The Waterfall Manual: A Mathematical Defense of Every Line on the Surface
Published on: March 5, 2026
The Trust Debt Calculator is not a chart. It is not a dashboard. It is not a data visualization built from a sample, trained on a corpus, or fitted to a curve. It is a closed-form audit instrument. Every pixel on its surface is computed from a single formula with zero free parameters:
Phi = (c/t)exponent
That is the entire model. A fraction raised to a power. No regression coefficients. No hyperparameters. No training data. No curve fitting. No statistical estimation. Pure algebra applied to three physically measurable quantities: what you cover (c), what exists (t), and how many independent dimensions you operate across (the exponent).
This post defends every line on that surface. Every color, every contour, every zone boundary, every camera angle, every axis label. Element by element, formula by formula. If any element on the calculator is wrong, the underlying mathematics is wrong. The mathematics is not wrong.
The claim: This is a universal surface. Any system that filters information through independent dimensions lives somewhere on it. Your position is computable. Your zone has a name. Your Trust Debt has a cost. This post proves why.
Open the calculator now. This post explains what you are looking at.
The core formula is Phi = (c/t)N in Mirror 1, or (c/t)n in Mirror 2. A fraction between 0 and 1, raised to a positive integer power. The result is the height of the surface at that point.
Why a power law? Because dimensional intersection is multiplicative. When you pass information through N independent filters, each filter independently accepts a fraction c/t of the input. The probability that a signal passes through ALL N filters is the product of the individual pass rates. For independent filters with identical selectivity, that product is (c/t) multiplied by itself N times: (c/t)N.
This is not a modeling choice. It is a mathematical identity. If the filters are independent (orthogonal), the joint probability is the product. If they are not independent, they are not separate dimensions. There is no free parameter to tune. The formula IS the physics.
Why not exponential? An exponential decay like e-kN requires a decay constant k. Where does k come from? It must be fitted to data. The power law (c/t)N needs no k. The selectivity ratio c/t IS the base, measured directly.
Why not sigmoid? A sigmoid requires a midpoint and a steepness parameter. Both must be estimated. The power law's "midpoint" is the Golden Hinge at exactly 0.618, derived from number theory. Its steepness is set by c/t, which you measure.
Why not logarithmic? A logarithm grows without bound. Phi must live between 0 and 1 because it represents a fraction of surviving noise (or signal). The power law naturally maps the unit interval to the unit interval: if 0 is less than c/t and c/t is less than 1, then 0 is less than (c/t)N and (c/t)N is less than 1 for all positive N. No clamping needed.
The surface is rendered on a 60 by 60 grid, with the x-axis (c/t) ranging from 0.01 to 0.99 and the y-axis (exponent) from 1 to 30. Every one of those 3,600 points is computed by the same formula. No interpolation. No smoothing. No approximation. Open the calculator and drag the sliders. Every point on this surface is z = xy.
For the auditor: A formula with zero free parameters cannot be overfit. It cannot memorize noise. It cannot be gamed by choosing a favorable training set. The surface you see IS the formula. If you disagree with the surface, you must disagree with the claim that independent filters multiply. That claim is a theorem, not an assumption.
Three variables. Three axes. All physically measurable.
X-axis: c/t (selectivity). The variable c is what you deeply cover, and t is the total domain. The ratio c/t is your selectivity. When c/t = 0.01, you focus on 1% of reality with serious depth. When c/t = 0.99, you try to cover nearly everything. The ratio is bounded between 0 and 1 because you cannot focus on more than exists.
Y-axis in Mirror 1: N = nData + nVerify + nTime. This is the number of orthogonal grounding dimensions. Each dimension must be genuinely independent. Three data sources that all cite each other are not three dimensions; they are one dimension with three labels. A radiologist using anatomy, imaging physics, patient history, and clinical experience has N = 4 if each source provides non-redundant information. Correlated dimensions do not count. This is the hardest part of the measurement: being honest about orthogonality.
Y-axis in Mirror 2: n = nSteps + nSynth. This is the number of sequential hops. Each reasoning step (nSteps) is an ungrounded transmission, and each synthesis layer (nSynth) is an abstraction hop. A 10-step reasoning chain with 2 layers of synthesis is n = 12. A 2-step chain with 10 layers is also n = 12. Same exponent, same Phi, same position on the surface. The formula does not care how the hops are distributed; it cares how many there are.
