Short version: prediction-market books are thin. Taking a full position in one click routinely moves the price two to four cents against you — which, on a contract where your edge was five cents, hands most of your thesis to whoever was resting on the other side. Splitting the order across a window fixes it, and costs nothing.
This is strategy six of ten in Algorithmic Trading Strategies, Translated for Prediction Markets. If you only adopt one thing from the whole series, adopt this one — it improves every trade you will ever make, on every strategy, forever.
The equities version
TWAP splits an order evenly over a set time period regardless of volume, aiming for an average execution price close to the start-to-finish price. It is the simplest execution algorithm there is, and institutions use it constantly, for one reason: their size moves the market and they want to stop paying for the privilege.
Why retail suddenly has an institutional problem
Here is the thing that surprises people coming from equities. In a liquid stock, a retail-sized order is invisible — you are a drop in an ocean and execution algorithms are somebody else's concern.
On a prediction market outside the flagship series, you are frequently one of the largest orders in the book. Your two hundred contracts are not a drop in anything. When you take the offer, you clear the top level, then the next, then the next, and your average fill is several cents worse than the price you clicked on.
That makes a retail trader's execution problem structurally institutional. Not because you are large in absolute terms, but because the book is small.
The arithmetic
Say your model puts fair value at 60 and the contract shows 55. Five cents of edge, and you want two hundred contracts.
- One clip: you clear three price levels and fill at an average of 58. Your realised edge is two cents, not five. You gave away sixty percent of the thesis to enter.
- Scaled: four clips of fifty, worked as resting limits across a couple of hours. You fill closer to 55–56. Your realised edge is four cents.
Same thesis. Same market. Twice the money. The difference is entirely execution, and it cost nothing but patience.
The tools
- Kelly — fix your total target size first, before you start working orders. Scaling into an undecided size is how you end up bigger than you meant to be. The reasoning is in the Kelly criterion for prediction markets.
- Kalshi Fee Calculator — check how many separate fills your size can absorb before per-trade costs eat the improvement you gained. There is a point where more clips is worse.
- Theta Edge Calculator — the counterweight. Patience has a price on a decaying contract. What an options trader sees in theta edge covers the shape of it.
- Probability Converter and EV Calculator — so the edge you are protecting is a measured number rather than a feeling.
How to actually run it
- Set total size before the first order. Non-negotiable. This is where most scale-ins go wrong.
- Choose the window from the contract's clock, not your patience. Days to expiry: hours are fine. Hours to expiry: minutes.
- Work limits inside the spread. You are trying to be the resting order that someone else pays to hit — the same position a market maker occupies, just for one trade.
- Re-check the thesis at the halfway mark. A scaled entry has a built-in advantage: if the reason to be in the trade evaporates, you are only half-sized when you find out.
- Accept a partial fill. Not getting the last fifty contracts at a good price is a better outcome than getting them at a bad one. This is the discipline people skip.
Where it goes wrong
Two ways.
Scaling into a decaying contract. On a short-dated market, the value you lose to time can exceed the slippage you saved. That is not discipline; it is a leak wearing discipline's clothes. Check the clock first.
Signalling. A perfectly regular clip every ten minutes on a thin book is legible to anyone watching, and someone will trade in front of it. Vary the size and the interval. Institutions solved this decades ago with randomisation; you can solve it by not being a metronome.
Next
The volume-aware cousin of this is VWAP and catalyst timing — trading with the liquidity instead of spreading evenly through the desert between catalysts. If you want to be on the resting side permanently, read market making on Kalshi. And smart order routing is the venue-level version of the same instinct.
Trade responsibly, and stop paying the book to let you in.