On performance, Ansan Greeners are the weakest side in the league. Yet scoring first was not rare for them — eleven times in the 25 matches with event data. They won two of those eleven. The seven home matches in which they scored first brought one point. I put all 212 league matches on one scale and opened what happened after Ansan went ahead, phase by phase and by match status.
On 20 September Ansan hosted Chungbuk Cheongju and scored after 12 minutes. This is how it ended — 12′ 1-0 · 33′ 1-1 · 71′ 1-2 · 76′ 1-3 · 90+5′ 2-3.
It was a familiar scene. Ansan scored first in 11 of the 25 matches with event data this season. The results of those 11: 2W 2D 7L (add the R25 1-0 win at Busan, which has no event data, and it is three wins in 12). Narrow it to the 7 home matches in which they scored first: 1D 6L, one point.
In the September table Ansan are 6W 4D 16L, 15th on 22 points. This piece follows what happened after Ansan went ahead.
One number first. On performance, Ansan are the weakest side in the league. Re-take every shot 20,000 times at its xG probability and they are worth 0.71 expected points a match, 17th of 17.
They create 0.79 non-penalty xG a match and allow 1.65. The difference (−0.86) is the worst in the league. 8.4 shots a match is the fewest; 16.8 shots faced is the most.
And yet the results matched the performances. Over the 25 matches with event data: 17.8 expected points, 19 taken, a gap of +1.2. Across the season Ansan are not an unlucky team. The question is where the points leaked out and where they came in.
In K League 2 a team that scores first takes 2.00 points from that match on average. Ansan take 0.73. That is the lowest in the league, and well short of the next team up, Chungbuk Cheongju (1.33).
Look at the eleven one by one and home and away split apart. The 4 away matches in which they scored first went 2W 1D 1L, 7 points. The 7 home matches went 0W 1D 6L — 1 point.
Widen it to every match in which they led at any point: 14. They won 5 and dropped 25 points in the rest, the most in the league (second: Chungbuk Cheongju, 20).
The 14 matches in which Ansan led, one per row. Left: the flow of the match. Middle: the xG each side created while Ansan were ahead.
In 383 minutes ahead, xG was Ansan 1.8, opponents 5.1 — 0.43 against 1.21 per 90. While leading, Ansan took 4.0 shots per 90, the fewest in the league; opponents took 16.0, the most. They conceded 9 in that time from 7.6 opponent xG — less bad luck than simply what they allowed.
The equalisers tell the same story. Of the 9 goals that turned an Ansan lead into a draw, 8 came from open play, and those shots averaged only 0.15 xG. Ansan did not give away big chances; they allowed many low-probability shots until one went in.
The same Ansan, level and then ahead. Every team drops off a little when it leads. Ansan do it to a different degree.
The moment they go ahead, the average height of Ansan's defensive actions drops from 32.7 to 25.7, the lowest of any team when leading. 65% of their defensive actions happen in their own third (first). Their pass share is 33%, the lowest; opponents reach the final third 47.5 times per 90, the most; opponents' set-piece xG, 0.59 per 90, is also the most.
This may be a choice, or it may be being pushed back. The data cannot tell them apart. The outcome is clear, though: once ahead, Ansan gave up the ball and the territory at the same time, and opponents turned that time into shots.
Ansan struggle to reach the opponents' half. 28.2 final-third entries, 8.5 box entries and 8.4 shots a match are all the fewest in the league. Only 23.4% of attacks that start in their own half reach the final third, also the lowest.
Yet settled attacks — moves that began with the ball — produced 10 goals from 7.1 xG: +2.9 against expectation, the best in the league. They do not get there often, but they finish what they get. Lima (6) and Machop (5) led the scoring.
Split by match status, settled-attack goals were 4 from 2.2 xG when level, 6 from 4.5 when trailing (best against expectation in the league) and 0 from 0.4 when leading. When ahead there were barely any settled attacks at all — 93 chains.
What the data shows — attack
Ansan defend deeper than anyone. The average height of their defensive actions, 33.9, is 17th; their PPDA, a measure of how little they press, is 2nd. So opponents get into the box 14.6 times a match (the most) and take 16.8 shots (the most). The 1.93 xG they allow a match is also the most.
Against opponents building with the ball, though, they held up. Opponents' settled xG is 0.28 a match, 2nd-lowest in the league. Ansan's goals conceded came another way — attacks after a regain, set pieces, and penalties.
Split by match status, opponents' settled attacks scored 6 from 3.0 xG when level, 2 from 2.8 when Ansan trailed and 1 from 1.3 when they led. From outside the box they conceded 10 from 7.1 xG, 2.9 above expectation, the second-largest gap in the league. Inside the six-yard box it went the other way: 7 from 10.4 xG, 3.4 fewer.
Goalkeeper Seung-bin Lee conceded 1.5 fewer than the placement expectation (10th of 20 keepers in the league). The goals conceded are hard to pin on the goalkeeper.
What the data shows — defensive block
This is where Ansan get hurt most. They lose the ball in their own half 22.6 times a match, the most in the league. Goals conceded within 15 seconds of losing it: 16 this season, also the most.
