basketball, with or without the timeout.

I

Every game has a run

II

Is there momentum?

What the scoring team does next

Points over the next two minutes for the team that just scored, vs what team strength predicts, by the run it is on. No timeout called. Bars are 95% ranges.

III

Do timeouts stop runs?

After the run: timeout vs play on

Points over the next two minutes for the team that was run on

Every way we checked

Extra points over the next two minutes from calling a timeout, with 95% ranges from resampling games

How coaches use them

Team timeouts per game, and the share of run moments (opponent up 6+ over two minutes, timeout available) where the coach called one

← All research

Article 05 · Rest, travel and the schedule · 1979 to today

The NBL killed the back-to-back

Playing on consecutive days used to cost a team about two points. The league has since scheduled them out of existence, and swapped them for four times as much travel, which barely matters.

I

The cost of a back-to-back

Rest before the game

Tired vs rested

From the more tired team's side (home team when equal)

Back-to-back cost, every way we checked

II

A different schedule

Back-to-backs and travel, every season

Share of team-games played the day after another game (bright orange) or with one day between (dark), and average distance travelled since the previous game

III

Does travel matter?

Travel and time zones

Points vs the Elo prediction, compared with no travel since the last game. All seasons (orange) and since 2016-17 (blue).

Who travels most

← All research

Article 06 · Ageing curves and imports

Imports shine, locals decide

NBL players peak at 26. Imports produce more per minute than anyone, but two in three are gone within a year, and it's the quality of the local players that tracks the ladder.

I

When NBL players peak

The NBL ageing curve

Production (Game Score per 40 minutes) relative to a player's peak, and minutes per game relative to peak role. Shaded: 95% range.

II

What imports bring

Imports vs locals

The best import seasons

Game Score per 40 minutes, 2021-22 to 2024-25 (600+ minutes for full weight)

III

Here today, gone tomorrow

Back next season?

Share of players with 200+ minutes who played in the NBL the following season

The Next Stars generation

Imports aged 21 or under with 200+ minutes, by Game Score per 40

The NBL,by the numbers.

Next up

Title race

The game right now

Scoring through the eras

The three-point revolution

Pace

Ask is locked.

Ask the NBL anything.

Teams

Offence vs defence

Franchise history

Team ratings & four factors

Players

Shot charts

Predictions

Season projection

Upcoming games

Results so far vs the model

Is the model honest?

Strongest team-seasons ever

Player projections

The questions fans argue about, tested.

Coming next

Win probability & clutch

How safe is a lead?

Game flow

Is it honest?

How early can you call it?

The biggest blown leads

Who closes games out?

The biggest shots since 2015

Win probability added

Clutch shooting

Is the hot hand real?

After a streak of makes

Why the raw numbers mislead

Individual shooters

What changes after a make

The free-throw test

Who really moves the needle?

Player impact (RAPM)

Impact vs scoring

More than their points

Points without the impact

Every rated player

Does it predict?

Does it carry over?

Lineup data quality

Forty-seven seasons of home cooking

Home edge by season

By era

Is it the travel?

Extra home edge by visitor's trip

Which clubs get the most from home?

Is it the crowd?

Crowd size vs a usual crowd

"Home" games away from home

Follow the whistle

Home minus away, by period

Checking the result

What a good shot looks like

Make rate by distance

Points per shot by zone

The quiet analytics revolution

Where NBL shots come from

Threes: NBL vs NBA

Taking good shots vs making them

Every team-season

What carries over?

Is shot-making a skill?

Shot-making, last three seasons

Every player

Checking the model

Every game has a run

Is there momentum?

What the scoring team does next

Do timeouts stop runs?

After the run: timeout vs play on

Every way we checked

The NBL,
by the numbers.