How we source our data

What we start with

• We start with real tune files and logs.
• A tune file holds the tables the computer runs.
• A log holds what the engine did.
• We read both, table by table and channel by channel, on the same software version you use.
• We also read the platform software's own documentation for table names, units, axis labels and limits.
• That is how a guide can name a table the way your editor names it.
• We test procedures on research machines before we publish them.
• A research machine is a test computer with the tuning software, a known file and a log.
• If a menu path, a unit or a table does not match, the guide waits.
• Every guide is reviewed and fact-checked before it publishes.
• The reviewer checks the tables, the units, the ranges, the safety notes and the links.
• Tune service files are compared against our 3 million+ tune values, with a final check-off by our lead tuner.

How ranges come from real files

• A range on TuneBible is a summary of values in real files.
• We group files by platform, engine, computer, cam, fuel, induction and transmission where the file records those details.
• We then show the low end, the high end and the middle of the group.
• We count files, not cars.
• One car can have several revisions, and each revision is one file.
• A group with a small sample is labelled with its file count.
• If the spread is wide, we show the spread instead of a single target.
• Example: a group of files with a similar cam idles between 750 and 850 rpm.
• The guide shows that range and the number of files behind it.
• It does not turn that into a rule for every engine.
• A value from one file is an example, not a range.
• We say which one it is.

What confirmed and unconfirmed mean

• Confirmed means we have more than one reason to trust the claim.
• The platform software's documentation describes the table or function, and at least one real file or log shows the same behavior.
• Where we can, we reproduce it on a research machine.
• A confirmed table name, unit or range has been checked against the software and a file.
• A confirmed behavior has been seen in a log and matched to the table that controls it.
• Unconfirmed means one file, one log or one document supports the claim and we have not reproduced it.
• The page says unconfirmed in plain words and states what is missing.
• We do not fill the gap with a guess.
• When a claim moves from unconfirmed to confirmed, the page is updated and the date changes.
• When a claim cannot be confirmed, it stays marked or it is removed.

How a guide is checked before publishing

1. A tuner writes the guide. The draft comes from the files, logs, software documentation and research-machine tests, not from a rewrite of another site.
2. A review checks it. Every table name, unit, range and safety note is checked against the same material. A claim without support does not pass.
3. We test the steps. The procedure is run on a research machine with the software version named in the guide. Menu paths and table labels must match what a reader sees.
4. We check the page. Links, figures, captions and tables are checked before the guide publishes. Every figure needs a caption and every table needs a header row.
5. We date it. The guide carries a published date. A later change is noted on the page and the updated date moves.

How the 3 million+ tune values are counted

• A tune value is one interior cell or one line item in a stored tune table.
• It is not a tune file, a table or a car.
• Axis rows and columns are excluded.
• We count the values in the LS, Mopar, Ford, and Honda & Acura reads we hold.
• The figure is one combined total across all of them, LS included: 3 million+ tune values, last updated 6 November 2026.
• The count starts with the LS library at a 3,000,000-value floor.
• It adds 6,929 Mopar, 2,001 Ford and 17,603 Honda & Acura values.
• That is at least 3,026,533 counted values, rounded down to 3 million+.
• The LS component is a floor, so the total is conservative, not an exact sum.
• It is not a promise that a close match exists for every car.
• We publish one combined figure rather than four separate totals.
• The figure describes the size of the stored library.
• It does not describe the size of the comparison cohort for a particular file.
• A report names the comparison set it actually used, including when that set is thin.

What we do not use

• We do not copy posts from discussion sites, social media comments or another site's guide into a TuneBible page.
• A claim rests on the file, the log, the platform software's documentation and our own test.
• Where a claim depends on one of those, the guide names which one and what it shows.
• We do not invent a number to fill a table.
• We do not publish a range from a single file.
• We do not present a value from one car as a platform rule.

Limits

A guide can describe what a table does and what a range looks like in the files we hold. It cannot inspect your engine, your fuel system or your sensor calibration. Those need a file, a log and, where the fault is mechanical, a physical check.

Remote work has the same limit. We read what you send. We say when the file or log does not show enough to judge.

Corrections

If a page is wrong, tell us. The corrections page explains what to send and how the fix is noted on the page.