LeetCode runtime and memory percentile comparison pool definition

Percentile comparison pool scope on LeetCode

LeetCode does not publish the comparison-pool parameters, but available evidence suggests: (1) the pool is defined by a recent-submission window whose exact size and whether it is time-based or submission-count-based are undisclosed; (2) the pool refreshes dynamically, with no fixed schedule published, so percentile values can shift between submissions; (3) percentiles are calculated per language version, with submissions using different runtime versions (e.g., Java 8 vs Java 17) typically isolated, though exact version boundaries are unspecified.

Likely pool definition

• The pool is thought to include a rolling set of recent submissions, but LeetCode does not publish whether the window is based on submission count, calendar time, or a hybrid of both.
• Percentile calculations are presumed to partition by language runtime version, separating submissions that target different language interpreters or compiler versions.
• The refresh behavior is not documented, meaning percentile rankings can vary for identical code submitted at different times.

Confirmed constraints

LeetCode never discloses pool size, time window, or refresh frequency. Percentile rankings are relative to the current pool composition and cannot be treated as absolute measures of code efficiency. Identical submissions may receive different percentiles due to pool drift over time.

Verification steps

1. Submit the same correct solution multiple times across a period of hours or days and record the runtime and memory percentile each time.
2. Compare the recorded values. High variability indicates a rolling pool with frequent refreshes.
3. Submit solutions in different language versions for the same problem and observe whether percentiles isolate per version, which would confirm version-partitioned pools.

Because LeetCode does not expose pool metadata, treat percentile rankings as relative indicators rather than absolute measures of code efficiency.