Gimhae FC have taken 12 points from a performance worth 22.1. The −10.1 gap is the largest in the league, and almost all of it was lost at home, while the score was still 0-0. Split the game into phases and you can see exactly how far they get, and where it stops.
Gimhae FC have 2W 8D 16L from 26 matches: 14 points, 20 scored, 50 conceded, 17th in the table. Take the shots from the 24 matches with event data, replay each match 20,000 times, and the expected points come to 22.1. The chance of finishing at or below their actual total is 2.4%.
The shortfall of −10.1 is the largest in the league. This piece asks where those points went. The short answer is two words — at home, and 0-0. The rest of it is phase by phase: what works, and where it breaks.
What this piece covers
In the 11 home matches with event data Gimhae are 0W 2D 9L for 2 points. The same matches were worth 11.2, and the chance of taking 2 or fewer is 0.3%. Away from home: 10 points from 13 matches, expected 10.9 — almost exactly as deserved.
So the shortfall is a home-only phenomenon. That is not the same as saying they played well at home. Per home match they create 1.08 non-penalty xG and allow 1.63, which is a losing profile. It is just not a 2-point one. By points per match they are 0.25 at home and 0.77 away — one of the few sides better away.
Split each match by score state and the picture sharpens. While the score is level Gimhae allow 1.51 xG per 90, the highest in the league (average 0.85; Ansan are second at 1.13). In the same minutes they create 0.57, second lowest.
So they score the next goal from level only 17.2% of the time, against 27.5% implied by shot quality — 10.3 percentage points below. From level they have scored 5 (xG 9.0) and conceded 24 (xGA 23.7). They concede first, and that one sentence is the shape of the season: they have opened the scoring in only 16% of matches.
A side that concedes first rarely leads. Gimhae have held a lead for 206 minutes, the fewest in the league and two thirds of Cheonan's 312. Of the 24 matches analysed they led in 5 (they also scored first in R25 and R28, so the season figure is 7).
In those 5 matches they scored not one goal while ahead (xG 1.3). They lost a lead 3 times, conceding 3 goals from 3.5 xGA, so this is not a story of bad luck while protecting a lead. They have barely had a lead to protect. When they do score first they average 1.75 points a match, against 0.29 when they concede first — the first goal matters unusually much to this team.
They have spent 1101 minutes behind. The attacking side of it is reasonable — 0.86 xG per 90 (mid-table), pass share up to 56%, the defensive line pushed up to 40.3, and final-third passes at 84.4 per 90, nearly double their baseline.
The problem is at the other end. Chasing, they have conceded 20 goals from 15.5 xGA. The +4.5 gap is the largest in the league. Going after an equaliser has repeatedly cost them a second and a third. In other words, the football they play when behind is ordinary; the price they pay for it is not.
The goals arrive at the end. Across the 24 matches analysed, the 15 non-penalty goals conceded after 75 minutes are the most in the league and +4.6 above the 10.4 xGA of those situations. That is 38% of their 40 non-penalty concessions, all in the last fifteen minutes and stoppage time.
In fifteen-minute blocks it is starker: 6 scored and 16 conceded between 76′ and the end; 21 conceded in the first half against 28 in the second. Goal difference per match is −0.56 before the break and −0.64 after.
These numbers cannot separate fitness from substitutions from concentration. They do say where to look first — the next job is a scene-by-scene list of those 16 goals.
Take the journey from back to front in two steps. Of the attacks that start in their own third, 61.1% reach the middle third (15th in the league) and 26.9% reach the final third (16th). League averages are 62.9% and 30.0%.
To the middle third they are close to average; the gap opens on the last step. This is less a side that cannot circulate the ball than a side whose last forward pass keeps getting cut off. Forward passes make up 25.0% of their passing and long balls 15.4%, both at the low end of the league — they are not launching it either.
What the data says — build-up and progression
Strengths and weaknesses of this phase
Of the attacks that reach the final third, 23.6% produce a shot (12th in the league), and each possession there is worth 0.024 xG (12th). They get in rarely, and they are below average once in.
The finishing is worse. Open-play xG runs at 0.69 per 90 (15th) against 0.45 goals (17th), a gap of −0.25 — 17th in the league. Over the season their non-penalty shots were worth 23.3 xG and produced 15 goals, −8.3. Shot locations are not unusual: average distance 17.3m, 61.4% of shots inside the box. With 24 matches it is too early to split luck from finishing, but one thing is clear — this is not a side creating plenty and missing. The chances themselves are scarce.
