the lead-lag edge in prediction markets · built on TxLINE
Prediction markets trade
a step behind fair.
TxLINE strips the vig, so its odds are the true price. A prediction market lags it, and the cheap side is underpriced. Lagisalpha catches the gap the moment it opens, tells you which side to take, and shows it snap back 82% of the time. A repeatable edge you can act on.
scanning the market against the vig-free fair…
_
the edge, on one match
One match. $1.03M traded 10+ points below fair.
We aligned a prediction market's real fills to TxLINE's vig-free fair for all 90 minutes of Paraguay v France. The two prices sat within 1 to 2 points, the spread, the whole match. Then France scored, the sharp line jumped, and for the minutes the market took to catch up, over a million dollars traded 10 to 37 points below fair. Every one of those fills was a cheap side sitting there to be taken. Not a claim: a measurement, every fill on-chain.
$8.6M
traded in-play on this one match, all of it measured against the sharp fair.
1.25pp
median gap. The market tracks fair to the spread until news hits, then it lags.
$1.03M
changed hands 10+ points below fair, the cheap side left on the table around one goal.
on-chain
every fill a Polygon tx, every outcome a TxLINE on-chain settlement. Verify it.
Paraguay v France, in-play. Prediction market fills read on-chain from Polygon, aligned to TxLINE's vig-free 1X2. The full track record, across 16 matches, is on /proof.
why the edge is there
A slow market, a fast reference, a gap that repeats.
The market is slow
A prediction market reprices only when someone trades. New information outruns it, so its price lags the truth for real, tradeable seconds.
The reference is sharp
TxLINE's de-vig line is the consensus fair with the margin stripped out. It already holds the true price the market has not reached yet.
The gap repeats
Not luck on one match. The divergence opens every time news hits and closes 82% of the time. A pattern you can trade, not a story.
The edge is validated on 16 matches so far, so the return is a pilot, not a promise. The reach rate is the firmer read, and both tighten as matches accrue. See the numbers, with the confidence interval, on /proof.
the edge
The lag is the alpha.
A prediction market can only move its price by trading, so it is always a step behind the sharp, vig-free line. In the seconds around a goal the market sits below fair and the cheap side is there for the taking. That is not a one-off: across 16 matches, every call published and none filtered, the gap closed 82% of the time. The trade is the convergence, not the outcome: exit at fair, Kelly-sized on the gap; the same bets held to the final result hand the convergence profit back. Find the divergence, take the cheap side, and ride it to the fair.
how you trade it
01
The sharp line leads
TxLINE strips the vig from a live odds feed, so its price is the true probability. It moves the instant news hits, seconds before a traded market can follow.
02
The market lags
A prediction market only reprices when someone trades, so it sits behind. When it falls past the threshold below fair, the cheap side is underpriced, and we flag it: which side, how far off, how much size is there.
03
You take the cheap side
The gap closes 82% of the time as the market catches up. Take the cheap side and take profit when it reaches TxLINE fair, sized by Kelly on the gap. Ride it to the fair, not the final whistle.
04
Verify every entry
Each entry is a Polygon fill you can open; each outcome settles on TxLINE's on-chain scores. Don't trust the track record, check it.
we were handed a hypothesis. we tested it, then did one better.
The sharp line moving is not the edge. The market being slow to follow it is.
The obvious idea is a sharp-movement detector: watch TxLINE for significant odds shifts and see if they call the result. We built it and it is a coin flip, 58%. The edge is not the line moving; it is the market being slow to follow. A goal is new information: TxLINE reprices it instantly, but a prediction market only moves when someone trades, so for a window the cheap side sits below fair. That is the lead-lag, and it converges about 82% of the time. It is our strongest, most proven signal.
And the record rolls on its own. Every divergence the detector fires is published and scored: either side, any size, any minute, each side named by its team. There is no exclusion filter and no curated subset. Sizing is the only risk control: Kelly on the gap, capped at 30% of the balance per call so no single bet can ruin the account, and the ledger on /proof recomputes as each match settles. The mechanism, with nothing cut out of it.
58%
their sharp-movement detector at calling the result, a coin flip
82%
of lead-lag divergences converge to fair, the proven edge
~4×
a high-danger possession makes a goal that likely; the divergence it flags converges 84% vs 75%
On the pilot sample of 16 settled matches, in-sample. A promising signal, not a settled result.
why it runs on TxLINE
The only feed that shows the true price early.
Strip the vig and a price move stops being noise and becomes a measurable distance from the true probability. That distance is the whole signal, and only TxLINE's de-vig stream exposes it. Without it, a gap between two prices is just two prices; with it, the gap is exactly how far the market is behind, and which way to trade. To catch it live you want the fastest, most direct feed, which is TxLINE.
The leading indicator
We difference the market's book against TxLINE's vig-free fair, tick for tick. The two agree at the spread until news hits, then the fair moves first and the gap is your entry. No feed but the de-vig stream makes that visible.
More markets, more entries
Today the signal runs on the goals and match-result markets that stream de-vigged. The more of the book TxLINE streams beyond goals, cards, corners, the more divergences there are to trade.
check it yourself
Both sides of the edge are anchored on-chain.
The fair line is TxLINE's World Cup feed: odds and scores anchored on Solana. The market side is real fills read straight from Polygon. Nothing in the track record is asserted; both legs are public, so you can recompute the edge yourself.
TxLINE Solana-anchored World Cup feed.