Изменения рейтингов за последние раунды временно удалены. Скоро они будут возвращены. ×

Блог пользователя Not_so_funny

Автор Not_so_funny, 7 часов назад, По-английски

Has anyone encountered situations where published ICPC regional standings metadata appeared difficult to reconcile with the operational compensation procedures announced during the contest?

I am currently trying to better understand how operational contest adjustments are typically reflected in final standings reconstruction and penalty calculations.

This concerns the ICPC Asia Peradeniya Regional 2025/26.

The Context

The official contest duration for the regional was 4 hours (240 minutes).

During the contest, the judging system for Problem A was returning incorrect "Wrong Answer" verdicts on valid solutions. To address this issue, the organizers announced an onsite extra-time compensation procedure for Problem A.

Under this operational procedure, the published "Extra Time" value allocated to a team represented the amount of contest time effectively spent on Problem A before obtaining an accepted solution. This was calculated using the timestamps of a team's submissions for Problem A:

Case 1: If a team made multiple submissions for Problem A, where the second-to-last submission occurred at time T1 and the accepted submission occurred at time T2, the awarded extra time was calculated as T2 − T1.

Case 2: If a team made only a single submission for Problem A, the awarded extra time was equal to the full timestamp of that accepted submission.

For example:

  • If a team submitted a valid solution to Problem A at T=30 (which incorrectly received a "Wrong Answer" verdict due to the checker issue), and their very next attempt at T=50 received an "Accepted" verdict, the team would receive 20 minutes of extra time.
  • If a team's first accepted submission for Problem A occurred at T=45 with no prior submissions to the problem, the team would receive the full 45 minutes as extra time.

These published "Extra Time" values were then incorporated directly into the final penalty and standings calculations.

The Reconstruction Discrepancy

After reviewing the final published standings and associated compensation values, some of the published timestamps and extra-time values appear mathematically difficult to reconcile with the announced operational procedure.

To illustrate the specific structural inconsistency:

  • Base contest duration: 240 minutes
  • Published extra time for a specific team: 28 minutes
  • Effective adjusted contest limit implied by the published metadata: 268 minutes (240 + 28)

However, the certified standings include an accepted submission for the same team timestamped at 272 minutes.

Under standard DOMjudge behavior and the announced operational compensation rules, I have been unable to mathematically reconcile how a submission timestamped at T=272 maps onto an adjusted contest limit of T=268.

Relevant published standings metadata referenced above: Standings Metadata

Why This Matters

Our concern is not limited to a single submission timestamp.

Because the published extra-time values were incorporated into the final penalty calculations, even small reconstruction inconsistencies can materially affect competitive outcomes when standings margins are extremely narrow.

Over the past several months, we have exhausted the standard administrative channels and escalated this inquiry directly to ICPC Judging and Systems leadership. Despite these multiple follow-ups and escalation attempts, we have not yet received a technical explanation reconciling the published metadata and standings calculations.

The Ask

I am not posting this to make accusations against any team or individual organizer.

I am simply trying to understand whether a known technical or administrative explanation exists for this type of reconstruction discrepancy.

In particular, I would be interested in hearing from people familiar with:

  • DOMjudge architecture and contest locking behavior
  • ICPC regional standings reconstruction workflows
  • Post-contest compensation adjustment handling
  • Operational contest administration practices

Is there a standard technical or administrative mechanism in DOMjudge or ICPC operations that can explain how a certified accepted submission timestamp exceeds the adjusted contest limit implied by the published standings metadata?

Official ICPC Global Results: https://icpc.global/regionals/results

Referenced published standings spreadsheet: (Note: To independently verify this source, navigate to the Official ICPC Global Results link above, locate "The 2025 ICPC Asia Peradeniya Regional Contest," and click the globe icon in the standings column. It redirects directly to this document). https://docs.google.com/spreadsheets/d/1-utxaWV7frrWjOTRJuNoGmFtrSwXQbQN/edit?gid=1483018665#gid=1483018665

Полный текст и комментарии »

  • Проголосовать: нравится
  • +1
  • Проголосовать: не нравится