Game Game Dojo is an online dojo for learning thinking games like chess and Othello by playing them. This page explains how we make our articles and analysis — our commitment to you as a reader.
About the Site
Game Dojo is run and edited by one person, Mitchie. Articles are published under that name, and the author is responsible for their content (about the author). For questions or to report an error, please use the contact form.
How We Write Articles
We care more about things you can verify on an actual board than about neatly summarizing general knowledge. Where possible, we follow a process like this:
- Choose a topic and a concrete position that illustrates it well.
- Check that position with a game engine or exact analysis.
- Confirm the numbers and conclusions, and edit the text by hand.
- For scientific or historical claims, verify sources where we can.
How We Use Engine Analysis
Game Dojo's strength is that you can actually play and analyze positions. When an article states an evaluation or conclusion, we base it on analysis rather than opinion.
- Chess: move evaluations and best-move calls come from Stockfish 18 Lite WASM run at depth 20. The home page "Daily Tactics Quiz" draws from the Lichess open puzzle database — public domain (CC0) and already engine-verified.
- Othello: the home page "Daily Othello Puzzle" is produced by our own Othello solver, which exactly solves the endgame to compute the true best move and final disc margin. The answer shown is a calculated best move, not just a rough guide.
Engines have assumptions and limits, though. An evaluation is one indicator — it doesn't mean "every other move is bad."
On Science and Health Topics
For topics touching science or health (such as "how chess affects the brain"), we avoid absolute claims. Even where research exists, results vary by subject and method. Rather than saying chess "will make you smarter" or "prevents dementia," we distinguish what is and isn't known — for example, "some studies show an association" or "this has not been proven."
On Our Use of AI
We sometimes use AI to draft or structure text. However, we do not publish AI output as-is. Choosing positions, checking analysis results, verifying facts, and final editing are all done by a person. An article's value comes not from "an AI wrote it" but from "we actually verified it on the board."
The analysis data is published
The numbers in our articles are the raw output of the analysers. So that a reader can check them, we publish the data itself.
- chess-analysis.json — everything behind the Chess Analysis Lab: positions (FEN), candidate moves, evaluations and the engine's principal variations.
- othello-analysis.json — the Othello positions we discuss, with every legal move solved to the end of the game.
- othello-stats.json — the 400-position counts: how often each rule of thumb was best, and the median, maximum and win-losing share of its cost.
What it takes to reproduce them: for chess, stockfish-18-lite-single.js from npm [email protected] (single-threaded) with Hash 256MB, ucinewgame before each position, and a fixed go depth 20. Each candidate move is played first and the resulting position searched, then the score is negated so that it reads from the point of view of the side that moved. One thread, a cleared hash and a fixed depth make the run repeatable; a different engine version or hash size will give different numbers.
For Othello, our own solver (negamax with alpha-beta) reads the position exactly, to the end of the game, so nothing depends on search settings. Positions are sampled with a fixed random seed, so the same script produces the same positions.
Where we have a record of the day an analysis ran, the article shows it as its last-updated date. Where we do not, no date is shown.
Corrections
When we find an error, we fix it. We deliberately choose evergreen topics so articles stay accurate over time. When we do cover information that can change (records, rankings), we make the "as of" date clear in the text itself. If you spot something, please let us know.
Contact
For questions, bug reports, or feedback, please use the contact form.