Strategies·November 17, 2026·7 min read

Sharp Money and Line Moves on Polymarket: How to Follow Without Chasing

How to tell a real repricing from a thin-book print, which confirmation signals matter, and when the right trade is to fade the move.

sharp moneyline movementmarket microstructurepolymarketexecution
Business newspapers — illustrative image
Photo: StockSnap · CC0Source

A line move on Polymarket is information only if it is backed by size, a catalyst, and a new level that holds. Everything else is noise: a single print in a thin book, a quote that widened, or your own order moving the price. The difference decides whether you follow, fade, or pass.

Most visible moves are small, temporary, and caused by impatient flow. The traders who profit are not the fastest chasers; they can tell a repricing from a liquidity event before they click.

Real repricing versus noise

Start with the book. A real repricing usually has more than one trade in the same direction, volume large relative to the recent average, a new level that holds after the first order fills, both sides of the book moving, and confirmation from a related market.

A noise move has one print, a thin book, a spread that widens after the trade, and no sibling confirmation. The classic case is a single buy that lifts the ask from 0.50 to 0.56 while the best bid stays at 0.48. The midpoint jumped, but nobody will buy at the new level.

Polymarket's display rule compounds this. The quote is the midpoint unless the spread is wider than $0.10, when the last trade is shown. A widening book can move the screen with no trade behind it.

The information in a trade

A trade is a fact about willingness to pay, not about the truth. Ask who traded, how much, and why now.

Wallets are public on Polymarket. A long history or a pattern of being early on similar markets is a signal, but a weak one: wallets can be sybils, histories can be farmed, and a large buyer can be wrong.

Size is more informative. A 20,000 share buy that walks three levels changes the book; a 500 share clip changes the print. If you cannot see depth before and after, you cannot tell which happened.

The catalyst is the third input. A move seconds after a court ruling or data release is likely information. A move at 3 a.m. with no news is more likely flow.

Confirmation signals

The best confirmation is a second market. If the same event or a mechanically related contract moves the same way at the same time, the move is more likely real. If the sibling is unchanged, it is more likely idiosyncratic.

In a negative-risk event, mutually exclusive outcomes must sum to $1. If one jumps from 0.30 to 0.38 and the others do not fall, the event is temporarily overpriced. That is either an unpropagated repricing or a liquidity event; sibling prices tell you which.

The book rebuilding is a second confirmation. If the swept ask is replaced at 0.56 and the bid moves to 0.54, the market accepted the price. If the book stays hollow, the move is fragile. A third is an external venue: a sportsbook line or exchange price that agrees with the move.

Latency is part of the edge

In a liquid market the useful life of a move is short. Professionals watch the same book, and the price can return to fair within seconds. If your data is a minute late, you are not following sharp money; you are providing exit liquidity.

Define the half-life: if a 4 cent jump retraces half within 10 minutes, a 15 minute delay makes you a buyer of the retrace.

False moves and manipulation

Thin markets produce fake moves mechanically. A large order sweeps the book, the midpoint jumps, and the next trade reverses it. A maker pulls a quote, the spread widens, and the display changes without a trade. A trader posts a large visible order, then cancels before it fills.

Resolution ambiguity is another source. A report that a candidate is "considering" a run is not a filing, and a story about a related company is not the named company. The price can move on the headline and reverse when traders read the rule.

Manipulation is real in thin markets. A holder can buy a small amount to move the mark. Public trade data makes some of it visible after the fact, but a large print is not informed just because it is large.

When to fade a move instead

Fading means selling the side that jumped, or buying the other side, on the view that the move overshot. It works when the move is liquidity-driven, the news is ambiguous, or the new price is inconsistent with a related market.

A setup has four parts: the move is large relative to the book and recent volume; there is no catalyst, or the catalyst does not meet the rule; a related market has not confirmed; and you can enter with at least two cents of edge after fees. Define the invalidation before entry. If the price continues 2 to 3 cents, your read is wrong.

A worked example

Suppose a politics contract is 0.48 bid and 0.50 ask, midpoint 0.49. A single buy sweeps the ask to 0.56. The best bid stays 0.48 and the related market stays 0.49.

Chase at 0.56: the fee is 0.04 times 0.56 times 0.44, or 0.99 cents, so the all-in cost is 56.99 cents. If your true probability is 60 percent, expected value is 3.01 cents per share. If it is 57 percent, the edge is zero. You need a strong view to pay 7 cents above the old midpoint.

Wait and buy at 0.535: the fee is about 1 cent, so the cost is 54.5 cents. The same 60 percent view now has 5.5 cents of value. A 2.5 cent retrace improves the trade by more than the entire fee.

Fade by buying No at 0.445, the same as selling Yes at 0.555. The fee is 0.99 cents, so the all-in cost is 45.49 cents. You break even if No wins 45.49 percent of the time, meaning Yes at most 54.51 percent. At a 50 percent fair value, expected value is 4.51 cents per share, about 10 percent. A stop at Yes 0.60 means No at 0.40, a loss of about 5.5 cents, or 12 percent. If the sweep was real news, the price can run to 0.65, which is why the stop and size matter more than the entry.

How to execute

  1. Snapshot the book before the move: best bid, best ask, midpoint, top three levels, and recent volume.
  1. Capture the prints during the move: size, side, wallet, and timestamp.
  1. Check related markets and an external reference before acting.
  1. Classify the move in writing: follow, fade, or pass.
  1. Use limit orders at fair value. Do not chase more than 2 to 3 cents beyond the old midpoint.
  1. Size at 0.5 to 1 percent of bankroll, and set the invalidation before entry.
  1. Record the outcome and closing price to measure whether your follows and fades have positive CLV.

Risks

The first risk is adverse selection. When you chase, the seller may know more than you. When you fade, the buyer may know more.

The second is latency. A strategy that works on minute bars can fail live if your feed is delayed or your order fills slowly. Measure live slippage, not the backtest.

The third is manipulation. Thin markets can be moved for the mark, and a large print can be a performance rather than a position.

The fourth is resolution. A move that is right about the probability can still lose if the market resolves 50/50 or the source is ambiguous.

Checklist

  • Did more than one trade print in the same direction, or was it a single sweep?
  • Did the book rebuild at the new level, or is it hollow?
  • Did a related market or an external venue confirm the move?
  • Is there a catalyst, and does it meet the resolution rule?
  • Is my total latency shorter than the half-life of the move?
  • If I am fading, do I have at least two cents of edge and a defined stop?

Sharp money is not a label you can read off a chart. It is a hypothesis about why the price moved, and it is useful only if you can confirm it before the move is over. The traders who follow well are selective, fast, and cheap. The traders who fade well are more selective still, because they are betting that the crowd is wrong at the moment it is most excited.

Sources

  1. 1.Prices & Orderbook— Polymarket Documentation
  2. 2.Resolution— Polymarket Documentation
  3. 3.Fees— Polymarket Documentation
  4. 4.Beating the bookies with their own numbers— arXiv
  5. 5.Unravelling the Probabilistic Forest: Arbitrage in Prediction Markets— arXiv
  6. 6.Explaining the Favorite-Longshot Bias: Is it Risk-Love or Misperceptions?— National Bureau of Economic Research

More from the journal

All articles