# The Quad Slip, held

> **THE READING PLAQUE.** *The Quad Slip, held*, draft 0.4 — **made readable on the SX
> surface by its covering persona's reading grant; a draft, not the published form.**
> The grant stands by the operator's word (2026-09-15), verbatim: "ok - that works well.
> I would now want something of this kind for the quad slip." Covering: **starl3n** ·
> endorsing: **Link Digital** (the standing pair). The Foundations · 2: the second reading
> of the foundations series — an experiment of the observatory node, held on this node.
> The construction record stands in the open workshop at
> [/sx/workshop/the-quad-slip-held](/sx/workshop/the-quad-slip-held); the experiment's
> own door, [opendata.ai/quad-slip](https://opendata.ai/quad-slip), wears the word
> *grown · held* and links back here while this grant stands. Reading is free; the way
> across is a statement of intent ([/sx/matters](/sx/matters)). This file's sha256 is
> pinned in the reading register.

*Draft state — this block records the state of the draft; the reading begins after the
rule.* Draft 0.3 — 16 September 2026: the baseline's place in the series set out at the
operator's word ("I'd like it to cover why the baseline is the setting of most interest
for the follow on experiments in the series of 10"); restamped the same day at his word
("it is specifically 1.1e-7") so that the baseline's detune is stated precisely wherever
the baseline is meant, the record's 1e-7 probe named as a probe — the first 0.3 mint
(pin 14855922…) superseded by this restamp, never erased; draft 0.2 (16 September 2026,
pin 43604668…, the explainer trued to the experiment) and draft 0.1 (15 September 2026,
pin 55a92d4b…) superseded, never erased. Author of record: Steven De Costa; composed at his
word from the experiment's record. Standing: a reading, stated never asserted; the record
it reads stands whole without it. The later locks and the walk land in a further draft. Draft 0.4 (16 September 2026) places 3 figures at the operator's word, each drawn from the record's own numbers and captioned with its source; the screenshots the operator will supply are not yet placed.

---

## What the Quad Slip is

Four figure-2 slips in three dimensions, sharing one centre at (12.5, 0, 0) and the
vertical y-axis through it. Slips 1 and 2 are the Double Slip's facing pair in the x–y
plane: one at the origin facing +x, one at (25, 0, 0) facing −x. Slips 3 and 4 are that
pair turned a quarter-turn about the y-axis, so they face each other along z in the y–z
plane, at (12.5, 0, ±12.5). Two facing pairs in perpendicular planes, meeting along the
shared vertical.

Each of the four is a full slip — the Slip's A and B loops, its own catches — computed
by the same unchanged generator. The four are independent by construction: each is its
own pass, and nothing couples one to another. Under the congruent placement they are
identical: the same T, D, slip and beat, and four catches each at 21 laps with the mask
(2,0,3,1).

![The four slips in three dimensions: the x-pair at (0, 0, 0) and (25, 0, 0) facing each other along x, the z-pair at (12.5, 0, ±12.5) facing along z, the shared vertical through the centre (12.5, 0, 0), and the central gauge where the x-gap and the z-gap are each lifted onto y. A schematic of the placement, not to scale.](/sx/assets/foundations/the-quad-slip-held/quad-slip-geometry.svg)

## What it observes

Two gaps at the centre. The x-gap is the distance between the B points of slips 1 and 2;
the z-gap the same for slips 3 and 4. Each is lifted onto a central gauge as a height in
y, as the Double Slip lifted its single gap. While the four are congruent the two gaps
are equal: every reading sits on the diagonal of the (x-gap, z-gap) gauge, and the
y-differential x − z is zero. The record's hypothesis, held to be seen in the data and
not asserted, is that x and z confine y into a projected differential: flat while the
four agree, emerging as their placements diverge. The experiment is the set of ways to
make them diverge, and the readout of what y does.

## The controls

- **x_sep and z_sep** (25 each at baseline) set the placement of each pair. A placement
  difference gives a constant y-projection of x_sep − z_sep, a level line across every lap.
- **z_detune**, δ, lengthens the z-pair's dyadic bus and so detunes its slip while the
  x-pair is left untouched. The z-gap drifts and the y-projection oscillates: at δ = 0.5
  the z-pair's catch count moves (571 → 550) and the y-projection lifts to about 21 at one
  beat; at δ = 16 the z-gap flips sign almost every lap, swinging between +24 and −29,
  and the y-projection swings between about +46 and −20 and never settles. The two kinds
  of difference compose: a placement offset shifts a detuned trace up or down.
- **readout_dp**, which the record calls Pi: the precision at which the y-projection is
  read, 4 decimal places at baseline.
- **laps**, up to 4096.

The page carries the drag-to-orbit scene of the four slips around their gauge, the
(x-gap, z-gap) gauge with its diagonal, the y-projection traced lap by lap with each
catch coloured facing or crossed, a z-detune sweep card, and a button that runs the
baseline calibration below.