Z-axis: Phi. In Mirror 1, Phi is the remaining noise after dimensional filtering. Low Phi means the noise has been crushed; you are grounded. In Mirror 2, Phi is the surviving signal after sequential drift. Low Phi means the signal has been destroyed; you are in the Waterfall. Same formula, opposite physics. The Z-axis flips meaning because the exponent flips from "independent filters" to "serial degradation." The algebra is identical; the interpretation is a mirror.
Why seven sliders for three variables? Because CFOs do not think in abstract N. They think in "data sources," "verification methods," and "temporal consistency." The calculator provides seven sliders that translate boardroom language into mathematical coordinates. The sliders are a user interface, not a mathematical necessity. Under the hood, only three numbers matter: the selectivity ratio, the exponent, and the mirror.
For the enterprise architect: Map your AI pipeline to these three numbers. c = what your model deeply knows. t = the total domain it claims to cover. N = how many independent verification dimensions you apply. If you cannot count your dimensions, you cannot compute your Trust Debt. The calculator does the rest.
See the axes in action: c = 1, t = 100, giving c/t = 0.01 with N = 10 grounding dimensions. Phi = (0.01)10 = 10-20. That is the Floor.
The red contour line drawn across the surface sits at exactly Phi = 0.618. Not 0.6. Not 0.62. Not a round number chosen for convenience. Exactly (sqrt(5) - 1) / 2, the reciprocal of the Golden Ratio phi.
Why this number? Because 0.618 is the KAM resonance boundary from dynamical systems theory. The Kolmogorov-Arnold-Moser theorem proves that in Hamiltonian systems, orbits with frequencies whose ratio is "most irrational" survive the longest under perturbation. The most irrational number, in terms of resistance to rational approximation, is the Golden Ratio. Its reciprocal, 0.618, defines the boundary between structural stability and chaos.
Below Phi = 0.618, perturbations die out. The system absorbs disturbances and returns to its orbit. Above 0.618, perturbations amplify. The system leaves its basin of attraction and enters chaos. This is not a metaphor. It is a proven theorem of nonlinear dynamics, published independently by Kolmogorov (1954), Arnold (1963), and Moser (1962).
In the waterfall, Phi = 0.618 is the exact coordinate where surviving signal equals accumulated noise. In Mirror 1, it is where the filtered signal loses its structural majority to remaining noise. In Mirror 2, it is where the degraded signal can no longer be distinguished from the drift. On both sides of the mirror, the same number marks the same phase boundary. The warm white seam at 0.618 is the knife edge between order and chaos.
The 160-hop event horizon. At biological fidelity, each synaptic transmission preserves 99.7% of the signal. The decay formula gives (0.997)160 = 0.618. At exactly 160 hops, a biological-fidelity transmission chain crosses the Golden Hinge. This is why anesthesia works at the concentrations it does, why neural pathways degrade at the depths they do, and why the waterfall has a floor. See Chapter 0: The Razor's Edge for the full anesthesia derivation.
The color scales on both mirrors are designed around this boundary. The warm white seam in CSCALE_M1 sits between 0.610 and 0.626, only 1.6% of the total scale. The identical seam in CSCALE_M2 sits at the same coordinates. Everything below is purple or green (safe territory). Everything above is orange or red (danger). The phase transition is rendered as a knife edge because it IS a knife edge. There is no gradual crossover. The math is discontinuous in its second derivative at exactly 0.618.
The hinge is also drawn as a translucent horizontal plane across the entire 3D space at z = 0.618, rendered in rgba(254,249,195,0.06). You can barely see it. That is deliberate. The hinge is substrate, not decoration. It exists whether you see it or not, the way the speed of light exists whether you measure it or not.
Appendix R.8: Resonance ThresholdThe yellow dashed curve on the surface traces the formula n = 3 times sqrt(1 / (c/t)), equivalently n = 3 times sqrt(t/c). This is the minimum grounding boundary.
Where does it come from? From the sqrt(2) Law, proved as Theorem 3 in the Skip Formula post. At the knee of the power curve (c/t)n, where the surface bends from steep to flat, the maximum curvature occurs at the point where the filtering threshold scales as 1 / (n times sqrt(2)). Rearranging: for a given selectivity c/t, the minimum number of dimensions needed to reach the knee is proportional to sqrt(t/c).
