Whale Position Lifecycle: Entry → Defense → Un…

Whale Position Lifecycle: Entry → Defense → Unwind (and How to Detect Each Phase)

Traders love talking about “signal half-life.” It’s comforting: a single number that tells you when a whale trade stops mattering.

Reality is messier and more useful.

Whales don’t buy once, post a tweet, and vanish. They manage a position like a system:

Entry: build exposure
Defense: add or reshape exposure when price moves against them
Unwind: reduce exposure into strength, or exit after the thesis changes

If you can detect which phase you’re in, you stop copying trades blindly and start copying behavior.

Internal links:

Whale profiles: /whales/0x94f199fb7789f1aef7fff6b758d6b375100f4c7a
Smart money leaderboard & methodology: /smart-money

1) Model It as a State Machine

The simplest lifecycle model that works is a 3-state machine with clear transition signals:

ENTRY  ->  DEFENSE  ->  UNWIND
  \_____________________/
        (thesis reset)

Your job is not to guess intent from one trade. Your job is to detect persistent net behavior:

net position change (size up/down)
aggressiveness (taker vs maker implied by execution patterns)
price context (are they adding after adverse moves or chasing momentum?)

2) What Data You Can Verify (Public APIs)

You do not need privileged data to do this well. Polymarket provides:

Gamma API for market metadata (slugs, condition IDs, outcomes)
Source: Gamma “Get markets” docs (fetch markets by slug, events, pagination). https://docs.polymarket.com/developers/gamma-markets-api/get-markets
Data API for positions, closed positions, and trade history (per wallet)
Source: Polymarket endpoints overview (Data API base + endpoints). https://docs.polymarket.com/quickstart/reference/endpoints

Two endpoints matter most for lifecycle:

Current positions:

curl "https://data-api.polymarket.com/positions?user=0xYOUR_WALLET"

This returns size, avgPrice, curPrice, and related fields for each open position.
Source: “Get current positions for a user”. https://docs.polymarket.com/api-reference/core/get-current-positions-for-a-user

Closed positions (settled / realized PnL):

curl "https://data-api.polymarket.com/closed-positions?user=0xYOUR_WALLET"

This returns realizedPnl, totalBought, avgPrice, timestamp, and market identifiers for positions that have closed.
Source: “Get closed positions for a user”. https://docs.polymarket.com/api-reference/core/get-closed-positions-for-a-user

From these, you can reconstruct:

whether a wallet is still exposed
whether it tends to scale entries
whether it typically unwinds into liquidity or exits abruptly

To detect phase in real-time, you’ll also want trades:

GET /trades is available under the Data API (listed in the endpoints overview).
Source: https://docs.polymarket.com/quickstart/reference/endpoints

3) Phase Definitions (Operational, Not Vibes)

Phase A — Entry

Entry is not “a big buy.” Entry is:

net position size increases over a window
with average entry price that stabilizes (suggesting structured execution)
often concentrated in one outcome (YES or NO) rather than ping-ponging

Operational signals:

Δsize > 0 in consecutive snapshots
avgPrice moves slowly relative to curPrice (scaling instead of chasing)
trade sizes cluster around a target notional (TWAP-ish behavior)

Phase B — Defense

Defense is where whales separate from tourists.

Defense is:

adding exposure when price moves against the existing position, or
swapping exposure between outcomes / markets to keep thesis exposure but reduce liquidation risk.

Operational signals:

Δsize > 0 while curPrice moved against the position since the last snapshot
avgPrice improves meaningfully (they are averaging down, not just adding)
position becomes less “clean”: multiple fills, sometimes across multiple nearby markets

Defense is also where sloppy copy-traders get destroyed, because they join after the whale has already improved their entry.

Phase C — Unwind

Unwind is not necessarily “sell everything.” It’s:

a persistent reduction in size, often into liquidity, often as the market moves toward the whale’s thesis

Operational signals:

Δsize < 0 consistently
realized PnL starts appearing in closed-positions
trade sizes are smaller and more frequent (working an exit) or one-time (thesis flip)

4) A Reproducible Classifier (Rules That Survive Backtests)

You can build a phase detector without ML. Start with rules that you can audit.

Define a rolling window (e.g., 6 hours) and compute:

net_size_change = size(t) - size(t - window)
price_change = curPrice(t) - curPrice(t - window)
avg_price_change = avgPrice(t) - avgPrice(t - window)

Then classify:

if net_size_change > +S_min:
  if price_change is adverse: DEFENSE
  else: ENTRY
elif net_size_change < -S_min:
  UNWIND
else:
  HOLD

Two parameters matter:

S_min: ignore noise (e.g., 200 shares or $500 notional equivalent)
“adverse”: define adverse relative to side (for YES, price down is adverse; for NO, price up is adverse)

This yields a detector that is:

deterministic
explainable
easy to backtest and refine

If you want to go beyond rules, treat the phase label as your target variable and do calibration (precision/recall) rather than chasing accuracy.

5) Why Phase Detection Beats “Copy the Last Trade”

Copying individual trades fails because:

you don’t know the whale’s remaining position
you don’t know whether they are entering or exiting
you don’t know whether the trade is hedge, defense, or unwind

Phase detection fixes this by shifting your question from:

“Did a whale buy?”

to:

“Is the whale still building exposure, defending it, or distributing it?”

That’s the difference between being early and being liquidity.

6) How We Use This in SightWhale

When you open a whale profile, the goal is not to show a single impressive trade. The goal is to give you:

what the wallet does repeatedly
how it behaves around adverse moves
whether it exits cleanly or gets trapped

Start with a profile page and watch how behavior changes across markets:

/whales/0x94f199fb7789f1aef7fff6b758d6b375100f4c7a

For a higher-level view across wallets, use:

/smart-money

7) Practical Takeaways

Entry is a process: one trade is not a position.
Defense is informational: averaging down is a signal only if it’s persistent and structured.
Unwind tells you more than entry: good wallets exit like professionals—quietly, into liquidity.
Treat phase as the product: copy behavior, not screenshots.

Sources (External)

Polymarket API endpoints overview (Gamma / CLOB / Data API): https://docs.polymarket.com/quickstart/reference/endpoints
Gamma Markets API — Get markets: https://docs.polymarket.com/developers/gamma-markets-api/get-markets
Data API — Get current positions for a user: https://docs.polymarket.com/api-reference/core/get-current-positions-for-a-user
Data API — Get closed positions for a user: https://docs.polymarket.com/api-reference/core/get-closed-positions-for-a-user