## The precision floor, twice

**The dissolved crossing.** An earlier reading had the x-gap cross 21 at lap 735 = 35×21.
It was an artefact of rounding the bus length to 4 decimal places: over some 700 laps the
rounding drifts the gap by about 0.04, enough to lift a true 20.96 to a spurious 21.02.
Computed at full precision, consistent with the catches, the gap approaches but never
crosses its geometric ceiling of 21 (25 − 2·2, the mask's inner edge); the true maximum
is about 20.98, at lap 229. The engine now reads the gap consistently, and the record
keeps the correction as it fell.

**The coherence horizon.** A near-zero detune with x_sep = z_sep reads as a perfect
y = 0 until the floor is crossed. The record's probe of this, at δ = 1e-7 with the wide
mask, held to lap 128 and "broke" at lap 129. The break is the readout floor, not a
dynamical confinement: the raw y-projection is nonzero from lap 1, a linear drift of
roughly k·δ·lap with k ≈ 4, but it sits below the 4-decimal readout until it crosses it.
So the break obeys a clean reciprocity — break_lap ∝ 1/δ (1e-6 → 13, 1e-7 → 129,
1e-8 → 1281) and the amplitude ∝ δ — and at δ = 1e-9 the drift never surfaces at all.
The 128 is not 2⁷; it is wherever that drift meets four decimals, and it slides with δ or
with the precision. The baseline below is not that probe: its detune is
specifically 1.1e-7 (0.00000011), and with the mask (2,0,3,1) its crossing is the bit
onset at lap 137. The record's reading, marked interpretive: the confinement is
epistemic, an observability horizon, not ontological; the coherence time of two
near-identical clocks is about floor/δ, a precision–duration reciprocity.

![The coherence horizon as the readout floor: the break lap against the detune δ with the wide mask, on logarithmic axes — the record's three probes, 1e-6 → 13, 1e-7 → 129, 1e-8 → 1281, on the line of the reciprocity break lap ∝ 1/δ. The record's probe values; the line is the law they obey.](/sx/assets/foundations/the-quad-slip-held/quad-slip-break-lap.svg)

## The mode-locking test — a null result

Two readings were put to the sweep. The tennis-racket reading took y for the
intermediate axis of a Dzhanibekov flip, the broken x/z symmetry resolving into it; and
at 21 laps the wide mask shows a 13:8 winding ratio, near the golden mean, which a
mode-locked pair would privilege. The z-detune sweep reads, at each step, the winding
ratio z/x and the y-amplitude. Across detune from 0 to 1.2 the ratio descends smoothly
(at 200 laps from 1.0 to 0.87, through fifteen or more distinct ratios, the widest
equal-ratio plateau about 0.03 of the range), and the count jitters as it falls — 50
drops against 45 rises — the aliasing of a continuous ratio, not the plateaus of a lock.
The 13:8 was adjacent-integer aliasing: at 21 laps the z-count passes through every
integer from 13 to 7, and 8 is one narrow step.

The reason is structural. Mode-locking needs coupling, and the pairs are independent by
construction, so there is nothing to lock: the y-projection is a linear beat between two
uncoupled clocks. That also refines the tennis-racket reading to an appearance rather
than a mechanism — the inversion is a beat's sign-flip. The fork the experiment surfaces
is that true mode-locking would require coupling the pairs, at the cost of the
independence the design chose.

![The winding ratio z/x across the z-detune from 0 to 1.2 at 200 laps, as the page's sweep card runs it: fifteen distinct ratios, the widest equal-ratio rung about 0.04 of the range, the ratio 1.0 at δ = 0 and 0.82 at δ = 1.2, no plateau — the null result for mode-locking. Measured: the sweep run on 16 September 2026.](/sx/assets/foundations/the-quad-slip-held/quad-slip-winding.svg)

## The series result — the baseline calibration

At one calibration the quad slip stops being a demonstration and becomes an instrument.
The calibration is pinned: mask (2,0,3,1), x_sep = z_sep = 1.575, z_detune
specifically 1.1e-7 (0.00000011), readout 4 decimal places, 1024 laps, the constructed
yellow dot. It is parameter-free
and deterministic, so it returns the same reference row every time: 90 catches, a
y-projection of 0.0004, and the six-node ground.

At this calibration the y-projection is a featureless linear ramp whose only content is
its rate, proportional to δ. Read through the 4-decimal floor it presents as a frame: an
a-priori origin at lap 1; five sub-readout zeros, laps on which was = is with no
differentiation — 1 + 5 = 6 nodes on y = 0; then the bit resolves, differentiation
begins at lap 137, and climbs in a staircase of about ten segments per epoch. The run
reports the frame as a field: the a-priori lap, six zero nodes, five pre-spacetime
zeros, the bit onset at 137, ten segments in the first epoch.