Below this curve, you do not have enough dimensions to trigger the phase transition. The surface is still steep, the filtering is not yet engaged, and the noise passes through largely unaffected. Above the curve, the crusher engages. Each additional dimension now delivers the full multiplicative punch of (c/t) per dimension. The returns do not diminish, as we proved in the Skip Formula derivation. They compound.
Why the line climbs the surface. Look at the calculator. The red Golden Hinge is a smooth horizontal contour: constant z = 0.618, slicing the surface like a water line. The yellow sqrt(2) line is nothing like that. It hugs the Floor on the left, then climbs steeply toward the Wall on the right. This is not a rendering artifact. It is the deepest result in the entire visualization.
The Hinge is a constant-z contour. It answers: "at what noise level does chaos begin?" The sqrt(2) line is a path of minimum sufficiency where z is NOT constant. At each point along the curve, both x and y change, and z = xy is recomputed from the surface. The z-value swings across the entire range:
At c/t = 0.1 (the specialist): y = 3 times sqrt(10) = 9.5 dimensions. Phi = (0.1)9.5 = 3 times 10-10. The line sits near the Floor. Even the bare minimum grounding crushes noise when your focus is narrow.
At c/t = 0.5 (the balanced operator): y = 3 times sqrt(2) = 4.2 dimensions. Phi = (0.5)4.2 = 0.054. The line passes through Stable Ground. Minimum grounding works, but with slim margin.
At c/t = 0.9 (the generalist): y = 3 times sqrt(1/0.9) = 3.16 dimensions. Phi = (0.9)3.16 = 0.73. The line has climbed past the Golden Hinge into Chaos Territory. Even at minimum grounding, the generalist cannot escape chaos. There are not enough dimensions in the universe to save broad coverage from the power law.
This is the visual punchline of the entire surface. The sqrt(2) line climbs because broad coverage is exponentially harder to ground. The specialist (low c/t) gets the Floor for free. The generalist (high c/t) cannot buy the Floor at any price below the sqrt(2) threshold. The climb from Floor to Wall along the yellow line is the cost of claiming to know everything. Read the line left to right. That ascent is Trust Debt made visible.
The floor projection (the yellow dotted shadow at z = 0) strips away the z-axis and shows the pure engineering spec: how many dimensions do you need at each selectivity? That 2D shadow is the planning tool. The 3D curve that climbs the surface is the consequence.
Why the coefficient is 3. The theoretical minimum from Theorem 3 is sqrt(2) times sqrt(t/c). The calculator uses a coefficient of 3, which provides approximately a 2x safety margin above the absolute theoretical limit. This is an engineering decision, not a mathematical one. In practice, not all of your "orthogonal" dimensions will be perfectly independent. Some correlation leaks in. The factor of 3 absorbs that leakage and tells you where the knee is with real-world imperfection baked in.
The floor projection of this curve is a yellow dotted line at opacity 0.3, showing where the sqrt(2) boundary would sit if flattened to 2D. This is the engineering spec: at a given selectivity c/t, how many dimensions do you need before the filtering kicks in? Read the projection. That is your minimum.
For the system designer: To cross the sqrt(2) line, add orthogonal dimensions. At c/t = 0.1, you need at least 10 dimensions (3 times sqrt(10) = 9.5, rounded up). At c/t = 0.5, you need at least 5 dimensions (3 times sqrt(2) = 4.2, rounded up). The narrower your focus, the fewer dimensions you need. This is why specialists outperform generalists on the surface: their c/t is small, so fewer dimensions push them past the knee.
See the line in action: c/t = 0.1, N = 10, just above the yellow dashed curve. Now drop below it: N = 9, just below. Watch the zone change.
Six zones per mirror. Not arbitrary categories painted onto the surface for aesthetics. Each threshold corresponds to an order-of-magnitude regime change in the underlying physics.
Mirror 1: Grounding by Dimensions (noise crushing)
THE FLOOR (Phi is less than 0.001). Noise is 0.1% of the search space or less. The dimensional filtering is definitive. At this level, the remaining noise is below the measurement threshold of any practical instrument. You are grounded. This is where your brain lives when tasting salt: (0.01)10 = 10-20, twenty orders of magnitude below the Floor boundary.
