Quantifying how pressure on the ball wins AFL quarters
A look at what captured AFL data tells us about pressure, winning quarters, and how to find your team's number.
A look at what captured AFL data tells us about pressure, winning quarters, and how to find your team's number.
Every coach knows the line: "Apply pressure on the ball." It's said at every level of football, from Auskick to Saturday seniors. Force errors, rush disposals, contest every possession. Pressure turns defence into offence.
But until recently, pressure was a feeling. A coach watches the contest and knows when their side is on top — or being smothered. The challenge has been putting a number on it.
Powercoach's pressure rating tries to do exactly that. It takes every possession the opposition has and grades what your team made of it — an intercept won high up the ground, a stoppage you trapped them into, a clearance you conceded with nobody within a metre — then averages those grades. A neutral contest sits at 100; forcing the opposition to rush the ball through their own goal is the ceiling at 300.
We just rebuilt the formula from first principles and recalculated every game in the system. Once we had clean numbers, we asked the obvious question: does pressure rating actually predict winning?
The answer is yes — strongly. Here are the three findings worth a coach's time.
Update — August 2026: we rebuilt it, and the scale moved
We changed the pressure formula after this article first went up, on the back of coach feedback that it "didn't match the feel". We did not re-tune it. We rebuilt it, because the problems were in what was being counted:
- Every moment is now graded from the defending side's point of view. The old formula let a team's own forward-line output — set shots it kicked, free kicks it won — nudge its own pressure rating upward. Pressure is what you do to the other side's football, so that is the only thing graded now.
- Pressure that wasn't applied no longer scores as pressure. A clearance conceded with no opponent within a metre of the ball used to count for most of an average contest. It now grades 40, well below the neutral 100 — because nothing was applied.
- Trapping the opposition into a stoppage counts, but counts as what it is. It used to score exactly zero, which was the single biggest hole in the rating. It now grades 140 — above a neutral contest, and a long way below winning the ball back (an intercept under real pressure grades 250). Locking the ball in is worth something. It is not worth as much as taking it off them.
- Early-quarter numbers are steadied with a prior instead of a "still stabilising" disclaimer, so the first few minutes of a quarter no longer throw out a wild 300.
There's a fifth change you'll see rather than read about. The gauge keeps the name Pressure Rating and stays cumulative — it answers "how has this quarter gone?", which is the number you want in a review or a printed report. The chart beside it is now genuinely rolling and is called Live Pressure — it answers "what is happening right now?". Same units, same scale, deliberately different windows: they will disagree mid-quarter, and that disagreement is the information.
The rebuild moved the whole scale down and tightened it. 150 is now the average team — not 100, and not the old 180. Across 66 real captured games (21,493 events, 528 team-quarters, 17,753 graded moments):
| Quarter winners | Quarter losers | Gap | |
|---|---|---|---|
| Old formula | 187.9 | 167.4 | 20.5 |
| New formula | 157.2 | 142.5 | 14.7 |
The gap looks smaller because the whole scale is tighter: half of all quarters now land between 139 and 161, so 15 points is a wide gap on this scale. What matters is whether the two groups separate, and on the fairest test of that — does the better-rated side actually win the quarter? — the rebuild is ahead: 78.3% against 76.3%.
The old 180 line no longer means what it did: the measured 50/50 line is now 150. Every chart, table and figure below has been recomputed on the new formula against those 66 games. Nothing on this page is on the old scale any more.
Finding #1: There's a clear gradient — and there's a swing point
We sorted every team's quarter rating into buckets — each game contributes eight observations (four quarters × both teams) — and checked how often each bucket ended up on the winning side of the quarter.
Chart figures are on the current (August 2026) scale, measured across 498 team-quarters.
The trend is unmistakable. At the bottom of the range, teams win about one quarter in twenty-five. In the 120s and 130s they're winning one in four. The 140s are close to a coin flip that lands the wrong way. From 150 they're favourites, and at 170 or better they're winning four quarters in five.
The 50/50 line — where pressure rating crosses from "more likely to lose the quarter" to "more likely to win it" — is 150. That isn't a rounded guess, and it isn't a coincidence either: 150 is also the average team's rating across the whole dataset, and the average team wins half its quarters. The buckets either side bracket it cleanly (140–149 wins 43.5%, 150–159 wins 58.4%).
Read the number against 150 — not the colour. The gauge's coloured bands do a different job: they sit at 125 and 175 and mark the normal range, which means almost every quarter you will ever look at sits in the amber middle. That's deliberate. A colour that changes on nearly every reading tells you nothing; dropping into green or climbing into red means this quarter left the range ordinary football occupies, and that is worth your attention on its own. The win-rate story lives in the number: past 150 and you're favourite to take the quarter; past 170 and you're winning it four times in five.
