TAC/AA-inspired benchmarks
Reproduce prepared mechanism and baseline-strategy cases.
Open benchmark catalogADVANCED RESEARCH · BEGINNER WALKTHROUGH
See how one research question is tested.
Follow the question, the one setting we changed, the result, and the limits of that result. You do not need a statistics or advertising background.
8 simulated advertisers16 kinds of shopper searches5 ad positionsUp to 60 simulated days
What is TAC/AA, and why is it mentioned?
TAC/AA was a 2009 academic competition in which computer programs managed simulated advertising campaigns. AdMechanica borrows its general market structure but does not claim to be the original competition server.
NEW · MODEL CHECKS
Open model checksBefore you trust a finding, check the model behind it.
Run prepared, repeatable checks of documented behavior and see the simulator’s limitations in plain language.
NEW · COMPLETE RESEARCH JOURNEY
Open the research journeyStart with a question, finish with a reusable finding.
The new journey connects strategies, market conditions, repeated tests, evidence comparison, and publishing in six plain-language steps.
REPRODUCIBLE RESEARCH LABS
Move from reading a result to testing it yourself.
Both laboratories save matched repeat plans, explain their limits, and keep every conclusion linked to its source evidence.
Strategy Evaluation Laboratory
Test where a server-owned bidding strategy stays stable or struggles.
Evaluate a strategyLoading released research evidence…