No obvious scam flags. No sanctioned counterparties. Nothing unusual in token composition.
What stood out wasn’t what it did — it was when it did it.
Instead of reading transactions, I plotted activity across time:
24 hours × 7 days
Color intensity = transaction volume
What I expected to see was a human rhythm:
Clear active hours
A multi-hour lull overnight
Some weekday/weekend bias
What I saw instead was perfectly even activity.
No quiet window. No multi-hour inactivity. No timezone bias. Just steady execution across all hours, every day.
That alone is a strong signal. Humans don’t behave evenly across time — we sleep, pause, and lose consistency. Software doesn’t.
To confirm, I checked for burst patterns:
Sudden activity spikes at regular intervals
Repeated across days
Consistent ratios relative to baseline
At that point, the conclusion was unavoidable: this wallet was automated.
What surprised me wasn’t that bots exist — that’s obvious — but how easy it was to identify one once I stopped looking at balances and started looking at timing.
Time-based behavior seems harder to fake than transaction structure:
Labels can be wrong
Funds can be mixed
Ownership can be obscured
Sleep patterns can’t
Curious if others here have used timing or activity rhythms as a primary signal in on-chain analysis, or if there are known false positives I should be aware of.