KOL Selection

Not every crypto influencer makes the WhaleClaw roster. The 128+ tracked KOLs are selected through a three-filter process designed to ensure only traders with demonstrable edge get included.

The three filters

Filter 1: Rolling performance

We track who has been right recently — not who had a lucky call last year.

Monthly rolling window — Performance is measured over the current month, not all-time
Automatic demotion — KOLs who drift (changing style, declining accuracy) get their tier weight reduced
Automatic removal — KOLs who go inactive or consistently underperform are dropped from the roster

This keeps the network fresh. Crypto markets evolve fast, and last quarter’s genius can be this quarter’s noise.

Filter 2: AI classification

Every KOL is continuously analyzed by AI to determine:

Classification	What it means
Timeframe	Higher-timeframe macro vs. short-term scalper
Style	Trend-following, mean-reversion, event-driven, etc.
Reliability	Signal clarity, consistency, noise ratio
Best asset	Which assets the KOL has edge on (BTC focus for WhaleClaw)

Classification is refreshed weekly based on recent output. A trader who shifts from scalping to swing trading gets reclassified automatically.

This ensures the weighting system always knows how to weight each KOL’s call — a macro trader’s “long BTC” means something different from a scalper’s.

Filter 3: Human review

Machines find patterns. Humans catch what machines miss.

Accredited crypto traders on the WhaleClaw team manually review the roster:

Context the AI can’t see — Is this KOL known for pump-and-dump? Are they running paid groups with conflicting signals? Are they copy-trading someone else?
Quality assessment — Some KOLs have great calls but terrible risk management. Some are right 70% of the time but the 30% are catastrophic losses.
Red flags — Sudden changes in posting frequency, suspicious timing relative to market moves, inconsistencies between public and private channels

Human review is the final gate. If a KOL passes the performance filter and AI classification but the human review raises concerns, they don’t make the roster.

Tier system

KOLs that pass all three filters are assigned a tier:

Tier	Weight	Criteria	Count
Tier 1	3–5x base	Walk-forward validated, consistently profitable, high signal clarity	~10–15
Tier 2	1.5–3x base	Strong recent performance, good signal quality	~30–40
Tier 3	1x base	Meets baseline criteria, adds consensus value	~80+

Tier 1 KOLs are the ones validated by the backtest — traders like Tareeq, Woods, Eliz, Muzzagin, Binance Killers, and Vivian. Their calls carry the most weight in the consensus.

Tier 3 KOLs individually have less influence, but collectively they matter. When 60+ tier-3 traders all lean the same direction, that’s a crowd signal worth knowing about.

Why 128+ KOLs?

Diversification of opinion. A single trader — even a great one — has cold streaks. When you aggregate 128+ traders:

Individual errors wash out
Genuine consensus emerges from noise
Transitions become visible (when multiple traders flip direction simultaneously, something real is happening)
Contrarian signals appear (when 90% agree but price goes the other way, the crowd is wrong — and that’s informative too)

Think of it like polling. One person’s opinion is anecdotal. A poll of 128 people weighted by expertise is data.

How it stays current

The roster is a living system, not a static list:

Weekly classification refresh — AI re-evaluates every KOL’s style and timeframe
Monthly performance review — Tiers are adjusted based on rolling accuracy
Continuous monitoring — New KOLs can be added; inactive ones are removed
Team review cycle — Human review happens on a regular cadence

No KOL has a guaranteed spot. Edge is earned continuously, not granted permanently.