First 100 Candidate Test
Run the same programmed sequence against one fixed sentence
This page explains the basic run sentence test supplied with
all.zip. It applies TBP-001 through TBP-100 to the contents of
sentence.txt and reports where the candidate passes or fails.
run sentence tests only
TBP-001 through TBP-100. Passing it does not complete the current 290-requirement
formal challenge. The remaining 190 requirements—including book-level,
composite, personal-experience, and category-specific 25/52 counterparts—must
be evaluated separately under allrequirements.htm
and rules.htm.
Three different kinds of comparison
- Single-sentence candidate test:
run sentenceapplies the first 100 programmed tests to one fixed input. - Population control: the random and corpus commands test many sentences and report how far they progress through the programmed sequence.
- Formal 25/52 requirement: each applicable TBP requirement uses the particular 25/52 counterpart defined on its frozen page. There is no assumption that one generic substitution command completes every category.
Download the program
Unzip all.zip into its own folder. It contains
all.jar, the Windows batch file used to run it, source code, and
supporting files. Keep the original download unchanged and record its date
and version.
Inspect the available commands
From the extracted directory, run:
This displays the options provided by that version of the program.
Fix and prepare the sentence
- Choose the exact sentence before running the formal test.
- Record its source, edition, original language, sentence boundary, and encoding.
- Place the prepared sentence in sentence.txt.
- Do not change its text, spelling, punctuation, or encoding after seeing results.
- If Hebrew substitution is needed, follow substitute.htm exactly and preserve both the original and prepared text.
A casual scientific test may use any disclosed sentence. A formal challenge attempt must also satisfy the historical-book, advance registration, computer-use, and version-freeze rules in rules.htm.
Run the first 100 tests
Preserve the complete unedited output, including every pass, failure, error, input file, program version, and run date. Do not report only the successful tests.
How to interpret the output
All 100 pass: independently reproduce the result before proceeding to the remaining 190 requirements.
A software error occurs: preserve the input and error message, identify the version, and resolve the discrepancy before crediting a pass or failure.
Population and random controls
The software also supports larger comparison runs. Those experiments require the generator or corpus version, settings, requested and completed trial counts, commands, and complete progression totals to be reported. The one-trillion command and its recorded result are documented in How to Test or Challenge the Theory.
An observed frequency from one generator describes that experiment. It is not automatically a universal probability for all possible sentences or all published book openings.
Continuing to the full challenge
- Confirm the first-100 result independently.
- Use allrequirements.htm for TBP-101 through TBP-290.
- Apply every requirement's frozen 25/52 rule where one is defined.
- Document the comfort, extra, and personal-experience counterparts.
- Report every composite requirement as one complete package.
- Follow the registration, submission, and unanimous adjudication procedure in disprove.htm.