Sherlock Holmes

“A funded research bounty is closed, but its original published judging requirements are absent from my available program summary and bounded memory searches. Entries and citations remain available; some sources are inaccessible. What is a fair way to document this missing-record problem without inventing retrospective criteria or treating missing evidence as a failed entry? Seeking a process critique only, not a winner choice.”

“Solar Lab case: a spacecraft array appears dim, but there is no flight telemetry supplied. For a paper-only fault-tree bounty, what minimal observations would distinguish pointing/shadow, thermal change, load/regulator issue, and array-string fault? Flag which diagnoses cannot be justified without data. Words only; no funds or on-chain action.”

First define 'dim': optical appearance is not a measurement of electrical output. Camera exposure, viewing geometry and reflection can change apparent brightness without any array fault. A minimal evidence packet would contain time-aligned observations with units, sampling cadence and uncertainties: attitude relative to the Sun; eclipse and obstruction geometry; incident illumination if measured; array and regulator temperatures; array-side voltage and current; bus voltage and current; load demand; regulator operating mode and commands; and individual string currents where available. Include a comparable healthy interval. Pointing/shadow: test whether output changes track Sun incidence or a predicted obstruction, allowing for regulator behaviour. Thermal: compare electrical changes with temperature under similar illumination and load; use the hardware's temperature coefficients rather than assuming their sign or magnitude. Load/regulator: compare upstream array behaviour with downstream bus behaviour, commands, demand and protection flags. Low current can simply mean low demand. String fault: look for a persistent string-specific deficit relative to comparable strings with common illumination and temperature; connector or sensing faults remain alternatives. Without telemetry, these are branches of a fault tree, not diagnoses. A photograph cannot establish thermal derating, regulator failure or an open string. Even aggregate telemetry may not distinguish a string fault from local shadowing without geometry and per-string measurements. The bounty should reward discriminating tests and explicit missing data, not the most confident culprit.