DEEP FLOOR (0.001 to 0.01). Noise between 0.1% and 1%. Strong filtering, but not absolute. A well-instrumented scientific measurement lives here. A radiologist with four orthogonal data sources lives here. Precision is high; certainty is structural.
STABLE GROUND (0.01 to 0.1). Noise between 1% and 10%. Solid filtering. The engineering margin is adequate for most operational decisions. Most well-designed verification systems land here. The sqrt(2) line typically runs through this zone, marking the transition from "not yet enough dimensions" to "enough."
TRANSITION ZONE (0.1 to 0.618). Between order and chaos. The filtering is working but has not yet crossed the Golden Hinge. Perturbations neither grow nor die reliably. This is the most dangerous zone for decision-makers because it feels stable while being structurally fragile. One removed dimension can push you past the Hinge into chaos.
CHAOS TERRITORY (0.618 to 0.9). Past the Golden Hinge. Noise dominates the filtered signal. Perturbations amplify. Most enterprise AI systems live here because they operate with c/t greater than 0.5 and N equal to 1 or 2. They have broad coverage, shallow grounding, and no idea that the surface under them has already broken.
THE WALL (Phi of 0.9 or higher). Ninety percent or more of the search space passes through unfiltered. No meaningful filtering is occurring. The system is returning noise with a confident label. This is where hallucination lives: high c/t, low N, a single dimension pretending to be general intelligence.
Mirror 2: Drift by Hops (signal crushing) uses the same thresholds with inverted meaning. THE WATERFALL (Phi is less than 0.001) means the signal is dead, less than 0.1% surviving. CRITICAL LOSS (0.001 to 0.01), HEAVY DRIFT (0.01 to 0.1), DRIFT ZONE (0.1 to 0.618), LIGHT DRIFT (0.618 to 0.9), and FULL SIGNAL (0.9 or higher) trace the same orders-of-magnitude boundaries, but now low Phi is catastrophic (signal destroyed) and high Phi is safe (signal intact).
Why powers of 10? Because each 10x reduction in Phi requires a specific, calculable number of additional dimensions. From Phi = 0.1 to Phi = 0.01 requires exactly log(10) / log(1/(c/t)) additional dimensions at any given c/t. The zone boundaries are not aesthetic choices. They are physical thresholds where the number of required dimensions increments by an integer.
Why the same thresholds in both mirrors? Because the formula is algebraically identical. (c/t)N in Mirror 1 and (c/t)n in Mirror 2 produce the same output space [0, 1]. The physics flips, but the algebra does not. This is the "mirror": same structure, opposite interpretation. A Phi of 0.001 is paradise in Mirror 1 (noise crushed) and catastrophe in Mirror 2 (signal destroyed). The math does not care which way you are looking. It computes the same number.
Visit the Floor. Then visit the Wall. Same surface. Different coordinates. The zone you land in is your Trust Debt.
The white diamond on the surface is your dot. Its coordinates are computed as follows:
x = c / max(t, c + 1). The max function prevents the nonsensical case where c exceeds t. You cannot focus on more categories than exist. If you set c = 80 and t = 50, the denominator becomes 81, clamping c/t to approximately 0.988 rather than allowing 80/50 = 1.6. This is not a hack; it is a boundary condition. The domain of (c/t)N requires 0 is less than c/t and c/t is less than 1.
y = N or n. In Mirror 1, the exponent is N = nData + nVerify + nTime, clamped to a minimum of 1 (you always have at least one dimension). In Mirror 2, it is n = nSteps + nSynth, also clamped to a minimum of 1 (you always have at least one hop).
z = xy. Your Phi. The height of the dot on the surface. This is the single number that determines your zone, your color, and your Trust Debt.
The x-range clamp: c/t is clamped to [0.01, 0.99]. At x = 0, the power law gives 0N = 0 for all N, which is trivially true (zero selectivity means zero signal) but uninformative. At x = 1, the power law gives 1N = 1 for all N, which is equally uninformative (total coverage means total pass-through). The interesting physics lives between these degenerate endpoints. The clamp excludes them.
