Load and sustainability

A sustainable deck is one whose returning reviews fit the time you can repeatedly give it. New cards are optional intake, while old cards return according to earlier study. Software documentation also warns that settings which demand higher retention shorten intervals and increase daily reviews.¹ The practical judgement is simple: control intake before the review queue controls you.

The available measurements show substantial use, but they do not define a normal workload. In an eleven-month log study of twenty-five first-year medical students using self-selected Anki decks, every card presentation, including repeats, counted toward the totals. Students above the class median on three exams completed more presentations and used Anki on more days, but the observational study did not establish that the additional use caused their scores.²

Other medical-education studies show why completion matters. Among 139 Anki users responding to one survey, 92 reported daily review, but the study did not measure actual review time.³ In a physical-examination course with 120 students, completion of repeated emailed-question cycles declined from 106 students in the first cycle to 91 and then 85.⁴ In a six-month ultrasound-image program, 29 of 36 medical students completed the study, yet participants completed only 38.6 percent of assigned cases on average.⁵

The diagram presents a reasonable operating procedure for setting new-card intake: observe returning reviews across ordinary days, then reduce optional intake when those reviews already consume the available study time.

<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 590 280" role="img" aria-labelledby="load-title" font-family="system-ui, sans-serif" color="currentColor">
  <title id="load-title">Illustrative procedure for adjusting new-card intake to observed review load</title>
  <g fill="none" stroke="currentColor" stroke-width="2">
    <rect x="25" y="55" width="160" height="170" rx="12"/>
    <rect x="215" y="55" width="160" height="170" rx="12"/>
    <rect x="405" y="55" width="160" height="170" rx="12"/>
    <path d="M185 140 H210"/>
    <path d="M375 140 H400"/>
  </g>
  <g fill="currentColor">
    <path d="M210 140 L200 134 L200 146 Z"/>
    <path d="M400 140 L390 134 L390 146 Z"/>
  </g>
  <g fill="currentColor" font-size="16">
    <text x="50" y="88">Observe reviews</text>
    <text x="50" y="116">on ordinary days</text>
    <text x="50" y="162">Record how much</text>
    <text x="50" y="188">session time they use</text>
    <text x="240" y="88">Compare with</text>
    <text x="240" y="116">available study time</text>
    <text x="240" y="162">Do reviews repeatedly</text>
    <text x="240" y="188">fill the session?</text>
    <text x="430" y="88">Adjust intake</text>
    <text x="430" y="134">Reduce or pause</text>
    <text x="430" y="160">optional new cards</text>
    <text x="430" y="198">Recheck later</text>
  </g>
</svg>

Figure: Observe reviews across ordinary days, then reduce optional new-card intake when they already fill the available session. With twenty minutes available, first observe how much returning work fits without rushing. If it already fills the session, pause or reduce new cards. Resume intake only when the resulting review load remains manageable across ordinary days, not merely during one unusually productive session.

Boundary. These are reasonable operating procedures, not experimentally validated limits. The cited sustainability measurements come largely from medical education and cannot supply a general daily card count, review duration, or abandonment rate for students, professionals, and language learners.