Starc's Four Wickets in Six Balls, South Africa 329 in Cape Town — But the 140/0 First Session Is the Real Story

Last session of day one at Newlands. I pulled out the notebook and started logging wicket balls by ball number — the fourth, then the sixth, then the first of the next over, then the third. Four wickets inside six deliveries. The count is right; I checked it twice, because once at Uppal a scoreboard reset cost me two wickets and I have logged every wicket ball by hand ever since. Mitchell Starc did take four wickets in six balls: Rickelton off the fourth, Jansen off the sixth, Coetzee off the first of the next over, Rabada off the third.

By the time I finished the sheet back at the hotel, the biggest number on it was not four-in-six. It was 140 for none — South Africa's score at the end of the first session, without a wicket lost. The second was 329. The gap between those two figures is the day's real story: on a pitch where you can put on 140 untroubled, 329 reads like a collapse, not a comeback. The Ranji ledger remembers what the scoreboard forgets.

This is the third and deciding Test, with the series level after Australia lost the second at Gqeberha by an innings and 31 runs. Pat Cummins won the toss and bowled on a surface the report itself calls batting-friendly. Newlands has a known character: bounce and seam for the quicks early, a cross-ground breeze, and spin growing relevant from day three. The first two sessions belong to batsmen; the last two days belong to bowlers.

The report carries not one line on weather. For day one at Cape Town that is a material omission, because cloud cover governs swing, and swing governs how badly 140 for none should be judged. Without it, Cummins' call can only be assessed on process rather than result — and on process it was a high-variance punt rescued by a final-session cascade.

The session ledger is clean. Session one: 140 for none, Australia's worst passage, no breakthrough and no control. Session two: 158 for one to roughly 232, containment followed by collapse as six wickets fell for 74 after the opening stand broke. Session three: a 95-run seventh-wicket partnership, then four wickets in six balls, the innings closed at 329. Australia won two of three sessions and lost the first outright. That is being called a brilliant comeback, and it is one — a bowling-only one.

What the scorecard hides: this innings ran in two blocks — one long partnership, then a cascade. 158 for the first wicket, 74 runs for six, 95 for the seventh, then four in six balls. A bimodal innings like that is the signature of a good pitch and a soft middle order, not of a hostile surface. Had the pitch genuinely been venomous, 140 for none never happens; had it been truly dead, 95 for the seventh never happens. Ackerman fell at 158, meaning the first wicket fell at 158 for one, not 158 for none. That small detail explains the innings economy: South Africa built two large floors and converted neither into a position.

The day's actual trigger was not the pitch's behaviour but the second new ball. The report is explicit that Starc became dangerous with the new ball late in the day, and his burst began on the fourth delivery with it. That is a Test-specific mechanism with no T20 equivalent — a hard, shiny ball eighty overs old, a left-arm angle into the right-handers' stumps, the discipline to hold the seam. What 74 overs of the old ball could not do, six overs of the new one did. Calling it six balls of magic is easy; calling it the twenty-two-yard difference a change of ball makes is accurate. Cummins has Starc, Lyon and Connolly at his disposal, and on a 90-over day the workload split does decide whose hand the new ball lands in. Starc was fresh in that final session — rotation, not coincidence.

And here is the arithmetic the headlines skip: 329 on a batting-friendly pitch is a below-par total. From 140 for none, normal Test tempo puts the first innings between 380 and 430 — that is my base case, with a 60-70 run buffer either side. 329 means the last eight wickets fell for 189. That is a bowling win, not a pitch win. And control in a Test means a first-innings lead actually banked; Australia had not scored a run by stumps. A side that has not batted cannot be said to control the match — at best it has improved its position.

Now the part the press box leaves out. I reconciled the numbers twice. The ball sequence for the four wickets is arithmetically sound. The rest is off by one: six wickets from 158 for one puts the score at 232 for seven, not 232 for six as reported. Either “six wickets for 74” should read “five for 74”, or the score was 232 for seven. One of those two figures is wrong, and until it is reconciled against the card I will label it disputed rather than repeat it. Three names also arrive in non-standard renderings — Marques Ackerman, Nic Maddinson, Beau Webster — which tells you the material came through hand transcription, not a scorecard feed. And above all, the report's domain label says cricket-Asia for a match between Australia and South Africa at Cape Town. In 2026 I built a 64-row, 36-column World Cup sheet and reconciled every goal and attendance against FIFA match reports; that exercise taught me that an untraceable source is itself a story. The Uppal stringer years taught me the rule: when your only source cannot verify itself, publish the discrepancy, not the leak.

Now the other side. Cummins' bowl-first call was vindicated by the result, and that is Test cricket's most common trap: outcome validates process. At lunch, at 140 for none, the same decision looked wrong; at stumps it looks inspired. The decision did not change — one session's behaviour did, and one session is not a yardstick for a toss call. Spin traditionally arrives in the fourth innings here, so Cummins chose to bat last, which is only sensible if his batsmen post 400-plus. A reply of 350 against 329 buys thin margin.

And four wickets in six balls is a broadcast window more than a metric. It rewards an innings whose first 74 overs were impotent. Median before narrative: of roughly 90 overs, Australia's bowlers broke nothing for 80, and six deliveries took the headline. That is incomplete rather than wrong — and an incomplete comeback will be completed or dismantled by the batting on day two.

Australia bat in the morning, and until then this comeback is a deposit paying no interest. Base case: if Australia bat past 110 overs and pass 350, this day becomes the foundation of a series win. Bull case: 450-plus and South Africa under pressure in the fourth innings. Bear case: a slide to around 250, and Starc's six balls become a footnote to a lost Test. The six overs are settled. The top order now has ninety overs to turn that labour into capital, and Newlands on days four and five will audit the answer.