Does the ball swing when the cloud comes over the mountain?

One day at Newlands

This week's series decider was played in the shadow of Table Mountain, and its first day split neatly in two. Australia chose to bowl. In the morning South Africa cruised to 140 for 0 at lunch, while the ABC's live blog reported that "the sun is beating down". Mitchell Starc said afterwards there was "not as much sideways movement and swing as we thought there might be on offer".

Then, in the 74th over, the same blog noted "the tablecloth having slipped off the mountain to blanket the suburbs on its lower slopes". Australia took the second new ball in the 81st. A few minutes later the blog read:

"The lights are on. The clouds are overhead. And the new cherry is swinging around corners."

South Africa lost their last four wickets for two runs. Starc finished with 6 for 72. If you wanted a picture of the old belief, you couldn't draw a better one.

But the day had a third part, and it's the one that gets left out. We looked up the hourly weather for all of it.

Cloud and wind are modelled hourly values from the Open-Meteo historical archive, for a grid point a few kilometres from the ground, with wind at 10 m. They aren't readings from the ground, and a model this coarse can't see a cloud sitting on one mountain. The afternoon's cloud is high cloud in the model; the tablecloth comes from the ABC's report.

The new ball alone didn't do it. A cloudy sky with an old ball didn't obviously do it either. Together, they did. That's a tempting result, but look at what else changed. The second new ball is a different ball, and our own simulator lesson warns that real balls vary a lot, even new ones of the same make. The floodlights had come on. And the low cloud and the new ball arrived only about seven overs apart, so this evening can't tell them apart. One evening is a story, not an experiment.

Then the cloud came back

Four days later, Australia needed 288 to win the series, and the sky looked the part. The model has 93–100% cloud for the first hour of the last day, and Kagiso Rabada had a new ball. He struck twice in his first over, and by the seventh over of the innings Australia were 22 for 3. It looked like Day 1 all over again. But in the fifth over the ABC's live blog noted "very little sideways movement through the air", and two overs later:

"The clouds are not delivering the swing that they helped Mitch Starc bend it like Beckham under lights earlier in the Test."

The wickets came another way. Matt Renshaw feathered one down the leg side. Marnus Labuschagne left one that angled in, and was lbw. Travis Head edged a "good bit of nip away off the scrambled seam". Then Simon Harmer took over, "getting drift and getting turn from straight", and finished with 6 for 16 in 21.1 overs, 14 of them maidens, and the first 10-wicket match of his Test career. Australia were all out for 111. South Africa won by 176 runs, took the series 2-1 and went back to the top of the Test rankings for the first time since January 2016. An epic finish, played out under cloud, and the cloud seems to have had little to do with it.

That was the seventh new ball of the match, which gives us something better than one evening. Here are all seven.

Cloud is modelled, from the same archive, for the hour or two after each new ball was taken, and the last day's figures are the model's first pass. The movement is what the ABC's live blog reported at the time: commentators' impressions, not measurements. We counted each innings' new ball and the second new ball separately.

Cloud and swing don't line up in that table. The ball that swung most did have cloud above it, but the next best came under an empty sky on Day 3, and the last morning, just as grey by the model, gave very little. There's a bowler hiding in there too: Starc had four of the seven, and the only two that clearly swung were his.

That doesn't show cloud does nothing. The last morning brought no mention of floodlights or the tablecloth, and Rabada and Jansen aren't Starc. But it does take away the simple version of the belief. A new ball and a grey sky weren't enough.

History agrees, and that's the problem

The famous examples all point the same way. Bob Massie took 16 wickets on debut at Lord's in 1972, in cloudy conditions. England sent Australia in under overcast skies at Trent Bridge in 2015, and Stuart Broad took 8 for 15 as they fell for 60. Captains plan around it too. Mitchell Johnson wrote of England: "When there was cloud cover they would attack, anticipating that the ball would swing. When the sun shone, they looked to tie us down."

