Fatigue does not announce itself.
It accumulates in microfractures.
Invisible.
Until propagation begins.
Twelve months since the first cascade.
Nine months in compression regime.
Capital buffers intact.
Margins reduced.
Volatility subdued.
No singular shock.
Just repetition.
Repeated refinancing at thinner spreads.
Repeated asset repricing at lower multiples.
Repeated strategic restraint.
Repetition consumes resilience.
Han Zhe introduced a new overlay:
Cumulative Stress Function.
Not peak stress.
Integrated stress over time.
In physics, fatigue damage often scales with cycles and amplitude:
Where
n = number of cycles
σ = stress magnitude
m = material sensitivity exponent
Even low stress,
repeated enough,
approaches failure.
Markets displayed a subtle behavioral shift.
Risk appetite not returning despite stability.
Liquidity present but hesitant.
Participants slower to deploy capital.
Fatigue influencing decision latency.
Decision velocity declined 14% quarter over quarter.
Small number.
Behaviorally meaningful.
Elastic Duration Index stable.
Synchronization velocity near zero.
Yet Cumulative Stress Function trending upward steadily.
This was not a system accelerating.
It was a system tiring.
A credit event emerged in a secondary market.
Previously ignored.
Now amplified modestly.
Not because scale large—
But because tolerance lower.
Tolerance narrows under fatigue.
Margin for error shrinks.
Gu Chengyi framed it clearly:
"Fatigue reduces elastic bandwidth."
"Bandwidth narrows, not suddenly—progressively."
The danger is not cascade.
The danger is miscalibration.
Under fatigue, reaction may be late.
Or excessive.
Both destabilizing.
He reviewed reaction time metrics from prior quarters.
Crisis phase: response within 40 minutes.
Compression phase: response within 2 hours.
Current fatigue phase simulations:
Projected response delay creeping toward 3.5 hours.
Decision friction increasing.
Friction is silent risk.
In dynamical systems, response lag introduces phase shift:
If τ grows too large,
feedback destabilizes.
Delay can transform stable system into oscillatory one.
Timing again.
Always timing.
Treasury adjusted governance protocol.
Reduced approval layers.
Pre-authorized buffer deployment thresholds.
Shortened τ intentionally.
Architecture compensating for human fatigue.
Meanwhile, markets remained outwardly calm.
Yield curve still inverted.
Credit spreads moderately wide but stable.
Volatility low.
Externally uneventful.
Internally cumulative.
Late one afternoon, liquidity in a narrow asset class thinned sharply.
No news trigger.
Order books shallow.
Participants cautious.
Fatigue visible as withdrawal from marginal engagement.
Within hours, liquidity partially returned.
But signal noted.
Elasticity exists.
But snap-back slower.
Gu Chengyi summarized in his nightly brief:
"Fatigue does not break systems."
"It delays correction."
And delay—
Is the most dangerous variable.
Because when correction finally occurs,
it compensates for accumulated drift.
And compensation is rarely linear.
The architecture is stronger than before.
The buffers thicker.
The monitoring sharper.
But people operate it.
And people tire.
The mathematics of resilience now includes psychology.
Bandwidth.
Latency.
Tolerance.
Chapter 173 will not test liquidity.
Nor duration.
It will test discipline under fatigue.
Because when pressure persists long enough—
The next move is not panic.
It is error.
And error—
In tightly coupled systems—
Propagates faster
Than stress ever did.