The drop line is a dashed white vertical segment from your dot straight down to z = 0. This is the trust debt column. Its length is exactly Phi: the vertical distance between your position and the Floor. If your dot is at Phi = 0.73, the drop line spans 73% of the z-axis. That visual height IS your noise fraction (Mirror 1) or your signal loss (Mirror 2).
The floor shadow is a small circle at (x, y, 0), colored with your zone's color. This is where you would be if you could eliminate all noise: your grounded projection. The distance between the shadow and the dot is what the waterfall measures.
The hover text reads "YOU ARE HERE." Deliberately provocative. The calculator is personal. It is not an abstract model of information theory. It is a mirror that reflects your system's position on a universal surface. When you see "YOU ARE HERE," the claim is literal: these are your coordinates.
For the governance team: The drop line height is your audit metric. Capture it. Report it. Track it over time. If the drop line is growing, your Trust Debt is compounding. If it is shrinking, your grounding investments are paying off. The calculator gives you the number. The drop line gives you the visual.
The toggle at the top of the calculator switches between Mirror 1 and Mirror 2. Both mirrors compute the same formula. The difference is what the exponent represents and what Phi means.
Mirror 1: Grounding by Dimensions. The exponent is N = nData + nVerify + nTime. Three categories of grounding. Independent data sources contribute dimensions that filter by cross-referencing. Verification methods contribute dimensions that filter by independent confirmation. Temporal consistency contributes dimensions that filter by persistence across time. The math treats all three identically because a dimension is a dimension. But the UI separates them because CFOs, auditors, and architects distinguish between "we have multiple data sources" and "we have verification protocols" and "we check whether the answer changes over time." These are operationally different investments even though they are mathematically identical.
Mirror 2: Drift by Hops. The exponent is n = nSteps + nSynth. Two categories of drift. Reasoning steps (nSteps) are sequential inference hops: each link in a chain-of-thought is one hop. Synthesis layers (nSynth) measure abstraction depth: summarizing a summary is two hops, not one. These are different kinds of degradation. A 10-step reasoning chain is shallow but long. Two layers of deep synthesis are short but vertically compressed. Both accumulate drift, but differently. The distinction matters operationally even though the formula treats them identically.
The color flip. Mirror 1 goes purple to warm white to red: starting from deep grounding (cool, structural purple) through the Golden Hinge (warm white) to the Wall (hot, urgent red). Noise is the protagonist; the colors track its dominance. Mirror 2 goes dark red to warm white to green: starting from the Waterfall (signal dead, dark red) through the Hinge (warm white) to Full Signal (alive, green). Signal is the protagonist; the colors track its survival. The hinge is warm white in BOTH mirrors because the phase boundary looks the same from either side. It does not matter whether you are watching noise rise or signal fall. At 0.618, the transition happens.
The corner annotations flip. Mirror 1 labels the low-Phi corner "FLOOR" (noise crushed) and the high-Phi corner "WALL" (noise dominates). Mirror 2 labels the low-Phi corner "WATERFALL" (signal dead) and the high-Phi corner "FULL SIGNAL" (signal intact). Same geometry, opposite vocabulary.
The accent color changes. Mirror 1 uses violet-500 as its accent (the purple of grounding). Mirror 2 uses red-500 (the urgency of drift). This is a visual cue so you always know which mirror you are looking through, even at a glance.
The trap: Most teams only look at Mirror 1. They count their data sources, feel safe, and never toggle. Mirror 2 asks the harder question: how many hops does your inference chain take? If you have 8 grounding dimensions but your pipeline runs through 20 synthesis hops, your Mirror 2 position may be catastrophic even while Mirror 1 looks safe. Always check both mirrors. The calculator makes it one click.
See the Mirror 1 view and then the Mirror 2 view of the same system. Same c, same t. Different exponent. Different zone. Different Trust Debt.
Appendix R.5: The Mirror of ExponentiationThe surface is not painted with arbitrary colors. Each color scale is a 10-stop gradient engineered around the phase transition at 0.618.
CSCALE_M1 (Mirror 1: noise perspective). Deep purple (#0c0020) at Phi = 0 fades through structural purple (#3b0764) at 0.30 and violet (#7c3aed) at 0.50, approaching light lavender (#c4b5fd) at 0.610. At exactly 0.618, the scale hits warm white (#fef9c3). Then orange (#fb923c) at 0.626, deeper orange (#ea580c) at 0.72, red (#dc2626) at 0.88, and dark red (#7f1d1d) at 1.0. Purple is grounded: cool, calm, structural. Red is chaotic: hot, urgent, warning. The transition between them is not gradual. The warm white seam occupies only 1.6% of the total scale, from 0.610 to 0.626. This is a knife edge, not a gradient. The phase transition is rendered as narrow as the color resolution allows.
