Promotion Race

Outlook After Round 26: What One Defeat Did to the Season

Eleven days ago this site published three pieces built on a 20,000-run simulation of the rest of the season: the revised predictions, the race for two, and the best and worst cases. Those pieces are frozen as published, and...

Português

Eleven days ago this site published three pieces built on a 20,000-run simulation of the rest of the season: the revised predictions, the race for two, and the best and worst cases. Those pieces are frozen as published, and the calls in them are being graded on the Tracker. This is the first refresh of the underlying model, rerun with every Round 26 result across the division and everything learned since: Park Chang-hwan’s suspension, Frizzo’s month out, Christian’s ban, Hwang Jae-won’s return. It is shorter than the originals because most of what they said still holds. What changed is the number at the top.

The headline numbers, then and now

Sep 7 (after R25)Sep 18 (after R26)
Seoul position2nd, 45 pts4th, 45 pts
Seoul projected final6058
Seoul top two37%21%
Seoul title11%2%
Most likely Seoul finish3rd (43%)4th (34%), 3rd (31%)
Suwon Samsung title60%85%
Favorite for 2ndSuwon FC 43%Suwon FC 47%
Daegu top two7%23%
Pair most likely to go upBoth Suwons 56%Both Suwons 55%

One defeat, at home, to the leaders, moved Seoul’s chance of automatic promotion from better than one in three to about one in five. Three things did it, in order of size.

The result itself. A home draw against Suwon Samsung was the model’s call; a defeat was the 31 percent branch, and in that branch the top-two probability was always going to be around 20 percent. The model was not surprised. It was just on the wrong side of the coin.

Daegu. Daegu were a 7 percent afterthought on September 7 because the model expected them to keep dropping points without Edgar. They then beat Yongin 3-1 with two from Deckers, got Kim Dae-woo and Hwang Jae-won back from suspension, and went second. Their projection rose from 56 to 58, level with Seoul’s, and they now take one of the two places in 23 percent of simulations, almost all of them at Seoul’s expense rather than Suwon FC’s.

Goals scored. Seoul and Suwon FC are level on 45 points, and Suwon FC are above on the first tiebreaker by four goals. Seoul finish level with Suwon FC in about 7 percent of simulations and are above them in one in ten of those. The tiebreak was against Seoul before; now it is against them from the first day of the run-in rather than as a hypothetical.

Seoul’s seven, rerun

RoundDateFixtureWinDrawLossModel sheetSep 7 callExpected points
27Sat Sep 19Daegu (H)41%29%30%LossDraw1.51
28Fri Oct 9Gimpo (A)53%26%21%WinWin1.84
29Sun Oct 18Gimhae (H)82%14%4%WinWin2.59
31Sat Oct 31Yongin (A)50%26%24%LossLoss1.76
32Sat Nov 7Jeonnam (H)65%21%14%WinWin2.15
33Sun Nov 22Chungnam Asan (H)62%25%13%WinWin2.12
34Sun Nov 29Gyeongnam (A)38%33%30%DrawDraw1.46

Two notes on the sheet. The September 7 calls on the Tracker do not change; that line was frozen on publication and will be graded as written, and it has already taken its first miss. The “model sheet” column is a fresh read for this article only, built the same way: expected record first (which rounds to 4-1-2 and 13 points, for 58), each result placed where it is most likely. The model has Seoul as the favorite in every one of the seven and still cannot make the sheet say more than four wins, because that is what a 1.9-point-per-game team does over seven matches.

The Daegu match is now a loss in the sheet where it was a draw eleven days ago. That is not because Seoul got worse; the win probability barely moved. It is because Park Chang-hwan’s suspension took a few points off Seoul, Daegu’s return of two suspended starters added a few to Daegu, and the sheet has to carry two losses somewhere. Yongin, where Seoul are now a 50 percent favorite, keeps the other one for the same reason it did before: it is the away match with the highest loss probability after Gyeongnam, and Gyeongnam is a better place for the draw.

Saturday, in numbers

Round 27 is the single largest swing left in Seoul’s season, larger than the Suwon match was, because it is worth three points to Seoul and minus three to a direct rival for the same place.

