The Laboratory
Run counterfactual experiments on the observed world. Choose one structural change and one period of history; the Laboratory replays that period with the change in place and reports how the result differs from what actually happened.
select intervention
one condition is changed; nothing elseselect historical period
reading the observed record…The changed condition is put back to exactly the value the observed world runs under, at the end of the period above. Everything the experiment has accumulated by then is kept. From that year onwards both worlds run under identical conditions, so any remaining difference is a difference the history itself carries — not a difference in conditions.
A period can only begin in a year the observed world still holds a complete state for, so the experiment intervenes on the observed state itself rather than on a reconstruction of it. No such year is available yet.
run experiment
the result is compared against the observed worldno intervention selected