Rubric 1.0.0 · column 13 of 13
Touch points
navigator.maxTouchPoints — 0 on a plain desktop, 5 or 10 on touch hardware.
Under rubric 1.0.0 this column is worth between 0.56 and 6.64 bits, and the likeliest outcome the table knows — 0 — costs 0.56. Those figures are assumptions printed before your run. The counted box beside them is not: it is how many of the devices these sites have read share your value.
How a browser produces the touch point count
The browser reports how many simultaneous contact points the input stack will deliver, which comes from the pointer devices the operating system has enumerated rather than from the screen. A touchscreen laptop reports a non-zero value while sitting on a desk with a mouse attached; a tablet in a keyboard case reports the same value it always did. The number is a property of the hardware the OS found, not of how the machine is being used.
How rubric 1.0.0 prices the touch point count
Five outcomes with almost everything at one end. 0 is assumed at 0.68 and costs 0.56 bits. 5 is 0.16 and costs 2.64, 10 is 0.12 and costs 3.06, and 1 is 0.02 for 5.64. Anything else shares 0.01 and costs 6.64 bits. Touch points sit in the capability group with cores and device memory, so in most passes this reading is discounted to 35% and contributes a fifth of a bit.
The spread between the cheapest and the dearest outcome is 6.09 bits, and that spread is the whole of what this column can ever do to a total. Where the four band edges sit →
Every outcome this rule can produce
This is the printed envelope, written out. The pen cannot land outside it, because these are
all the answers rule touch has.
| Outcome | Assumed p | −log₂p |
|---|---|---|
| 0 modal outcome — the printed line | 0.680 | 0.56 |
| 5 | 0.160 | 2.64 |
| 10 | 0.120 | 3.06 |
| 1 | 0.020 | 5.64 |
| 2 any other count | 0.010 | 6.64 |
What changing the touch point count costs
There is nothing worth buying here — the whole column is under a bit for the common case. Its value is entirely as a cross-check, and that is where changing it is expensive: a user agent naming iOS beside zero touch points, or a desktop platform reporting 10, is a contradiction with no upside at all. This is the clearest example on the site of a signal whose importance is not what the bit count says it is.
The rubric's own reasoning, verbatim
Effectively binary on desktop (0) versus touch devices (5 or 10).
What moves it
The hardware. Touchscreen laptops report non-zero on Windows and confuse anything that treats this as a mobile test.
Where this prior is weakest
Worth under a bit for the common case, which makes it nearly worthless as a uniqueness signal. Its real value is as a consistency check — a mobile user agent reporting zero touch points is a contradiction, and contradictions are what vendors act on. This rubric scores rarity, so it does not price that at all.
Correlation
Grouped with CPU cores and Device memory under "capability". Whichever of them carries the most surprisal in a given pass takes full weight; the others take 35%. The chart draws the short stroke rather than describing it.
How the correlation discount works →
Challenge it
If you have a source that contradicts this assumption, the honest thing is to say so in
public against this specific prior. Open a challenge
against touch →
What people ask about the touch point count
Answered from rubric 1.0.0 itself. Where an answer quotes a probability, that probability is a constant in the shipped scoring code and is printed in the table above.
- Why is touch points worth so little?
- Because zero is assumed at 0.68 and costs 0.56 bits, and it is usually discounted to 35% of that against a louder member of the capability group. As a rarity signal it is nearly worthless, which the rubric states rather than disguises.
- Why does my laptop report touch points?
- Because the operating system enumerated a touchscreen, whether or not you use it. Touchscreen laptops commonly report 10, which is priced at 3.06 bits here and confuses any check that reads a non-zero value as evidence of a phone.
- Can touch points give away a spoofed profile?
- Easily, and not through this score. A user agent claiming a phone alongside zero touch points is a contradiction between two signals that always agree on real hardware. This rubric prices rarity and cannot see it at all.
Where to read next
- The prior for CPU cores — the other reading the capability group discounts this one against.
- The prior for device memory — the other reading the capability group discounts this one against.
- The complete prior table for all 13 columns — every rule, every outcome, and the rationale the rubric ships with each one.
- Assumed and counted: what a prior is, what the counted column beside it measures, and why the two are never added.
- Compare two passes for the delta on this column — change one setting and read what actually moved, instead of guessing from a total.
Where the pens move
Knowing which prior is weak is only useful if something can actually move the signal.
Engine level, not JS level. Donut Browser spoofs through CDP injection in its Wayfern Chromium fork, so the value a page reads is the value the engine reports — there is no patched accessor to catch.
It refuses to contradict itself. A profile whose proxy exit geography, time zone or language disagrees with its fingerprint does not launch. That is a pre-launch gate in the app, not advice in a manual.
Open, and quiet. AGPL-3.0, desktop app and sync server both in the repository, and no analytics SDK anywhere in the source. macOS, Windows, Linux.