The gap between quarter winners and quarter losers averages just under 15 points — winners 157.2, losers 142.5. For coaches who like to know whether a number is "real" or just luck of the sample: the gap holds up to standard statistical tests, comfortably clear of zero. The size of the effect (Cohen's d = 0.91, where 0.2 is small, 0.5 medium, 0.8 large) is large — about as strong a signal as any single match-state stat we've measured against winning.
For a coach reviewing a game, 150 is the most useful single number in this analysis. It tells you whether your team's pressure was meaningful — not just better or worse than another quarter, but actually high enough to start swinging outcomes. The top bucket is the thinnest in the sample (66 observations), so treat "80% above 170" as a direction of travel rather than a precise odds line.
One thing the rebuild deliberately changed: this is not a congestion count. Under the old formula, games full of stoppages read as games full of pressure — the most congested side in our dataset came out as the best pressure team in it. Trapping the ball in still counts, but at 140 against 250 for an intercept won under pressure. Sides that scrap the ball into repeated ball-ups now rate slightly below average, which is what that football actually is.
Finding #2: Pressure predicts winning in every quarter
If you've been around football long enough, you've heard someone call Q3 the "premiership quarter" — the moment when good teams pull away. Our data gives that idea a gentle nod: Q3 is the quarter where pressure separates winners from losers most cleanly. But the more useful finding is how evenly pressure predicts the result across all four quarters, and that the gap between the winning and losing side widens steadily as the game goes on. Wherever you are in the game, the side applying more pressure is the side more likely to win the next break.
The numbers per quarter (current scale, 249 decided team-quarters):
| Quarter winners | Quarter losers | Gap | How clearly does pressure predict the winner? | |
|---|---|---|---|---|
| Q1 | 157.5 | 144.6 | 12.9 | Strongly (d = 0.85) |
| Q2 | 158.4 | 143.6 | 14.9 | Very strongly (d = 0.96) |
| Q3 | 157.1 | 141.7 | 15.4 | Very strongly (d = 0.98) |
| Q4 | 156.0 | 140.1 | 15.9 | Strongly (d = 0.88) |
(That d number is just a way of expressing how cleanly the two groups separate. Higher = the winners and losers don't overlap much. A 0.80 means the two groups have clear separation; a 0.98 means they barely touch. Each quarter here rests on roughly 60 decided quarters, so treat the differences between quarters as indicative rather than settled.)
The striking thing is the consistency. Every quarter tells the same story, and the differences between them are small. Q3 edges ahead on how cleanly the two groups separate — a quiet vindication of the old "premiership quarter" wisdom — but only just. What does move steadily is the size of the gap: winners are about 13 points clear in Q1 and about 16 clear by Q4, and almost all of that comes from losing sides sliding backwards (144.6 down to 140.1) rather than winners lifting. Pressure isn't a stat that switches on at a particular moment in the game; it's a lever that works from the first bounce to the final siren, and the side that stops pulling it late is the side that falls away.
There's a coaching read in that, too. You might expect pressure to matter less late, when scoreboard pressure, fatigue, set plays and clutch shooting all start competing for influence. The data says otherwise: pressure holds its predictive power right to the end, and the raw gap between the two sides is at its widest in Q4. A side that keeps the heat on in the last quarter is still the side most likely to win it.
So if you're looking for one point to make at any break, pressure on the ball travels. It isn't a third-quarter trick — it's the through-line of the whole game.
Finding #3: Pressure isn't destiny
Here's the caveat that keeps it honest. Across the 249 decided quarters in the recalculated set, the team with the higher pressure rating won the quarter 78% of the time. Which means roughly one quarter in five was won by the team applying less pressure. (On the old formula, over exactly the same games, it was 76% — the rebuild made the stat a better guide, not a perfect one.)
Why? Usually, efficient finishing. A team that gets fewer scoring opportunities but converts the ones they get can absolutely beat a team that grinds pressure across the field but kicks 2.6.
So pressure rating isn't a deterministic stat. It's a strong tendency — about as strong as any single match-state stat we've measured against winning quarters. The correlation between a team's pressure rating and their net score in the same quarter sits at 0.49 on a –1 to +1 scale (where 0 means no relationship and 1 means perfect): a clear positive link, even if not a guarantee. More pressure tends to mean more scoring; it doesn't promise it.