What the data says — final third and finishing
Strengths and weaknesses of this phase
Attacks that begin with a regain come at 52.1 per 90, 4th in the league. That is not few. The problem is where and how fast. 78.7% of those regains happen in their own third (3rd highest). Only 2.2% of them produce a shot inside ten seconds (16th), and the xG created within fifteen seconds is 0.154 per 90 (14th).
The other direction is the opposite. When Gimhae lose the ball, opponents get a shot away within ten seconds 5.1% of the time — the highest in the league — and create 0.451 xG per 90 within fifteen seconds, again the highest. In short, their transition exchange rate is the worst in the division: slow going forward, fastest and sharpest coming back.
What the data says — transition
Strengths and weaknesses of this phase
Set pieces are worth 0.149 xG per 90 to them (8th), close to average, and have produced 0.112 goals per 90 (13th). Over the season: 3 set-piece goals from 3.98 xG (−1.0) and 5 conceded from 3.58 xGA (+1.4). By type, 0 from corners, 2 from free kicks, 1 from throw-ins.
Penalties are the clearest gap: 1 won against 6 conceded, with the penalties they give away worth 0.175 xG per 90 (2nd in the league). Set pieces and penalties together are worth several points over a season.
What the data says — set pieces and penalties
Strengths and weaknesses of this phase
Read those phases again by score state and Gimhae look like three sides.
In other words, they are at their most aggressive when losing and at their most passive when level. The trouble is that level is where most of the season happens — 46% of all minutes.
They have faced 322 shots worth 38.0 xG, including 106 from the centre of the box and 38 from the six-yard area. Once an opponent reaches the final third, 29.2% of those possessions end in a shot (2nd in the league) and each is worth 0.040 xG — the highest in the league. Arriving is the same as creating against this defence.
Adjust for where the shots on target went and the expected concession is 37.3 against 46 actual: a gap of +3.7, 5th in the league. The adjustment ignores pace, distance and deflection; it answers only "were the shots that came to this part of the goal saved at league-average rate". And that +3.7 is only part of the −10.1 overall shortfall.
What the data says — defence
Strengths and weaknesses of this phase
Against the top six they have 0 points from 9 matches (expected 6.4). That part is explained by the gap in quality.
The part that is not is elsewhere. At home to sides outside the top six they have 2 points from 6 matches, expected 7.9. Away to the same teams they took 10 from 9 (expected 7.8). More points in the harder setting. They draw 28% of their matches and have had 0 goalless draws — so this is not a team grinding out stalemates. They lost or drew home games they could have won.
The scoring list, in goals, reads Je-uk Woo 3 · Seung-jae Lee 3 · Beka 3 · Maisa Paul 2 · Rae-jun Lee 2 · Bruno Costa 1 · Kang-wook Lee 1 · Ho-young Seong 1. No one is in double figures and 8 players have scored. The largest xG total belongs to Je-uk Woo (21 shots · 3.39 xG · 3 goals). Among players with 10 shots or more, only Beka (+0.4) and Rae-jun Lee (+0.9) finished above expectation — the finishing loss is spread across the attack rather than down to one player.
In goal, Pil-soo Choi played 21 matches. He faced 111 shots on target and conceded 36 against a placement-adjusted expectation of 33.5, a loss of +2.5 goals. His save percentage of 67.6% ranks 13th of the 20 keepers with ten or more appearances. It is worth adding that he keeps goal for the side allowing the most non-penalty xG in the league, 1.56 a match.
On shot quality they were ahead 1.49 to 0.46 and lost 0-2. The 2.20 expected points is their season high.
Three conceded in a match where the chances allowed were worth 0.61. This is the placement story at its clearest.
Their best performance of the season (2.15 expected points). The lead went in second-half stoppage time.
They took a point from a match worth 0.08 expected points, having allowed 2.67 xG. Luck runs both ways.
They move it from their own third to the middle third 61.1% of the time against a league average of 62.9%. The break happens on the next step.
Regain-started attacks come at 52.1 per 90, 4th in the league. The issue is not the count but the position (78.7% of regains in their own third).
3 goals conceded while ahead against 3.5 xGA. There is almost no loss in that phase. They have simply been ahead for only 206 minutes.
The placement-adjusted +3.7 is a real loss and 5th in the league, but it is a fraction of the −10.1 shortfall. The rest happens while the game is level.
10 points away against 10.9 expected — exactly as deserved. The shortfall is at home.
Next: (1) a scene-by-scene account of the 16 goals conceded late at home, including substitution timing and set-piece share; (2) a match-by-match review of the six home games against sides outside the top six; (3) whether the xG allowed while level is concentrated in a few matches; (4) video on where final-third entries break down.