The Moment They Lead

On 20 September, in R27, Seongnam fell behind at home to Hwaseong after four minutes. They levelled on 11 and went ahead on 16. Then they led for 84 minutes. Across those 84 minutes open-play xG read Seongnam 0.23, Hwaseong 1.46. Final-third entries: 21 to 49. In the fifth minute of stoppage time a cross from the left was headed in. 2-2.

It was not a one-off. This season Seongnam went ahead in 13 matches and conceded while leading in 11 of them. The only match in which they never gave a goal away after taking the lead was the 1-0 against Busan in R18 (in R23 at Ansan they went ahead on 90 minutes and led for only 5). Add up the points dropped from winning positions and you get 14. In the September table Seongnam are 10th on 31 points.

One question, then: why does Seongnam get weaker the moment it goes ahead? The data answers in three layers — how far they retreat once ahead, what they give away from that deeper position, and what kind of team they were before going ahead, at level. Finally, what the change of head coach in between did and did not change.

I unrolled all 25 matches minute by minute — when Seongnam led, when it was level, when they trailed, and the moments the goals went in.

Grey dominates. Seongnam spent 61% of their minutes level, 23% ahead and 17% behind. Here are the 13 matches in which a blue band appears.

In 11 of 13 matches they conceded while ahead. The median time from first going ahead to the first goal conceded was 30 minutes; in 3 matches the lead lasted less than 15. Across the 583 minutes they led, Seongnam’s open-play xG per 90 was 0.45; the opponents’ was 1.11. They conceded 13 goals while ahead, from shots worth 12.6 xG — they conceded what they allowed. This was less bad luck than how much they gave away while leading.

Every team gets pegged back sometimes. The issue is frequency. For all 17 clubs I counted, among matches in which they led at any point, the share in which they conceded while ahead.

At 85%, Seongnam are the highest in the league. Chungbuk Cheongju are next on 79%; the league average is 62% and leaders Suwon Samsung are on 32%. The 14 points dropped from leading positions rank 5th; goals conceded per 90 minutes ahead (2.01) rank 6th.

Nor do they spend unusually long ahead. Seongnam led for 583 minutes this season, slightly below the league average (605). They lead briefly, and concede within that brief time.

I put the same Seongnam side next to itself — level and ahead. Every team drops back when it leads, so the comparison is between the size of Seongnam’s shift and the shifts of all 17 clubs.

The moment Seongnam go ahead, their defensive actions move 10.5 points closer to their own goal on a 0–100 scale. That is nearly twice the league average and the largest of the 17 clubs (next: Gimpo, −7.1). A team whose defensive actions were the 3rd-highest in the league at level drops to 16th when ahead.

The attack disappears with it. Open-play xG per 90 falls from 0.99 at level to 0.45 ahead. For most teams it barely moves (league average change +0.03). The opponents’ xG rises from 0.66 to 1.11 — the biggest rise in the league. xG created after winning the ball back falls to 0.27, the lowest in the league: when they win it back that deep, nobody is far enough forward to go with it.

This is not an average of one or two matches. In 11 of the 12 matches they led for ten minutes or more, their defensive actions while ahead sat deeper than at level.

So what does dropping that deep cost? Opponents restart their attacks close to Seongnam’s goal. Clearances do not travel, and the opponent picks the ball up again.

At level, opponent attacks starting within 30 m of Seongnam’s goal come 8.1 times per 90. Ahead, that rises to 13.1 — the 2nd-most in the league — and those attacks produced 6 goals even without penalties. Opponent corners double from 3.1 to 7.1 per 90; Seongnam’s fouls rise from 8.2 to 10.3 and yellow cards from 0.9 to 1.7. While ahead, Seongnam conceded 16.2 set pieces per 90 — the most in the league — and their 4 penalties conceded is also the league high.

Of the 13 goals conceded while ahead, 7 came from attacks that began within 30 m of Seongnam’s goal. 5 were headers and 2 were penalties. Only 3 came from a long counter starting in the opponent’s half. More than half came from balls won back close to Seongnam’s goal — clearances, second balls after set pieces. The 90+5 equaliser in R27 began with a ball won back in Seongnam’s half that became a cross.