But we remember those days because they fit. As the sociologist Nick Allum pointed out in 2015, we're far more likely to recall the ball swinging under cloud than the same swing on a sunny day, or a grey day when nothing happened. Nobody writes a match report about that.

What the science has tested

• Humidity is out. A 2012 study from Sheffield Hallam and Auckland found no link, and a 2024 review in Cambridge's journal Flow says humidity changes the air's behaviour around the ball by less than 1%. Damp air is actually lighter than dry air.
• Cloud itself barely gets studied. The review leaves it out because it's hard to measure and "does not directly affect the air flow around the ball". Any effect would have to come through something else.
• The leading "something else" is a hot pitch. The idea is that a sun-baked pitch stirs up the air, and turbulent air spoils swing. It's plausible: in wind tunnels, turbulence of about 5% stopped a new ball's conventional swing. But when the review's authors measured two English grounds over four days, the temperature difference passed the threshold only once. They say hotter countries still need testing, and that atmospheric turbulence has "not yet been studied in detail or quantified".

So the honest position is the one in our "Weather: myth and fact" lesson: on cloud cover, no firm evidence either way.

Is it really the wind?

Here's a thought that deserves a proper look: maybe when the cloud comes in, so does the wind, and the wind is doing the work. At Newlands it's more than a thought, because the tablecloth is a product of wind. It forms when a south-easterly is pushed up the mountain's slopes into colder air, and the moisture condenses. "Cloud on the mountain" can be a wind report in disguise.

So we checked. This Test gave a first answer, and it isn't the one we expected.

Modelled averages for 11:00–17:00 local time, from the same archive. Day 5 stops at 15:00, with the match nearly over; its cloud stayed at 87–100% until the last of those hours.

The biggest swing came on the calmest day, when the modelled wind was about the same in the morning and the evening. The south-easterly blew on Days 2 to 4 with mixed results: almost nothing through the air on Day 2, some swing on Day 3, a little on Day 4, the breeziest of them. By the last day it had eased and come round towards the south, and there was very little. Much of the bowling on those days was with an older ball. The wind's direction did shift during Day 1, and these are modelled values, so this doesn't clear it. But wind doesn't explain the contrast on its own.

Is that just this Test? We ran the same comparison over about eight years of match-season daylight hours.

Hours from 11:00 to 18:00 local time, 2018 to 2026. Clear is 0–20% cloud, overcast 80–100%. Our own quick analysis of modelled data, not a published study. It describes the sky over the area, so it says nothing about a tablecloth on the mountain itself; testing that needs observations from the ground.

At these two grounds, cloud and wind barely move together, and they don't even lean the same way. So cloud can't simply be standing in for wind. A correlation of −0.10 is close to no relationship at all.

But wind does move the ball

None of that makes wind unimportant. Swing depends on how the seam meets the air, not the ground. The Cambridge review works through a 4 m/s crosswind (about 14 km/h) on a 120 km/h ball: it shifts the ball about 18 cm sideways, and also swings the oncoming air round by 6.8°, which has the same effect as changing the seam angle. That second effect, the authors say, is likely to matter more.

You can see all of this in the swing simulator in our Academy. Here's the same new-ball outswinger, bowled at 120 km/h, in different air.

Humidity from 20% to 95% moves the ball about 2 cm. A breeze from the off side takes away 14 cm of swing, because it closes the seam's angle to the air from 20° to about 12°, and then pushes the ball about 18 cm the other way. No wonder bowlers care which end they get.

The headwind is the striking one. It raises the speed of the air over the ball to about 140 km/h, past the point where a new ball switches to reverse swing in our simulator, so the same grip now swings the ball the other way.

Or is it our eyes?

Some batters argue for a different explanation altogether: a ball you pick up late only looks as if it moved more. Dim light could do that, and it would explain why a grey day feels so dangerous even if the ball is doing nothing unusual.

