The Macro Engine's Cycle Clock sorts the business cycle into eight fixed phases, Goldilocks through Recovery, assigned by a decision tree rather than a continuous score. On 2026-09-30 it read Overheat, phase three of the eight, late cycle with growth peaking and policy tightening.
The Cycle Clock is one of the Macro Engine's modules: a business-cycle classifier that reads the rates regime, credit and volatility conditions, and curve shape, and matches that combination against a fixed decision tree until one of eight phases fits. It is not a continuous score sliding along a scale. A reading is always exactly one of eight discrete labels, never a blend of two.
The eight phases, by their regime index in the classifier's own code: 1 Goldilocks, 2 Reflation, 3 Overheat, 4 Stagflation, 5 Slowdown, 6 Contraction, 7 Crisis, 8 Recovery.
The classifier does not simply pick whichever phase looks closest. It works through the possibilities in a fixed order and returns the first one that matches. Overheat, for instance, is checked only after Crisis, Contraction, Slowdown and Stagflation have already failed to match, so reaching an Overheat reading already rules those four phases out first. Each phase also carries its own fixed sector overlay, a favored and avoided pairing set in advance, independent of that day's specific reading.
On 2026-09-30, the Cycle Clock read Overheat, phase three of the eight: late cycle, growth peaking, policy tightening.
On 2026-09-30 the Cycle Clock's reading was Overheat, phase three of the eight-phase taxonomy, described by the engine as late cycle, growth peaking, policy tightening.
A single reading only ever names one of the eight phases at a time. It does not say how close the classifier came to a neighboring phase, or how long the current one has held. Overheat specifically sits behind four other phases in the classifier's own check order, so reaching it already carries information about what the cycle is not.
The engine's guidance changes by phase. Overheat's is shorter holds, faster profit-taking and tighter stops. Each phase also carries a fixed sector overlay, drawn from the same cyclical and defensive buckets this overview does not recompute phase by phase.
This taxonomy has no single threshold to cross. It is eight discrete labels, not a score. What moves a reading from one phase to another is a change in the underlying combination the classifier reads: the rates regime, the structural bias, credit and volatility conditions, and curve shape. Overheat's own three exits, toward Recovery, Reflation, or back toward Stagflation, Slowdown or Contraction, depending on which of those inputs changes, are named in that phase's own piece.
The classifier has no built-in delay to confirm a move before changing the reading, for any of the eight phases. A shift in the underlying inputs can move the label from one classifier run to the next without lingering first to confirm it.
This piece names the eight phases and how the classifier works, not what each of the other seven means in practice. Only Overheat, today's actual reading, is covered in depth here. The other seven each have their own mechanism and their own sector overlay, and describing them without a dated, sourced example would be exactly the kind of invented content this pipeline exists to avoid.
It also does not say how long a phase typically holds, or how often the classifier flips between neighbors. The taxonomy is fixed; the frequency and duration of any one phase are separate, unanswered questions here.
Naming a phase's own sector overlay is not the same as saying which individual stocks inside those sectors are actually leading. That ranking is a separate question the sector leadership score answers.
Goldilocks, Reflation, Overheat, Stagflation, Slowdown, Contraction, Crisis and Recovery, numbered one through eight by the classifier's own regime index. A reading is always exactly one of these eight labels, assigned by a fixed decision tree rather than a continuous score.
It reads the rates regime, credit and volatility conditions, and curve shape, and works through the eight possibilities in a fixed order, returning the first one that matches. Overheat, for example, is checked only after Crisis, Contraction, Slowdown and Stagflation have already failed to match.
It can. The classifier has no built-in delay to confirm a move before changing the reading, so a shift in the underlying rates, credit, volatility or curve inputs can move the label from one classifier run to the next without lingering first.
Yes. Each of the eight phases carries a fixed favored and avoided sector overlay set in advance, the same pairing every time that phase is read, independent of that day's specific market conditions.
No. It describes which combination of current rates, credit, volatility and curve conditions the classifier matched, not a prediction of what comes next. The classifier can move to a different phase on the next run without warning.
The eight-phase taxonomy is fixed, but a reading is not sticky: the classifier can move from one phase to a very different one across a single run if the underlying inputs shift, without confirming the move first. Treat any one reading as that run's classification, not a durable regime call.