The other direction is weak. They win the ball centrally 17.1 times a match, the fewest, and create only 0.17 xG a match within 15 seconds of a regain. Not a team that wins it and strikes, but one that loses it and is struck.
Split by match status, this phase shows how the leads fell. When leading, Ansan conceded 7 to opponents' attacks after a regain from 3.8 xG — +3.2 against expectation, 2nd in the league. When trailing it was the reverse: just 7 conceded from 11.3 xG, 4.3 fewer than expected, the luckiest in the league. When level, 10 from 12.1.
What the data shows — transition
Ansan have given away 9 penalties this season, the most in the league. 7 were scored (4 when level, 3 when trailing). That is more than double the league average for penalty goals conceded (3.1).
Set pieces work better at the other end. Non-penalty set-piece chains brought 3 goals from 5.0 xG (+2.0), 4 of them from throw-ins. At the back they conceded 6 from 6.7 xG, about as expected, though no team allows more set-piece xG a match. When leading, opponents' set-piece xG was 0.59 per 90, the most in the league — yet only 1 set-piece goal went in during that time.
What the data shows — set pieces and penalties
I recounted the goals by match status. In 959 minutes behind, Ansan scored 10 from 10.4 xG and conceded just 15 from 20.5 opponent xG. Goal difference while trailing was +5.2 against expectation, second in the league — thanks to conceding fewer than expected.
Level, it went the other way: 25 conceded from 20.0 opponent xG, the most in the league and 5.0 above expectation (including 4 penalties). Leading, they conceded 9 from 7.6 xG — close to expectation, but a lot for so little time ahead.
Put together, Ansan's season total matched expectation because two things cancelled out. When behind, opponents scored less than expected (a gain); when level and ahead, Ansan conceded more than expected (a loss). The gain went into coming back; the loss went into wiping out leads.
Ansan have played 16 matches at home and 10 away. At home: 2W 2D 12L, 8 points. Away: 4W 2D 4L, 14 points.
The performances were, if anything, slightly better at home: 0.76 expected points a match at home against 0.63 away (matches with event data). But at home they took 4.2 points fewer than expected and away 5.3 more. The chance of taking this many or more from those nine away matches was 4.4%.
Most of the difference is the scene we have already seen. Of the 7 defeats after scoring first, 6 came at home. The away wins, by contrast, were mostly hung on to without much of the ball — R7 at Seongnam (1-0, xG 0.53 : 1.62), R19 at Paju (1-0, 0.09 : 1.58). Rather than Ansan being strong away, their away 1-0s held and their home 1-0s did not.
The 25 matches side by side on non-penalty xG, labelled Edge, Even or Outplayed with a ±0.3 threshold.
With the edge in 3 matches Ansan went 1W 0D 2L; even in 2, 1W 0D 1L; outplayed in 20, 3W 4D 13L. Wins from matches in which they were outplayed make up a large share of their points.
1–4 are about what happens after going ahead, 5 and 8 about how the goals came in, 6–7 about Ansan's baseline, 9–10 about the rest.
The series picks its four matches by rule: (1) the match closest to the season average (2) the match most typical of the team (for Ansan, the longest lead that still did not end in a win) (3) the match where the result most exceeded the performance (4) the match where the performance most exceeded the result.
Ahead from the 22nd minute and held on for 62 minutes. Then the 77th, 83rd and 99th — the longer the lead, the longer the collapse.
Thoroughly outplayed, and turned it round anyway. A good share of Ansan's points came from matches like this.
Behind from the 5th minute for 105 minutes. Lost despite the better performance.
Over the season their points are +1.2 against expectation. The performances are the league's weakest, and the table reflects that.
The volume is the league's lowest, but their finishing from settled attacks is the best (+2.9 against expectation). The problem is how rarely they get there.
Seung-bin Lee conceded 1.5 fewer than the placement expectation. They concede a lot because they are shot at a lot.
While ahead they allowed 7.6 xG and conceded 9 — close to expectation. They faced more shots while leading than anyone.
Away performances (0.63 expected points a match) were worse than at home (0.76). The away points came in above what the performances deserved.
Ansan produced the league's weakest performances and took the points those performances deserved. But the way those points came in and went out was uneven. They won two of the eleven matches in which they scored first, and the 25 points they dropped from leading positions are the most in the league. Once ahead, Ansan dropped deeper than anyone and faced more shots than anyone.
Still to play: R28 Suwon Samsung · R29 Yongin · R30 Cheonan · R31 Suwon FC · R32 Gimpo · R33 Jeonnam — all away. Three things to watch.
Limits · The portal's xG model is not published. Expected points treat shots as independent and ignore the order of goals. The portal gives penalties little or no xG. Match status uses live-feed goal times moved onto the event clock. R25 has no event data and is left out of event-based measures. The time spent leading is only 383 minutes, and home and away cover 16 and 9 matches, so samples are small. Whether dropping deep when ahead is a choice or being pushed back cannot be separated with this data. These values were found by scanning many measures (multiple comparisons).