The record's discipline travels with it: the 6 and the ~10 are calibration outcomes of
beat × z_detune × readout precision, robust over a band of detune, not asserted
constants; the invariant is the ramp, and the readout precision is the grain at which
the bit is cut from it. Pi sets that grain, and with it the pre-spacetime depth:

| Pi (dp) | ingress period | bit-onset | pre-spacetime depth |
|---|---|---|---|
| 3 | ~2327 laps | 1178 | 52 |
| 4 | ~233 laps | 137 | 6 = 1 + 5 |
| 5 | ~23 laps | 12 | 2 |
| 6 | ~2.3 laps | 2 | 1 |

The record reads this as the quintic threshold, and marks it interpretive: the quartic
is the last polynomial solvable by radicals and the quintic the first that is not, so
the continuous ramp, which has no closed form, is resolved into the discrete staircase
computationally, at the rate Pi. Its hypothesis, stated as a hypothesis: this is a
functional sub-spacetime calibration — a pure 'is' onto which the epistemic bit anchors,
so that a unit fixed on it makes the address of a step (epoch, position in the epoch,
amplitude, elapsed) a measure of relative time or space, addressable over a multiway
path selection.

## Why the baseline is the setting that matters

The page shows the reason. Run the baseline and look at the y-projection over laps. Every
other setting of this experiment draws a wobble — a gap drifting, a sign flipping, a
line lifted off zero by a placement. The baseline draws a ruler. Three pairs of points
sit flat on the zero line: the a-priori origin and the five zeros, the six nodes of the
pre-spacetime. Then, at lap 137, the swings begin, and they come in epochs — five swings
up and five down at one height, then five and five at the next, each epoch one step
higher than the last, the heights cut by the four-decimal floor into level steps. Nothing
in the run put that structure there. The rate of the ramp is the only free number — the
detune, pinned at specifically 1.1e-7 — and the floor turns it into a staircase whose
every landing is a catch lap of the substrate.

That is what makes it the calibration, and what the rest of the series of ten is built
on. Three properties do the work.

It is the same for everyone. The baseline is parameter-free and deterministic: any actor
who runs it, on any node, gets the identical row and the identical frame, whatever the
composer holds. The page's button exists so that no hand-set run can perturb the
reference. A ruler two parties cannot both read is no ruler.

It gives every later reading an address. The frame the run reports — epoch, position in
the epoch, amplitude, elapsed since initiation — is a place to stand, and its landmarks
are integers of the substrate's own comb: the first bit at 137, the horizon at 237, and
the ladder that follows, 499, 952, 1867. Later instruments do not choose their constants;
they inherit these.

And the instruments that follow read it as their ground, in the record's own words.
The Hinge is "the first experiment to use the pinned quad-slip baseline as an
instrument": two causal actors each instantiate the same 'is' and stand in a relational
difference, and every horizon that difference produces — 237, 137, 92, 47, 46, 12, 11, 2,
1 — is a catch lap of the baseline, a landmark rather than a point between. The Double
Hinge finds that at the unit difference the being is "lap 137 — the ground's own first
bit", and that the fold's coherent precision is Pi 4, "exactly the baseline's own
readout": the ground reads itself. The Quad Hinge folds two double hinges "on the same
baseline" and finds that they land on the same fold independently, on beat-harmonic
plateaus at 46, 92 and 137. The XOR reads "being = 137 = the baseline's own bit-onset"
and confirms the a-priori '1' up the fixed ladder 137 → 237 → 499 → 952, flipping to the
quintic at 952 — exactly where the horizon spills past the baseline's own 1024-lap
budget and can no longer close. The XOR Mirror measures the origin's being × rate as 137
against the fully confined floor near 116.7, which is where the residual of about 17.3%
comes from, and reads its confinements from the 2¹⁰ baseline window. The Channel makes
each of its instruments "a quad-slip (its y-projection flip-signal over loops)", finds
the optimal rate 0.25 "at the pinned 2¹⁰ baseline", and a peak capacity of log₂ 137 to
within 0.05%. And the Coco reads the substrate's whole catch comb on the baseline
instrument, run out to 131072 laps, with the epistemic '1' at lap 137 as the thirteenth
catch and the origin of its cooling ladder.

So the three numbers pinned here — the onset at 137, the budget of 1024 laps, the
four-decimal grain — are not the Quad Slip's alone. They are the ground the back half of
the series stands on, and the reason the series result is a setting rather than a
finding: a place every later reading can be brought back to and compared on.

## Where it sits

The third instrument of the series of ten and the result of the slip → double slip →
quad slip line. On the observatory node the door is
[opendata.ai/quad-slip](https://opendata.ai/quad-slip), state *live*: the instrument
runs on request, up to 4096 laps, as the platform singleton, and the page's baseline
button reproduces the reference row regardless of what the composer holds; the Quad
Slip predates the mldata anchoring practice and carries no external anchor.

## Standing

Held on this node as fact until someone stands up in the record to refute it; a
refutation enters as a statement of intent at [the matters of note](/sx/matters) and
goes to the record's own path. What has since been built on the Quad Slip, and the walk
a reader may take, are the operator's hand and land in a further draft.