Signal Half-Life in Prediction Markets: How Long Does Whale Alpha Last?

Most traders obsess over being first. In reality, how long a signal stays valid is often more important than how fast you see it.

On Polymarket, smart money (“whales”) place large, information-rich bets. Your edge doesn’t come from guessing what they know, but from understanding how quickly the edge they have decays.

Signal half-life: the amount of time it takes for a new information signal (like a whale trade) to lose half of its expected edge, because the rest of the market sees it, digests it, and re-prices.

If you assume every whale trade is timeless alpha, you’ll keep buying stale signals at the exact moment they flip from edge to exit liquidity.

This deep dive explains:

• the three phases of a whale signal’s life
• how to estimate signal half-life from on-chain and orderbook behavior
• practical rules you can apply when following whale alerts
• how a system like Whale Score™ can embed signal half-life into its weighting

Internal links:

• Whale leaderboard: /smart-money
• Whale profiles and Score: /whales
• Backtesting: /backtesting

1. The Three Phases of a Whale Signal

Phase 1 – Emergence: From Private Edge to Observable Flow

In the earliest phase, information is:

1. Private or tightly held (insiders, domain experts, quant teams)
2. Expressed via capital flow, not public statements
3. Sparse but high-conviction (few big tickets, not many small ones)

On Polymarket, you see this as:

• a single wallet or tight cluster lifting offers aggressively in a niche market
• orders that do not hesitate at mid or best ask when liquidity is thin
• little social chatter—the trade happens before the Twitter thread

Key traits of the Emergence phase:

• Low participation: one to three wallets doing most of the volume
• Directional: a clear push toward one outcome, not ping-pong scalping
• Price elasticity: the orderflow is willing to move the price several ticks

If you act in this window, you are effectively partnering with the originator of the edge. You are close to the source.

Phase 2 – Diffusion: The Market Notices

As time passes, more participants notice:

• the price has moved
• liquidity has been eaten away at a specific level
• a previously quiet market is suddenly active

Diffusion looks like:

• copycat flow: smaller wallets echoing the same direction, but with worse fills
• liquidity response: market makers reload the book; spreads adjust
• narrative lag: tweets, Discord messages, Telegram threads appear that explain a move already in progress

At this stage, the edge is no longer “pure information.” It is a combination of:

• residual informational edge
• orderbook microstructure
• emotional overreaction or underreaction by retail

The signal half-life often sits somewhere inside this phase. Past that point, you are no longer buying the original idea; you are trading crowd behavior around the idea.

Phase 3 – Exhaustion: From Signal to Exit Liquidity

Eventually, the signal decays:

• the price approaches a level that embeds the new information
• the original whales slow down or start offloading risk
• retail arrives late, with smaller, emotionally driven trades

Signs you are at or past exhaustion:

• volume is high but asymmetric size disappears (few big tickets, many tiny ones)
• the book shows thick size on the same side retail is piling into
• news has caught up: the “reason” for the move is now broadly known
• outcome odds have moved to a level where only extreme scenarios justify further edge

Acting here is often a negative-EV decision. You are no longer following information; you are funding the exit.

2. Approximating Signal Half-Life in Practice

You do not need a PhD to estimate whether a whale signal is still alive. You need a consistent way to measure time, volume, and structural change around a trade.

Step 1 – Timestamp the Origin

Define t₀ as the first moment where all three conditions hold:

• a wallet or cluster takes significantly larger size than the typical ticket
• the trade materially changes the orderbook or price (not just a nibble)
• no obvious public catalyst has just fired (the trade looks like initiative, not reaction)

This is your anchor.

Step 2 – Track Cumulative Volume and Price Drift

From t₀ forward, bucket time into intervals (for example, 5 minutes, 30 minutes, two hours) and measure:

• cumulative volume in the direction of the whale
• net price change relative to t₀
• concentration of that volume (few wallets vs. many)

A rough half-life estimate comes from the point where:

• price has realized most of the move implied by the original trade
• additional volume is dominated by copycats, not the initiator
• new information is visible in off-chain channels (tweets, news, group chats)

In many Polymarket-style event markets, this might be 30–180 minutes, not days.

Step 3 – Watch Wallet Behavior, Not Just Market Prints

A single large buy can be:

• part of a sustained build (multiple adds over hours or days)
• a one-off mispricing capture
• a hedge or cross-venue arbitrage leg

If the initiating wallet:

• keeps adding at higher prices → signal is still in Emergence/Diffusion
• starts unwinding or flipping sides → the informational edge is probably fully priced
• goes silent while retail continues in the same direction → exhaustion is close

Your half-life estimate must be conditional on the wallet’s behavior, not just the chart.

3. Embedding Signal Half-Life into a Whale Scoring System

A static score (“this wallet is 84/100 forever”) misses the dynamic part of edge.

A serious system should:

• penalize signals that only briefly produce edge before reversing
• reward signals that retain predictive power over a measurable window
• down-weight trades that are crowded instantly and decay fast

Conceptually, let:

• S(t) be the informativeness of a given whale signal over time
• half-life T½ be where S(t) ≈ ½ · S(0)

A good Whale Score weights wallets whose S(t) decays slowly, not just those with large S(0).

In practice, this can translate into:

• higher scores for wallets whose trades age well (PnL measured one hour, six hours, 24 hours post-trade)
• lower scores for wallets whose trades are mostly short-lived spikes followed by mean reversion
• time-decay factors baked into rankings and alert filters (for example, “ignore signals fired more than X minutes ago unless price has barely moved”)

4. Practical Rules You Can Use Today

Even without full infrastructure, you can adopt operational rules:

• Rule 1: Define your freshness window.
  Decide a maximum age (for example, 60–90 minutes) beyond which you treat a signal as “old” unless price has barely moved.

• Rule 2: Track initiator activity, not just symbols.
  Map alerts back to wallets and ask: Is this wallet still building, or already exiting?

• Rule 3: Avoid narrative-only trades.
  If you discovered the trade through social media, you are probably in Phase 2 or 3 already.

• Rule 4: Respect structural barriers.
  If price already moved into a range where only extreme outcomes justify it, treat new buys as exit-risk, not fresh alpha.

Prediction markets reward people who understand when a signal is still a signal, and when it has become a story. Half-life thinking is how you stay on the right side of that line.