Active study · Next review December 15, 2026

TrainingPlan.dev search and AI discovery follow-up

Track fixed search periods, page groups, safe totals, AI citation runs, leads, and poor results against the public baseline.

Inspect the active study →

Studies and open protocols

Plans that can be checked before results exist

Status and next-review dates show that these are not completed studies.

Planned · Review Dec. 15, 2026

Public bot access test

Protocol: Select sites before requests. Record bot purpose, status, rules, edge block, HTML, sitemap, and limits.

Open the test record →

Baseline next · Review Dec. 15, 2026

Fixed prompt citation test

Protocol: Freeze questions and products. Save every trial, answer, citation, fact check, place, account state, and date.

Open the run log →

Planned · Review Dec. 15, 2026

Utah answer results

Protocol: Build the candidate set before runs. Separate local facts, citations, rank, and reviewer judgment.

Open the scoring rubric →

Planned · Review Dec. 15, 2026

Page feature citation test

Protocol: Approve pages and measures first. Compare direct sources, clear answers, original data, and set layouts.

Watching · Review Dec. 15, 2026

llms.txt use

Protocol: Track file requests and later citations. Do not treat a time link as proof that the file caused a citation.

Read the current note →

Research background

What kind of researcher is Bob Bodily?

Bob earned a PhD from Brigham Young University. He has worked as a data scientist and research leader. That training helps him ask clear questions, show his method, and fix errors. It does not make every claim right.

See Bob's Google Scholar record →

Rules for each study

  1. Set the question, sample, measures, scores, and limits first.
  2. Save the dates, inputs, sources, code, and allowed outputs.
  3. Separate what happened from what caused it.
  4. Report missing, weak, bad, and mixed results.
  5. Share ownership, sales interests, and reviewer ties.
  6. Give each release a version and fix major errors in public.