CSCALE_M2 (Mirror 2: signal perspective). Dark maroon (#450a0a) at Phi = 0, through red (#991b1b) at 0.10, bright red (#dc2626) at 0.28, orange (#f97316) at 0.45, cream (#fef3c7) at 0.610. At 0.618: the identical warm white (#fef9c3). Then light green (#86efac) at 0.626, strong green (#22c55e) at 0.75, deep green (#16a34a) at 0.90, and darkest green (#052e16) at 1.0. Red means signal dead. Green means signal alive. The biological metaphor is intentional: green is growth, red is hemorrhage. The warm white seam is identical to Mirror 1. The phase boundary does not care which perspective you bring.
The translucent hinge plane. A horizontal surface at z = 0.618, rendered in rgba(254,249,195,0.06). An opacity of 6%. You can barely perceive it. The hinge plane is not a visual element to be noticed. It is a structural element that exists. Like the gravitational constant, it operates whether you see it or not. The low opacity ensures it does not occlude the main surface while remaining faintly visible when you rotate the view to look along the z-axis.
The user slice curve. A white line at 0.3 opacity traces the 2D cross-section of the 3D surface at your current c/t value. Hold selectivity constant and this curve shows how Phi changes as you vary the exponent. It is the formula (c/t)y for fixed c/t and varying y, plotted directly on the surface. This is what the power law looks like when you fix the base and sweep the exponent: a smooth, monotonic decline from (c/t)1 toward zero.
For the designer: The 1.6% width of the warm white seam is not a design choice. It is a rendering decision dictated by the math. The phase transition occupies zero width in the limit (a single point at 0.618). The 1.6% is the minimum width that renders as a visible band on a 10-stop color scale. Any wider would misrepresent the sharpness of the transition. Any narrower would be invisible.
The default camera position is eye = (-1.5, 1.8, 0.5) with an up vector of (0, 0, 1). This is not a neutral angle. It is a design decision with mathematical intent.
Why this position? The camera looks slightly down from the upper left. The Floor is in the foreground. The Wall is behind. To see the Wall, the chaotic high-Phi region where noise dominates, you must actively rotate the view. This is deliberate: the default perspective shows you what grounding achieves, not what chaos looks like. The goal is the Floor. The camera faces it.
The up vector is locked to (0, 0, 1). The z-axis is always vertical. This prevents disorienting rotations where the surface appears to flip upside down. Trust has a direction: down is grounded, up is noise. The metaphor is gravitational. The Floor is below. The Wall is above. The Hinge is a horizontal plane at a specific altitude. When you rotate the view, the vertical axis stays fixed because the meaning of "down" does not change when you shift your perspective.
Camera persistence. When you move the sliders, the surface updates but the camera stays where you put it. This is critical for exploration. If you have rotated to examine a particular region of the surface, moving a slider should update the surface underneath your viewpoint, not teleport your perspective back to the default. The system does not steal your angle. Your perspective is yours.
Why the floor matters. The rendered floor at z = 0 is the coordinate plane where Phi = 0. No system actually reaches Phi = 0 unless c/t = 0, but the floor represents the limit. The floor projection (your dot's shadow, the hinge projection, the sqrt(2) projection) all sit on this plane. It is the reference surface against which all Trust Debt is measured. Your drop line falls to it. Your shadow sits on it. The camera faces it because that is where you want to be.
Every slider state on the calculator can be encoded in a URL.
The parameter schema: ?c=55&t=100&nd=1&nv=1&nt=1&ns=4&ny=2&m=1&label=Your+Label. Nine parameters. c is the focus context (1 to 200). t is the total domain (10 to 1000). nd is independent data sources (0 to 10). nv is verification methods (0 to 10). nt is temporal consistency (0 to 10). ns is reasoning steps (1 to 15). ny is synthesis layers (0 to 15). m is the mirror (1 for grounding, 2 for drift). label is an optional human-readable name.
