The chapter walked two lists, every channel the sim streams and every control the garage offers, and traced each item to the patch it serves. This page is that chapter's lesson: the same ground one layer deeper, the why behind each claim it made, and a short test at the end. The numbers here are the chapter's numbers, from the session cited on that page.
Why counting channels is evidence
An inventory cannot prove physics. The physics case for the four patches, grip as the only force that ever turns, slows, or speeds a car, was made in the introduction and does not need the channel list. What the inventory tests is the claim's reach: if everything on the car really routes through the patches, then nothing the simulator streams and nothing the garage adjusts should resist the trace, and the test only means something if every item is on the table. That is why the audit matters more than the count. An inventory that silently drops what it does not recognize would look complete while proving nothing; this one accounts for every channel by name and fails loudly on anything new, a discipline this project adopted after a family of per-corner channels sat unnoticed in the stream for months. A clean audit is what upgrades "here is what I found" to "here is everything there is."
What the budget itself says is narrower but still worth having: eleven channels per tire, as many as the entire engine gets, says what the simulator's builders chose to watch most closely. That is a fact about the instrument, not about nature. The two facts point the same way, and only one of them is physics.
The trace question, worked on three channels
The chapter sorted every channel into one of three kinds. Here is one of each, worked all the way through.
A channel that measures a patch directly. The left front's middle carcass temperature is rubber state at the print: how hot the tire's core is at its centerline. Nothing stands between this number and the patch; it is the patch, reported in Celsius.
A channel on its way to a patch. The left front brake line pressure is force in transit. Pedal force becomes hydraulic pressure, pressure becomes clamping force at the disc, clamping becomes a torque on the wheel, and the torque becomes rearward force at that one patch. Every link in that chain exists to deliver force to the rubber, so a squiggle in this channel is evidence of what the patch was asked to do, not yet of what it did.
A channel that measures a consequence. Yaw rate is how fast the body is rotating. No part of the car can rotate the body except the patches making a net moment about the center of mass, so this channel is the patches' work after the fact, summed. Cause channels and effect channels bracket the patch from both sides, and knowing which side a trace sits on decides what it can be evidence of.
The trace question, worked on one control
The chapter's garage table gave the front anti-roll bar one line. Here is the full chain behind it. In a corner the body rolls, the outside suspension compresses and the inside extends, and the bar, a torsion spring tying the two front wheels together, resists exactly that difference. Stiffen it and the front pair reacts a larger share of the car's total roll moment, which means more of the cornering load transfer happens across the front pair and less across the rear (Smith, Tune to Win, p. 36 to 37: stiffening a bar both reduces roll angle and increases lateral load transfer at its end). The introduction's bent load curve then prices the split: a pair carrying a bigger imbalance returns less combined grip, so the front bar is a lever that trades front grip against rear. That is the mechanism. For this car it stays a hypothesis until the setup experiments change one bar step and read the patches, which is exactly what the A/B chapters do.
The mechanisms behind the one-line rows
The garage table compressed each control family to a sentence. The sentences are hypotheses, and these are the mechanisms they compress.
Dampers time the load; they never set its amount. How much load transfers in a steady corner is fixed by geometry and mass; the dampers decide how fast it arrives and settles while the car is still pitching or rolling (Smith p. 74 to 76: shocks change the rate of transfer, not the amount). That is why the chapter's row says "the rate of transfer, never the amount," and why chapter 24 will look for a damper click in the data's transients, not in its steady states.
Toe is an angle the car runs before you steer. Toe settings point the wheels slightly in or out of parallel, so each tire travels at a small standing angle to its own direction of motion even on a straight. Toe-in makes an upset self-correcting: load shifting onto one wheel steers that wheel back toward straight; toe-out does the opposite, quicker to turn and nervous in a straight line (Smith p. 60). The angle between where a tire points and where it travels is the subject chapter 11 measures.
The differential decides which rear patch gets the torque. An open differential on corner exit feeds torque to the unloaded inside tire until it spins, and then all of it goes there (Smith p. 148). Locking the pair together trades that failure for coupled rotation: both rear patches turn together whether their loads match or not. The garage's ramp angles, friction faces, and preload are three ways of setting how strongly, and chapter 33 reads the result wheel by wheel.
Fuel is the load that changes itself. Every other control holds its value until the garage; fuel is mass on all four patches that falls every lap, moving the car's weight and its balance all stint without anyone touching a setting. That is arithmetic, not a hypothesis, and the stint chapters read it directly.
Why lap time is the last thing to watch
The chapter said a control's first checkpoint is the patch it serves, not the stopwatch. The reason is the length of the chain: control to patch to force to acceleration to speed to lap time. Each link adds noise from somewhere else, track temperature, fuel state, traffic, the driver's own variation, so by the end of the chain a real change can drown and a phantom change can appear. Reading the patch is reading one link from the cause. Reading the lap time is reading through the whole chain plus the whole day. The course's rule of changing one control at a time and reading the patches first is this chain, taken seriously.
Run it yourself
The chapter's experiment transfers to any sim that records data. Take one recording and list every channel it contains, with its unit, and sort the list by what each channel watches; a spreadsheet is enough. Then open your garage and restate every control as which patch it serves and by what mechanism. Anything that resists either list is not a failure, it is your open question, and it is worth keeping in print. Our toolchain runs the channel half as an audit that refuses to pass anything unclassified, but the method needs no tooling, only the one question asked of every item.
Sources
- The chapter this lesson deepens: [Everything routes through four
contact patches](/posts/2026-07-19-four-contact-patches), including the channel counts restated here, from telemetry session fc1c2bc75761511d cited on that page.
- The introduction and its lesson, for grip as a force and the bent
load curve: Grip Leaves a Fingerprint.
- Smith, Tune to Win: p. 36 to 37 (anti-roll bar stiffness, roll
angle, and lateral load transfer at its end), p. 60 (toe-in and toe-out response to upsets), p. 74 to 76 (dampers change the rate of load transfer, not the amount), p. 148 (open differential behavior on corner exit).