Anatomy of a Cupset – Part 5: The Findings

Ben Archer >> Blog , Statistical Analysis >> Anatomy of a Cupset – Part 5: The Findings

This series has covered four Australia Cup ties. Perth Glory went out to Lions FC on penalties. Newcastle Jets lost 3-2 to Preston Lions. Central Coast Mariners drew with Sydney United 58 and lost the shootout 11-10. Adelaide United lost 2-1 to South Melbourne, and then lost in almost identical fashion to Công An Hà Nội twelve days later.

Reading the four back together, what the beaten sides shared was a season profile rather than anything particular to the night. That profile was there in the numbers months before any of them played a cup tie. This part works from the data rather than the video, partly because the Central Coast footage I obtained was corrupted and I can’t legally source it another way, and partly because the tape has already done its work across the earlier parts.

Shots Conceded

I expected the common element to sit in the attacking numbers and it turned out to sit in the defensive stats.

Ten A-League clubs entered this year’s Australia Cup. Auckland and Wellington Phoenix are based in New Zealand and don’t take part, so they sit outside the population sampled below.

Rank the ten by shots faced per game across 2025/26 and the top four are Central Coast on 18.6, Newcastle on 16.2, Adelaide on 16.2 and Perth on 15.3. Those are the four clubs that lost to lower-division opposition. The four that met NPL opposition at full strength and came through are Sydney on 14.2, Western Sydney on 13.8, Melbourne Victory on 11.3 and Brisbane Roar on 11.2. Western Sydney and Brisbane have since gone out, but they lost to Melbourne Victory and Sydney, which is a different sort of result and doesn’t bear on the question.

That’s a complete separation in both directions. Every side that lost to a lower-division club was conceding more than fifteen shots a game across the season, and every side that beat one was conceding fewer than fifteen.

These four were letting opponents shoot at them more often than anyone else in the competition, week after week, against A-League attacks. An NPL side doesn’t need to be good to turn that into a result, it only needs to be organised enough to reach the same volume once.

Finishing Above Expectation

Sort the same ten clubs by goals scored minus expected goals across the season and the top three are Central Coast on plus 4.12, Adelaide on plus 4.11 and Newcastle on plus 3.32. No other side in the field finished above zero, and the closest was Melbourne Victory at minus 0.02.

Three of the four beaten sides were therefore the three biggest finishing overperformers in the competition. That’s the gap between a headline number and the number underneath it. Central Coast are the clearest version, scoring 1.3 a game from an expected goals figure of 1.2, while taking 10.0 shots and facing 18.6. Adelaide had the best shot accuracy in the league at 38.0 per cent on expected goals per shot of 0.106, which is middling. They were taking ordinary chances and converting them at an unusual rate.

Newcastle need a correction here, because I got them wrong earlier in the series. Their attack wasn’t a mirage. They created more than anyone else, with expected goals of 1.9 a game, expected goals per shot of 0.129 and 3.4 big chances a game, all of which led the league, and they finished above that as well. Their problem sits on the other side of halfway, with expected goals against of 1.8 and interceptions of 7.9, both second worst in the competition. The scoring covered that across a season and it didn’t cover it across ninety minutes in Melbourne.

Perth Glory

Perth finished the season at minus 2.29 on goals scored against expected goals, which puts them sixth of the ten and below three of the four clubs that came through. There’s no overperformance in their record, and their season was running behind their underlying numbers rather than ahead of them. They still faced the fourth-highest shot volume in the field and they still lost.

Perth are the control case within the four, and they separate the two numbers. Shots conceded holds for all four of the beaten sides, while the finishing residual holds for three and Perth went out without it. That puts shots conceded as the condition behind the defeat, and the finishing residual as the reason three of these results looked like upsets at the time.

The Big Picture

Macarthur and Melbourne City also lost to lower-division opposition and I’ve excluded both. Macarthur used seven non-regular first-team players in the fifteen they picked against SD Raiders, and Melbourne City used eight against Lions FC. Macarthur sit fifth on shots faced at 14.4, immediately behind the four.

One-route Attack

Each of these sides attacked through one channel. Newcastle were strong through short central combinations and dribbling, and sat at the floor of the league for crossing and long passing. Adelaide were the same shape from the other direction, bottom third for total crosses, accurate crosses and successful long passes. Central Coast were a direct long-ball side with very little short or creative output. Perth were similar, a low-possession side that relied on space in behind.

An NPL side won’t outplay an A-League squad across ninety minutes, but it can compress the pitch and take that channel away for a single match. Put that alongside sixteen shots conceded a game and the tie becomes even.

Some Limitations

The field is ten clubs, so every ranking here is a ranking out of ten. Eight of those ten met lower-division opposition at full strength, which is a small sample. A side can finish several goals above expectation across a season through noise alone, so the residual is a flag rather than a verdict.

Macarthur used seven non-regular first-team players out of fifteen and Melbourne City used eight, and those counts are what the exclusions rest on. Both sat below the four on shots faced, so including them doesn’t change the separation.

All figures come from one provider for the whole league, since Opta, FotMob and A-League Stats don’t agree with each other. Games played run from twenty six to thirty because finals are included and Australia Cup ties are not.

Asian Champions League: The Litmus Test

Newcastle Jets enter the AFC Champions League Elite league stage as A-League Premiers, hosting Shanghai Port, Johor Darul Ta’zim, Kashiwa Reysol and Gamba Osaka, and travelling to Kashima Antlers, Vissel Kobe, Beijing Guoan and Kyoto Sanga. That’s eight matches against eight different opponents with no return fixture to correct a bad night.

Newcastle go into it as the best attacking side in Australia and the second most exposed defensive one. I dare say that they’ll concede more shots per match in the league stage than the 16.2 they conceded domestically, and that they will likely get exposed by teams that like to control the ball. The finishing that covered the gap across twenty eight matches has eight games to show up in, and despite my hopes otherwise, I can’t see them moving beyond the league stage.

Adelaide United dropped into ACL Two after losing the Elite play-off 2-0 to Công An Hà Nội. They’ve been drawn in Group F with BG Pathum United, Lion City Sailors and Tai Po, playing each of them home and away between September and December, with the top two going through to the Round of 16.

Adelaide’s finishing residual of plus 4.11 was the second highest in the league, sitting on expected goals per shot of 0.106. Their 7.5 interceptions a game were the fewest of anyone, they faced 16.2 shots a game, level with Newcastle, and they didn’t score against CAHN.

The group format gives Adelaide a return fixture against every opponent, so a bad night is recoverable in a way the Elite league stage doesn’t allow, and Group F is a gentler draw than eight matches against J1 and Chinese Super League sides. I’d still expect them through to the Round of 16, conceding more than 16.2 shots a game and scoring at or below their expected goals across the six.

If both come through without the pattern showing, then the profile matters less than I’ve argued here, and I’ll write that up rather than quietly dropping the claim.

Statistics from ScoutingStats for the 2025/26 A-League Men season, per game, finals included and Australia Cup ties excluded. Cup results as at 22 August 2026, with Sydney FC and Melbourne Victory the only A-League clubs still in the competition.

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Post

Anatomy of a Cupset – Part 1

The 2026 Australia Cup Round of 32 delivered three results the form guide did not…

Mistakes, Mentality and the Player I’d Sign

The World Cup has reached the knockout stage, which means losing is suddenly everywhere. In…

Mind the Gap

Why character and culture belong in the recruitment model On the London Underground, "mind the…