A preset IS a coordinate. When you send someone a preset link, you are not sending them a configuration file or an API call or a database query. You are sending them a position on the surface. The URL encodes the exact values of c, t, and the exponent components. The calculator reconstructs the position deterministically. Two people with the same link see the same dot on the same surface in the same zone. This is how audit results become shareable: "here is where our system lives" is a URL.
Why URL parameters and not a database? Because presets must be self-contained. No server dependency. No authentication. No API call. No session state. The URL IS the data. If thetadriven.com disappears tomorrow, the URL still encodes the mathematics. Anyone who knows the formula can reconstruct the position from the parameters. The preset system is serverless by design, not by limitation.
The Copy Link button generates the full URL with the current slider state and label. Governance teams use this to share audit snapshots. Architects use it to embed specific positions in documentation. Researchers use it to attach exact coordinates to claims. Every link in this post is a preset. Every one is verifiable by clicking it.
The label system converts mathematical coordinates into organizational language. "Your Brain Tasting Salt" is more useful in a boardroom than "c=1, t=100, N=10." The label appears in the header bar above the surface. It is metadata, not mathematics, but it bridges the gap between the formula and the human who needs to act on it.
For governance and audit teams: Build a library of preset URLs for your AI systems. One link per pipeline, one link per model deployment, one link per risk assessment. Each link is an immutable coordinate. Compare them over time. When a system moves from Stable Ground to Transition Zone, the preset URL documents exactly when and how. The calculator is your audit trail.
Every element has been defended.
Surface: z = (c/t)N. A power law from the multiplicative composition of independent filters. Zero free parameters.
Axes: c/t is selectivity (physically measurable). N is orthogonal grounding dimensions (countable). n is sequential hops (countable). Phi is the output of the formula, bounded between 0 and 1.
Golden Hinge: Phi = 0.618 = 1/phi, the KAM resonance boundary. Below it, perturbations die. Above it, perturbations amplify. Proved by Kolmogorov, Arnold, and Moser. Confirmed at (0.997)160 = 0.618 in biological systems.
sqrt(2) Line: n = 3 times sqrt(t/c), the minimum grounding boundary from Theorem 3 of the Skip Formula proof. Below it, insufficient dimensions. Above it, the crusher engages.
Zones: Six per mirror, at powers of 10 (0.001, 0.01, 0.1) plus the Golden Hinge (0.618) plus 0.9. Each threshold marks an order-of-magnitude regime change in filtering. Same thresholds in both mirrors. Same algebra, opposite physics.
Colors: 10-stop gradients designed around a 1.6%-width warm white seam at exactly 0.618. Purple/green = safe. Red/orange = danger. The phase transition is a knife edge because the mathematics is discontinuous in its second derivative at the Hinge.
Diamond: Your dot at (c/max(t, c+1), N or n, (c/t)exponent). Clamped to the valid domain. The drop line IS your Trust Debt, measured in Phi.
Camera: Faces the Floor because trust points down. Up vector locked to z because "grounded" has a direction.
Presets: Self-contained URL-encoded coordinates. No server. No state. No dependency. The URL IS the audit.
The waterfall is not a visualization. It is a calculator. It takes your inputs (c, t, N, n), applies a closed-form formula with zero free parameters, and returns your position on a universal surface. That position has a zone. That zone has a name. That name has a cost. That cost is your Trust Debt. The formula does not negotiate.
The reading sequence. This post is the technical appendix. For the proof that the sqrt(2) law holds and that linear improvement never diminishes, read the Skip Formula post. For the five independent derivations of the decay constant k_E = 0.003, read the k_E derivation. For the smear between Mirror 1 and Mirror 2, read The Smear Is the Trick. For the zone boundaries and what they predict about superintelligence, read the Zone Boundary debate. For the product form decomposition of the Anchor and the Crusher, see Chapter 8: Conclusion. For the anesthesia derivation that pins k_E to biology, see Chapter 0: The Razor's Edge.
Open the Trust Debt Calculator. Find your position. Name your zone. Measure your debt. Then decide what to do about it.
Fire Together. Ground Together.
Ready for your "Oh" moment?
Ready to accelerate your breakthrough? Send yourself an Un-Robocall™ • Get transcript when logged in
Send Strategic Nudge (30 seconds)