# Short Deck practice chooser

Checked September 12, 2026 against Short Deck Poker Trainer 1.5.0 (source build 10).

## Purpose and files

The six cases in cases.json are original editorial examples for choosing a current app task. chooser.html presents the same cases as a responsive page with one-page A4 print styling. Save the HTML file for offline reading or use your browser's Print command. It has no scripts, form, app connection or tracking requests. The optional article link requires a connection.

The companion article is https://gtogecko.com/blog/short-deck-daily-challenge-mistake-deck . GTO Solutions AS publishes GTO Gecko and Short Deck Poker Trainer. The chooser is separate editorial material, not an in-app screen or an import file.

## Evidence

The current official US Apple product listing is https://apps.apple.com/us/app/short-deck-poker-trainer/id6760950166 . Current release and feature wording were checked on September 12, 2026. Current authorized source and local tests support the workflow distinctions; the store description alone does not establish every persistence detail.

We ran 180 distinct local tests using Flutter 3.47.2 and Dart 3.13.2: 95 initial tests, 53 in an extension run and 34 in a final run. Two tests were repeated after stronger assertions, so there were 182 successful test executions. Coverage included Daily generation/storage/completion, local-calendar streak handling, Mistake Deck scheduling, landing-screen states and History detail/retry behavior.

Two focused checks established the main distinction: five wrong Daily answers with zero score can persist as a completed day and one-day streak; and a scheduled correct review can change an item's due date while preserving original Daily answers/score, full original History records and the ordinary training meter. These checks use local fixtures and application source, not user analytics or live gambling results. They do not test whether the reference strategy is optimal.

## Manual inspection protocol

Use a version whose controls match this guide. These steps may create practice records in your app; they are optional. Keep any personal records on your own device. Do not clear a deck to carry out these checks.

1. On a new Daily Challenge, record the date and answer the first spot. Leave, return through Resume and check that the earlier answer remains recorded.
2. Continue through the remaining spots. After the final answer, choose See your result to finish. Record completion separately from score; a wrong answer is not an incomplete day.
3. Reopen See today's result. Confirm that it displays the completed record rather than letting you replace the original answers.
4. If a saved qualifying mistake is due, note its due state and the original result. Choose Mistake Deck, then Start review. After a saved review, compare the review state with the retained original record. Do not assume every nonmaximum grade creates a deck item.
5. If All caught up appears while saved items remain, inspect their next due times. Distinguish this from Review status unavailable, which requires reading the displayed status or retrying.
6. For an immediate retry of a retained History decision, open its detail and choose Try again · unrated. Compare the original result and the Mistake Deck schedule; this immediate retry does not advance scheduled-review progress.

An observed mismatch may be a version, content or storage issue. Note the version and visible labels, and use Short Deck support for the specific condition: https://gtogecko.com/shortdeck/support . Do not infer an empty deck from a failed read.

## Reproduction boundary

cases.json and chooser.html expose the full editorial mapping; SHA256SUMS checks file identity. The website's article generator and QA script are retained with its editorial research bundle. The application's private source, proprietary scenario rows, local test source and raw store payloads are not redistributed. This public protocol lets readers inspect behavior; it is not a public reproduction of the private automated test suite.

No random simulation, participant sample or learning intervention was used. The cases are constructed navigation examples, not learner data. No installed App Store binary, physical phone, dedicated screenreader or physical printer was tested. No guarantee of skill, learning, rankings or winnings follows from a completion, score, review-stage label or test pass. Calendar statements use local dates; they are not a fixed 48-hour promise across clock or timezone changes. Recheck after relevant app changes.