What that means in practice: pressure is necessary but not sufficient. A team that masters pressure but can't finish in front of goal will still lose. A team that finishes well but applies no pressure will give the opposition too many chances. Pressure rating is one of the most reliable single levers a coach can pull, but it lives alongside shot accuracy, scoring efficiency, and inside-50 conversion.
What this means for your team
The numbers in this post come from clubs across local Melbourne competitions. The threshold is real for this set of teams in these leagues. Your team's specific number will be different.
That's the case for capturing your own games:
- Find your team's threshold. After a few weeks of data you'll know your team's average rating in the quarters you win versus the quarters you lose. Some teams sit well below the league line, others well above. Knowing which side of your line you're on, in real time, is what turns a feeling into a system.
- Find your strongest quarter. League-wide, pressure predicts winning about evenly across the four quarters — but individual teams have their own shape. For some the gap concentrates in Q1 — they win on getting on top early. For others it's Q4 — they win on closing out. Capturing your own games tells you which type of team you are, and what to defend against from week to week.
- See the leakage. A team consistently leaking pressure rating against a particular opponent often has a structural problem — wing roles too narrow, mids losing the contest, defenders dropping off too early. The number forces the conversation.
The league-wide benchmark is a starting reference: quarter winners average around 157, quarter losers around 143, and the 50/50 line sits at 150. The gauge that carries that number keeps the name Pressure Rating and stays cumulative for the period you're looking at. Alongside it sits Live Pressure, a rolling read on the same number: the gauge summarises the period as a whole, while Live Pressure tracks the last stretch of play. They share a scale and they will differ mid-quarter; that's the point of having both. As you capture more games, the benchmark that actually matters is your own.
In short
- There is a threshold where teams start winning quarters more often than they lose them — 150, which is also what an average team rates. Below it they win under a third of their quarters; at 170 and above, four in five.
- Pressure predicts winning in every quarter — the side applying more pressure wins the next break about as reliably in Q1 as in Q4, and the gap between the two sides is widest late. Q3 separates most cleanly, but only just; there's no single "premiership quarter."
- Pressure isn't destiny — about one quarter in five is won by the team applying less pressure, usually through efficient scoring.
- It isn't congestion, either. Locking the ball into repeat stoppages counts, well below winning it back. Scrappy football no longer rates as heavy football.
- The benchmarks worth chasing are your own. Capture, review, adjust.
A note on the data
Every figure on this page was recomputed for the August 2026 rebuild against 66 captured games from local Melbourne AFL competitions (21,493 events, 264 quarter observations, 249 of them decided, 17,753 individually graded moments), after filtering out demo games, unfinished matches, and drawn quarters. Each game contributes eight observations — four quarters × both teams — for 498 team-quarters in the bucket chart. Every candidate version of the formula was measured on that same set before one was chosen.
The pressure rating itself is the average grade of what a team made of the opposition's possessions. Every possession you confront gets one or more grades: forcing the opposition to rush it through their own goal is the ceiling at 300, an intercept won under real closing pressure 250, an intercept mark 220, trapping them into a stoppage 140, a clearance conceded under pressure 120, an uncontested goal conceded 20. A neutral contest sits at 100 and the average team averages 150. Because it's an average, a 30-possession quarter compares fairly against a 20-possession one. Restarts nobody could pressure — kick-ins, free kicks — are left out of the average entirely rather than counted as failures.
Things worth knowing if you're using these numbers as a benchmark:
- The scale changed in August 2026, and it did not just shift. The formula was rebuilt, so old and new numbers are not convertible. If you kept notes from an earlier version of this article, or from the app before the change, recompute rather than adjust — the old 180 line is now 150, and a quarter that read 200 under the old formula does not reliably read anything in particular under the new one.
- The sample is the 66-game corpus the new model was validated on, because that is the only set the new numbers have actually been measured over. Expect them to move a little as the recalculation reaches every game in the system.
- The data is local-league football. The threshold is real for the teams in these competitions. Elite-level football may produce different numbers — typically tighter, with smaller gaps.
- The line isn't magic. It's where the data shifts from "you're more likely to lose the quarter" to "you're more likely to win it." A team a point below it isn't doomed and a point above it isn't guaranteed.
- One thing the rebuild could not fix. Because the rating is an average, conceding a clearance still drags the number down even when you pressured hard. That's inherent to any average, not a quirk of the model — what changed is the price. A clearance conceded under real pressure grades 120: above a neutral contest, below what an average team does.
- Drawn quarters were excluded from the win/loss splits since they don't tell us much about who applied more effective pressure.
- The dataset will keep growing. As more clubs use Powercoach across more competitions, the numbers in this post will be revisited and refined.