We like this idea because it can be tested, and because some of the pieces are real. A 2024 study on cricket ball visibility found many combinations of light, ball and background with little or no contrast, though sightscreens gave high contrast for every ball, and that's where a batter first picks it up. And swing already arrives late: about three-quarters of the movement comes in the second half of the flight. Seeing a ball late on top of that would feel like a lot.

Pink against red is a useful clue. It's sometimes said that red moves more than pink because pink is easier to see, but CricViz's tracking of day-night venues in Australia has both balls averaging 0.7° of swing and 0.7° of seam. Pink swings more for roughly the first 30 overs, then less. Batters don't all agree on which is easier, either: Mohammad Mithun said you can read a red ball's seam early but not a pink one's at twilight.

We found no test of the illusion under cloud. It should be simple to do, because ball-tracking measures the ball, not the batter.

Has the ball stopped reversing since the sandpaper?

Now to the other claim doing the rounds. Speaking on The Middle With Watto this month, former Australia coach Justin Langer said: "The ball hasn't reverse-swung anywhere in the world since that day." That day being the 2018 ball-tampering affair, at this same ground. He went further on the current series: "Including both teams, in the first two Test matches, the ball hasn't reverse swung." His wider point is about who else was doing what back then, and we can't judge that. But the cricket claim we can check.

So as stated, the claim is wrong. We found nothing either way on reverse swing in the second Test at Gqeberha, which was still being played when his comments were reported.

There's a fairer version of it, though. Yahoo's report says reverse swing has been "largely missing in recent years", and Pat Cummins said in 2022 that Australia "didn't really experience it" that summer. If it has become rarer, the sandpaper is only one suspect:

• Pitches and grounds. Hazlewood: "If the wicket is hard, firm, deteriorating, and not much seam movement, you're probably going to get that reverse swing." Cummins blamed "grassy wickets, short games".
• The saliva ban, which came in during COVID-19 and changed how teams look after one side of the ball.
• Closer policing. Umpires still watch for throws into the square, and Vernon Philander has argued that they should allow bowlers more freedom.
• Other weapons. One report argues that the wobble seam has taken over much of the job, and that pitches help the quicks enough already. The last day at Newlands fits that: little through the air, Head out to the "scrambled seam", and Rabada later beating the bat with the "wobbled-seam".

And there's a measurement problem. CricViz notes that ball-tracking can't separate conventional swing from reverse. So "it has gone away" and "it has dropped by half" are claims nobody can back with a simple published number. We looked and couldn't find one.

What would settle all this

The data almost certainly exists. Every Test ball is already tracked. Set each ball's measured swing against its age, its speed, the wind at the ground and a record of the sky, across a few hundred days of cricket. One Test's seven new balls show how far a single match can take you, and how much further the numbers could. Allum suggested this in 2015. Eleven years on, we can't find it done in public. If you hold that data, or you've done this work and we've missed it, we'd love to read it.

Try it yourself

This is why the CricCuts Academy exists: so a young player who hears "it'll swing today, look at those clouds" can ask why, and find out. There's no cloud slider in our simulator, because nobody knows yet what to put in it. Everything else in this article, you can try.

Main sources. Grimshaw, Briggs & Atkins (2024), A review and reassessment of the aerodynamics of cricket ball swing, Flow 4. James, MacDonald & Hart (2012), The effect of atmospheric conditions on the swing of a cricket ball, Procedia Engineering. Mehta (2014), Fluid mechanics of cricket ball swing. Allum (2015), The Conversation. Jones (2025), CricViz. The Johnson quote is as cited in Grimshaw et al. Weather values from Open-Meteo's historical archive, modelled rather than measured at the ground. We read the 2024 visibility study through its published excerpts. Match details are from the live coverage linked above and SA Cricket Mag's match report.

Related reading: how we write a lesson and the Academy is open. More on the CricCuts blog.