The same contract trades on two venues and the prices disagree. One book has an event at fifty-eight cents, the other has it at sixty-four. If you have looked at prediction markets for more than a week you have seen this, and the instinct is usually to buy the cheap one and sell the rich one and call it free money. Sometimes it is. More often the gap is telling you that one side has more information than the other, and if you take the naive arbitrage you end up short exactly the venue that is right.
The question I care about is not whether the gap exists. Gaps exist all the time because of fees, withdrawal friction, different resolution wording, and thin books that nobody bothered to update. The question is whether the gap is noise or whether it is smart money repricing one platform ahead of the other. And the cleanest tell I have found for that is the position layer underneath the price. Not the price itself, the flow.
Why the price alone lies to you
A price on a prediction market is just the last trade or the current midpoint. It does not tell you who is on each side or how much conviction is behind it. Two venues showing the same number can be there for completely different reasons. One might be pinned by a market maker who has not moved his quotes since the news broke. The other might have a large wallet that just walked the book up because they saw something.
So when I see a spread between platforms, the first thing I want is not the delta in cents. I want to know where the informed size sits. On a market like Polymarket, wallet flow is on-chain and readable, so you can actually see who took the other side and how big they went. That is the layer that resolves the ambiguity. If the richer venue got rich because a wallet with a real track record loaded the yes side, the gap is probably not an error. If the richer venue is just a stale quote and no size has moved through it, the gap is more likely something you can fade.
Here is the rough mental model I use. The price is the outcome of a negotiation. The whale is one of the negotiators. Reading the price without reading the flow is like reading the final offer in a deal without knowing which party had the leverage.
The overlay, step by step
The workflow is not complicated, it is just disciplined about the order you do things in. If you check flow before you have framed the gap, you will see a big wallet and talk yourself into whatever story fits.
- Frame the split first. Which venue is cheaper on the outcome, and by how much after fees and any withdrawal cost of moving capital between them. If the edge does not survive frictions, stop here.
- Pull the recent wallet flow on the venue where flow is visible. You are looking for size that is directional, one-sided, and clustered in time. A single large buy on the yes side over a short window matters more than a lot of small two-way churn.
- Check the wallet's history, not just the size. A wallet that is up over many resolved markets carries different weight than one that is large and new. Size without a record is often someone hedging an off-platform position, which tells you nothing about the outcome.
- Ask which direction the informed position points relative to the gap. This is the whole trick. If the big money is on the same side as the richer venue, the richer venue is probably right and the cheap venue is lagging. If the big money is on the cheap side, the cheap venue may be the one that already repriced and the rich venue is stale.
- Decide follow or fade. When flow confirms the richer venue, you follow by buying the cheap venue toward the informed price, expecting it to catch up. When flow is absent or contradicts the gap, you fade, because now it looks like a stale quote or a thin-book artifact rather than information.
The uncomfortable part is step three. Everyone wants to skip the history check because a big number is exciting. But size is not the same as information. I have been burned assuming a large wallet was smart when it was just a market maker laying off risk, and the position I read as a signal was actually the exhaust from a hedge.
Following versus fading, and when each breaks
Following the whale is the trade where the gap is real information and the slower venue has not caught up. You are betting that the cheap platform converges toward the informed one. This works best when the two venues resolve on genuinely the same wording, the informed wallet has a record, and the flow is recent rather than something from days ago that has already been priced.
Fading the gap is the trade where the spread is mechanical. No informed size, a stale quote, thin liquidity, or a resolution difference that makes the two contracts not actually the same bet. Here you are betting the gap closes because it never reflected information in the first place. The classic failure mode is fading a gap that was real, getting run over when the slow venue snaps to the informed price, and realizing the whale was ahead of you the whole time.
A few rules of thumb I keep coming back to:
- If you cannot see any meaningful flow behind the richer price, treat the gap as stale until proven otherwise. Absence of informed size is itself a data point.
- Confirm the two contracts resolve on identical criteria before you call anything a mispricing. A surprising number of cross-platform gaps are just two questions that sound the same and are not.
- Recent and one-sided beats large and old. A whale who loaded up before the last news cycle has already had their information absorbed.
- Never trust size alone from a wallet with no resolved history. Off-platform hedging looks exactly like conviction and means the opposite.
- Size the trade for being wrong about which venue is informed, because half the time the ambiguity does not fully resolve and you are taking a probabilistic edge, not a certainty.
The thing that makes this work at all is that prediction markets leak. Traditional order flow is hidden inside an exchange, but on-chain wallet activity is public, so you can see the negotiation happening rather than only the final number. That does not make you right. It just moves you from guessing why two prices disagree to having an actual reason to lean one way.
I still get it wrong plenty. But framing it as follow or fade, gated on whether real informed size is present, has kept me from the worst version of this, which is confidently shorting the venue that turned out to know something I did not. If you take nothing else from this, take the habit of reading the flow before you read the gap, and letting the wallet with a record break the tie.