Glossary

A market where shares pay $1 if an outcome happens and $0 if it does not. The price of a share is the market's probability estimate.

The two sides of a binary market. One YES plus one NO always redeems for exactly $1, so their prices always sum to $1.

The moment a market's outcome becomes final and winning shares pay out $1 each. On World this happens on-chain, typically within a minute of resolution.

The determination of which side won. Crypto rounds resolve from price feeds; sports resolve from results.

The Token-2022 stablecoin World markets settle in. All dollar figures on Foreteller are real CASH amounts.

A short repeating market, like BTC Up, that asks whether a price will be higher at the end of a few minutes. New rounds open continuously.

Depositing $1 of collateral mints one YES and one NO share. Burning a matched pair redeems the $1. This is what makes prices probabilities.

Exchanging winning shares for CASH after settlement. Losing shares burn worthless. On World, redemptions are keeper-executed, so winners do not sign their own payout.

The price read as a percentage. A 30c share implies 30%. If your own estimate is higher, the share is value; if lower, it is overpriced.

Whether prices match reality over many markets. Perfectly calibrated 70c contracts win exactly 70% of the time.

What a bet is worth on average: probability x payout minus cost. Positive EV is the only reason to enter a trade.

Realized outcome minus entry price, per share. The core skill measure: buy at 40c, outcome wins, edge is +60c; it loses, edge is -40c. Averaged over many calls, edge separates skill from luck.

The market's final price minus your entry price. Beating the close consistently means you saw fair value before the market did, the gold-standard skill signal from sports betting.

The margin a venue builds into prices. When YES plus NO cost more than $1, the difference is vig.

The tendency for low-probability outcomes to be over- or under-priced. On World we measured the opposite of the classic bias: 3-8c tails have historically won 2-3x their implied rate.

The side priced above 50c. Our data shows World favorites in the 60-72c band have won only about 48% of the time, far below what their price implies.

A program that quotes prices and takes the other side of every trade. On World, the Janus AMM is the sole counterparty; there is no order book.

The bot that updates the AMM's quotes. It reprices on new information in discrete steps, not continuously, which is why quotes can briefly go stale.

The gap between what you pay buying and receive selling. World has no explicit spread, but quote stickiness creates an effective spread we measured at roughly 5c round trip on mid-priced contracts.

Price movement caused by your own order size. Thin markets move more; our paper trading simulates this honestly because real fills would too.

How much size a market can absorb without moving. On World, depth is one balance sheet (the AMM pool), and edge has historically degraded above roughly $5k positions.

The price during a market's life, before it converges to the truth near close. This is where World prices are least reliable, and where the edge lives.

Trading while the underlying event is happening, like a match or an open crypto round. Prices move fast and stale quotes appear.

The source of truth a market resolves against, like a price feed for crypto rounds.

The capital you can afford to lose entirely, managed as a unit. Every sizing rule is expressed as a fraction of it.

The formula for optimal bet sizing given your edge: bet a fraction of bankroll proportional to your advantage. Most professionals bet a half or quarter Kelly because overestimating edge is easy.

The randomness in outcomes even when your decisions are right. Binary markets are high variance; a real 60% edge still loses 4 times in 10.

The distance from your bankroll's peak to its lowest point after. Survivable drawdown is the design constraint for sizing.

Losing the bankroll before your edge can express itself. Betting too large converts a winning strategy into a losing one.

Making bigger, worse bets after losing. Detectable in the data as sizing up after drawdowns; the discipline metric on Foreteller profiles measures the opposite.

Holding both sides of a market to reduce risk. Legitimate for locking profit; our skill scoring nets hedged pairs out so they cannot inflate confidence.

Foreteller's 0-100 forecasting skill score: the lower confidence bound of your average edge per call, mapped through a logistic. It answers 'how sure are we this is skill, not luck'. One lucky hit cannot fake it; many small good calls cannot hide.

One (market, side) position held to resolution, with entry price, size, and outcome recorded permanently. The unit of the skill dataset.

The lower 90% confidence bound on your mean edge. Positive worst-case edge over many calls is the strongest evidence of real skill.

How much size a trader's edge survives. Measured as size-weighted edge versus average edge: if your bigger bets do worse, your capacity is small, and funding decisions respect that.

The share of your active trading days with positive average edge. Skill shows up repeatedly; luck clusters.

The correlation between your stake and your realized edge. Positive means you bet bigger when you are right more, the Kelly instinct.

The number of scored calls behind a statistic. Small samples mean nothing: 10 wins proves less than 60% over 200 calls. All Foreteller scores shrink toward neutral on small samples.

Trading virtual capital at real market prices, settled by real outcomes. On Foreteller it is the evaluation layer: prove skill on paper, graduate to funded capital.

Capital allocated to a proven trader, sized by their scores, with profits split. The graduation prize of the whole ecosystem.

A refundable deposit a funded trader posts that absorbs initial losses, aligning incentives between trader and capital.

Trading with yourself or in circles to fake volume or records. The skill engine filters it: hedged pairs are netted, dust is floored, and confidence bounds punish thin samples.

Splitting activity across many wallets to farm rewards or reroll luck. Confidence-bound scoring makes this self-defeating: each wallet has fewer calls, so each scores lower.