The timing is worth a look too. Of the 13, 4 came from the 80th minute on, while 7 came in the first 15 minutes of the second half. Seongnam’s problem with leads looks less like “stamina at the end” and more like “where they hold” — and the most dangerous period was the quarter of an hour after half-time.

Look only at Seongnam ahead and they seem weak. But the state they spend most time in is level, and at level they play like a top-half side.

At level Seongnam create 0.99 open-play xG per 90 and concede only 0.66 — the 3rd-lowest figure conceded in the league. They have 54.8% of the passes, and their defensive actions sit 3rd-highest. xG after regains is 0.68, 3rd. Behind, they are more aggressive still — open-play xG 1.27 (3rd), average pass position 2nd-highest in the league.

The problem is finishing. At level Seongnam scored 17 goals from 19.7 xG. Over the season goals minus xG is −3.9, 16th in the league. Because they do not score more at level, their leads are short; once ahead they drop deep and it is level again. Their 10 draws (7.1 expected) break down as 4 after falling behind, 3 after leading, and 3 goalless.

After losing 0-2 at Chungnam Asan in R20 on 1 August, Seongnam parted with head coach Kyungjun Jeon on 4 August. They were 12th, on 20 points. R21 was played under a caretaker, and on 11 August general manager Haewoon Kim moved into the head-coach role. The season splits at that line.

In the 18 matches before the change Seongnam were a better side than their results. Non-penalty xG was 1.22 : 0.95 per match; re-take every shot 20,000 times and expected points come to 26.1 — 1.45 per match, the pace of the 8th-placed team in the league. They took 20, 6.1 fewer than expected. A large share of that gap is this piece’s subject: in the 7 matches they led before the change, they dropped 7 points.

The seven matches since run the other way. Points per match rose to 1.57, while performance slipped to 0.98 : 1.31 xG per match; they took 2.9 points more than expected. Pass share fell from 54.3% to 46.6%. And the problem with leads is unchanged: in 5 of the 6 matches they led after the change, they again conceded while ahead, and the open-play xG they allowed while leading went up.

The flow of each match was in the status chart above. Here is the performance of all 25, side by side in non-penalty xG, split into Edge, Even and Outplayed at an xG difference of ±0.3.

In the 9 matches where they had the edge, Seongnam went only 3W 3D 3L. Even matches: 2W 3D 1L from 6. Outplayed: 2W 4D 4L from 10. The 6 matches they had the edge in but did not win (R2 Chungbuk Cheongju 2-2 · R7 Ansan 0-1 · R14 Suwon FC 0-0 · R15 Gimhae 1-1 · R16 Suwon Samsung 0-1 · R26 Gyeongnam 0-1) are the biggest loss of Seongnam’s season.

1–6 are the moment they lead, 7–8 are level Seongnam, 9–10 are the season’s dividing line.

I chose four matches by this series’ rules (typical · signature · result over process · process over result), filtered through this piece’s subject.

③ is this piece in one match. Even the one lead Seongnam held looked no different, while they held it, from the leads they lost. What separated kept from lost was less how they defended than how the opponents finished.

At level Seongnam play like a top-half side; the moment they go ahead they retreat deeper than anyone in the league. From that deeper position opponents restart close to Seongnam’s goal and score from crosses, set pieces and penalties. 11 of 13 leads conceded in, 14 points dropped. A new head coach has not changed that shape.

The remaining fixtures: R28 Chungbuk Cheongju (A, 9 Oct) · R29 Suwon FC (H, 17 Oct) · R30 Suwon Samsung (H, 25 Oct) · R31 Jeonnam (A, 1 Nov) · R32 Daegu (H, 7 Nov) · R33 Yongin (H, 21 Nov) · R34 Gimhae (A, 29 Nov). First up are Chungbuk Cheongju, second in the league for conceding while ahead — two teams that struggle to hold a lead. Three things to watch.

• The first 15 minutes after the break — 7 of the 13 goals conceded while ahead came then. When they start the second half ahead, do they keep the defensive position they hold at level?
• The second ball in front of the box — while ahead, the opponents’ starting point moves inside Seongnam’s last 30 m. Who wins the cleared ball, and is the flank the crosses come from closed?
• Attacking while ahead — open-play xG falls to 0.45 per 90 once ahead. An attack that still threatens a second goal is what stops the opponents from committing everyone forward.