Saturday resultChanceSeoul top twoDaegu top twoSuwon FC top two
Seoul win40%35%11%56%
Draw29%16%22%63%
Daegu win30%7%42%54%

Win and Seoul are second going into the break at 48 points with a 35 percent chance of staying there. Draw and the race for second is effectively Daegu against Suwon FC with Seoul as the third wheel. Lose and Seoul’s automatic-promotion chance is a rounding error, with the playoffs the realistic season and third place the realistic seed.

What total is enough now

Seoul final totalPoints from last sevenChance of top two
58137%
591415%
601524%
611641%
621757%
631871%
641983%
662195%

The coin-flip total is still 62, which is now 17 points from seven matches rather than from eight: five wins and two draws, or five wins, a draw and a loss plus a favorable tiebreak. Seoul reach 62 or more in 16 percent of simulations, down from 35. Sixty-one, which was worth 39 percent eleven days ago, is worth 41 now because the field around Seoul is slightly weaker than it was; but Seoul reach 61 or more only 27 percent of the time.

From the other direction: for 60 Seoul points to be enough, Suwon FC need to finish on 59 or fewer (35 percent, up from 23 because of Frizzo) and Daegu on 59 or fewer (63 percent). Both, together, happen about one time in four.

The five seasons, refreshed

SeasonFinal totalFinishSep 7Sep 18
The ceiling66 (win out)1st or 2nd3%1%
Best realistic62 to 642nd, sometimes 3rd14%15%
The median58 to 613rd or 4th, home semifinal likely52%45%
The bad case56 to 574th, sometimes 5th23%20%
The floor55 or fewer4th to 6th, away semifinal possible9%19%

The bands have shifted down by about two points and the floor has doubled. Seoul finish fifth or sixth in 14 percent of simulations now, against 3 percent eleven days ago, because Hwaseong (projected 54) and Busan (52) are closer to Seoul’s 58 than they were to 60, and because a Seoul side that loses to Daegu and then drops points at Gimpo is suddenly within a bad fortnight of both. Finishing outside the top six still did not happen in 20,000 runs.

The playoff picture below Seoul is nearly settled: Hwaseong are in the top six 99 percent of the time and Busan 97, with Chungnam Asan and Gimpo sharing the remaining 4 percent. Whoever finishes third hosts sixth on December 3, which the model now makes Busan (54 percent) or Hwaseong (36).

Who goes up

PairSep 7Sep 18
Suwon Samsung and Suwon FC56%55%
Suwon Samsung and Daegu5%21%
Suwon Samsung and Seoul E-Land25%19%
Suwon FC and Daegu2%2%
Suwon FC and Seoul E-Land11%1%
Anyone else1%2%

Suwon Samsung are in the pair in 97 percent of simulations. The one-in-nine chance eleven days ago that they collapsed and Seoul went up with Suwon FC is gone: it needed them to lose at Mokdong, and they did not. The race is for the seat next to them, and Suwon FC’s grip on it has barely loosened despite Frizzo and despite two straight draws, because their remaining eight are still five at home against clubs in the bottom half.

What would change this again

  • Saturday. Covered above. It is worth about 28 points of top-two probability to Seoul across the win and loss branches.
  • Suwon FC at Jeonnam, also Saturday, 16:30. Without Frizzo, with the manager in the stands, against a side unbeaten in four at home under a new back four. A Suwon FC defeat there and a Seoul win two and a half hours later would put Seoul second by three points with the game in hand still to come; the model’s top-two number for Seoul in that combination is about 45 percent.
  • Baek Ji-woong’s fitness for Saturday, unreported since he went off on 78 minutes.
  • The break. Three weeks without a league match after Saturday. Frizzo comes back during it, Kim Jun-hong comes back to Suwon Samsung after it, and Edgar’s calf is the unknown that decides whether Daegu are a two-striker team or a one-man team in October.

The frozen calls

The Tracker’s September line now reads one miss from one (it called a draw against Suwon Samsung). Its call for Saturday is a draw. This article’s fresh sheet says a loss. Both are honest readings of the same probabilities, one from a week when Seoul had their captain and Daegu did not have two starters, one from now. If Seoul win, both were wrong in the good direction, and the whole table above moves back toward where it stood on September 7.