table

When to use

• The table IS the slide. A plain markdown table on any slide already gets the house treatment — same type, rails, hairlines and zebra. Reach for table when the rows need what only a component can declare: a capacity budget, autosplit, the portrait card reshape, and the row/col/cell focus axes.
• Three or more columns, or four or more rows. compare-prose maxes out at two options with prose bodies. A table scales past that, as long as the cells stay short.
• Cells are short phrases. Each cell is a value, a phrase, or a state marker — not a paragraph. If the cells need sentences, use verdict-grid or cards-stack.
• Stable column meaning. Every row reads the same way across columns. Mixing column meanings row-to-row breaks the table's scannability.

When not to use

• Cells full of prose. Long sentences in a cell wrap awkwardly and force the column wider. Move to verdict-grid for criteria with body text, or cards-stack for full prose rows.
• More than 6 rows. Past 6 rows the table crowds the slide. Split across two slides or summarize the rows that don't differentiate.
• A table that only supports the prose around it. Then it does not need this class at all — write the pipe table on a content or un-classed slide and the universal treatment styles it. The component is for a table that owns the slide.
• Reaching for it when a specialist fits better. Mostly pass/fail badges is obligation-matrix or verdict-grid; term/definition pairs are glossary; a dated plan is roadmap.

Common mistakes

• Writing a vague or duplicate first column, assuming it's just another data column. When the table overflows a narrow box, the Fit Ladder reshapes it into row-cards (column headers become in-card labels) instead of clipping — the FIRST column becomes each card's title in that reshape, so it needs to be a genuinely identifying label per row. This happens automatically and needs no opt-in.
• Expecting the first column's **bold** to show when the column is already emphasized as a row label. The row-label emphasis sets the same weight and ink **bold** would, so bolding inside it is a no-op. no-row-label turns the emphasis off and hands the column back to you.
• Reaching for a CSS override to make a short table fill the slide. table-fill is the switch — it grows the table AND centers each cell in its band. The default hugs the rows and centers the block, so a three-row table reads as three rows rather than three rows stretched over a screen. The centering is safe: a table too long for its slide or pane keeps its header row at the top and clips at the bottom, never at both ends.

Authoring

<!-- _class: table -->

## Heading framing the comparison.

| Criterion | Option A | Option B | Option C |
| --- | --- | --- | --- |
| First criterion | Value | Value | Value |
| Second criterion | Value | Value | Value |
| Third criterion | Value | Value | Value |

Slots

Anatomy

┌─────────────────────────────────────────┐
│  header                                 │
│  LABEL                                  │
│  Here are the numbers side by side.     │
│                                         │
│  ┌───────────┬───────────┬───────────┐  │
│  │           │ Option A  │ Option B  │  │
│  ├───────────┼───────────┼───────────┤  │
│  │ Row 1     │ ✓         │ ✕         │  │
│  │ Row 2     │ ✕         │ ✓         │  │
│  │ Row 3     │ ✓         │ ✓         │  │
│  │ Row 4     │ ⚠         │ ✓         │  │
│  └───────────┴───────────┴───────────┘  │
│  Footnote text for scope caveats.       │
│  footer                          11/19  │
└─────────────────────────────────────────┘

Guides

• The six answerswhat each state marker means
• Marks in layoutsthe state-cells switch that draws marks in table cells

Related

• compare-proseexactly two options with prose bodies
• verdict-gridoptions scored against criteria with pass/partial/fail badges
• obligation-matrixmany regimes compared against shared obligations
• cards-stackeach row needs full-prose breathing room rather than a tabular cell