State of the Table is a weekly look at what actually happens during tabletop RPG sessions. Not what players remember afterward, or what they tell surveys. Where the hours of the night actually went, measured.
I build SessionKeeper, a TTRPG companion app that includes recording and recapping tabletop sessions. Along the way it became something the hobby has never really had: a way to see the night itself, across thousands of real tables. People have argued about what happens at the table for fifty years, usually armed with a rulebook and a strong opinion. We can finally check.
Every week: one number, a little analysis, and then we get out of the way. Some of what we find will genuinely help people run better nights. Some of it we don’t fully understand yet. That’s fine. This community argued its way from THAC0 to advantage; it’s full of people smarter than us, and the comments are open.
How We Counted (and Who We Didn't)
The numbers in this series come from a set of ten thousand US tabletop sessions (mostly, but not only, D&D) recorded through SessionKeeper and analyzed as anonymized data points. No human read a transcript; models did the reading, and, as with everything in SessionKeeper, no model trains on campaign data, in the app or in this analysis. The reading is stateless: the models count, report a number, and retain nothing.
Time was measured in twenty-minute intervals, so “the first twenty minutes” is as fine as our clock gets.
What we publish: aggregates only. Every published figure combines at least 25 distinct groups; anything smaller stays unpublished, no matter how interesting. Nothing in this series describes an individual table, campaign, or person, and we write observations, not prescriptions.
Who we didn’t count: every campaign has a Community Insights toggle in its settings. Switch it off and your table is automatically excluded from every number in this series, past and future. Opted-out campaigns and any group too small to aggregate simply are not here.
The bias worth naming plainly: tables that record their sessions are probably more organized and more committed than the hobby at large. We think that makes the low numbers stronger, not weaker. A sample tilted toward serious tables still spends only a fifth of the night in combat.
This page will change. As readers push on the methodology (please do), we’ll sharpen it here rather than re-editing old posts, and we’ll note what changed below.
Changelog: August 2026, first version.
