What Changed Between Two Comparable Logs?

1. The Question

Does the revised airbox change the intake-air temperature trend during a fixed commute segment?

The owner wanted to know whether a hardware change affected intake temperatures. Rather than argue from memory, they recorded the same commute segment twice — once before, once after — and controlled the variables they could.

The design was deliberately boring, which is what made it strong. One variable, one route, one marked segment — nothing fancy, nothing left to argue about. The question was narrow enough that a clean answer was possible.

Most failed comparisons are designed after the fact: two existing logs forced to serve a question they were never recorded for. This one started with the question and built both sessions around it.

2. The Vehicle State

2020 turbo wagon, 56,000 km. Tyre pressures, fuel level, and payload were matched between sessions; the airbox change was the only deliberate hardware difference.

A deliberate single-variable change is what makes the comparison a comparison.

Matching vehicle state across sessions closed the usual escape routes. Tyre pressures, fuel level, and payload were noted as identical; the airbox was the only deliberate difference between the two files.

That single-variable discipline is what gives the observed delta its meaning. Had two things changed at once, the same data would have supported two stories and proven neither.

3. The Conditions

Both sessions ran the same route at the same time of day, a week apart, with ambient temperatures within 2°C. Warmup procedure was identical: ten minutes of mixed driving before the marked segment.

The documented conditions are what allowed the word 'comparable' to be used honestly.

The two-degree ambient gap is what 'comparable' actually looks like — not identical, just close enough that the residual is noise rather than signal. Documenting it is what lets a reviewer check the claim instead of trusting it.

Warmup matching mattered as much: both runs logged the same ten-minute lead-in, so the marked segment began from the same thermal state rather than from wherever each drive happened to be.

4. The Recorded Window

Each log carried a marked four-minute segment over the same road section, plus the unmarked approach for warmup verification.

Same road, same clock position, same recorded window — by design, not by luck.

The four-minute marked segment covered identical landmarks in identical direction — the windows were engineered to overlap. Whatever differed inside them could not be blamed on 'different part of the drive.'

Keeping the unmarked approach in the file was a quiet piece of good practice: it let the reviewer verify warmup state instead of taking it on faith. Evidence about the evidence.

5. The Observation

Within the marked windows the intake-temperature traces diverged by a small but consistent margin. The delta was real, modest, and — critically — confined to the variable that had actually been changed.

Nothing else in the logs moved. That is what made the small delta readable.

The delta was small, consistent, and confined — exactly the signature a real single-variable effect leaves. Nothing else in the two logs moved, which is the strongest support a modest difference can get.

The conclusion stays proportionate: 'the airbox change produced a small, repeatable shift in intake-temperature trend under these conditions.' Not a verdict for all conditions — a finding inside the tested frame.

6. The Follow-Up Question

A third session in hotter ambient conditions will test whether the observed margin scales with heat load — the next question the comparison raised.

Each comparable pair answers one question and writes the next.

The natural next question wrote itself: does the margin scale with heat load? A third session at higher ambient tests the interaction the first two could not reach — each comparable pair generating its own successor.

That chain — question, comparable pair, bounded answer, next question — is the rhythm this publication exists to document. One good comparison is worth more than a folder of unrelatable logs.