grade ranks a wallet on profit it has already realized, review_score ranks a single large trade, sharp_money says which side a market’s large-trade money is on, and the discovery scores order the explore feed. Pick the field that matches your question, because the others will not answer it.
Wallet scores
A Trader response carries all four of these. A leaderboard entry carries every one exceptrank.
Every field in this table is left out of the response when 0xinsider has no value for it, so read each one with a default rather than expecting
null.
grade and streak_tier measure different things. An A wallet can be cooling, and a B wallet can be hot.
The review score on one trade
Every large trade carriesreview_score, a number from 0 to 1. A higher score means read this trade first; it does not predict how the market resolves. Four parts add up to a raw score out of 80, which is then divided by 80 and held inside 0 to 1.
review_score is computed fresh on every read, so it falls as the trade ages and moves with the wallet’s current win rate. recorded_review_score is the same formula frozen when the trade was first stored, and it is null on trades recorded before that field shipped.
signal_score and recorded_signal_score are the older names of the same two numbers. They stay in every response with the same values, and the spec marks them deprecated.
Which side the money is on
Market intel and Sharp money flows both return asharp_money object. smart_money is a deprecated alias that carries the identical bytes.
The two routes count different wallets. Market intel adds up every tracked large trade with no grade filter at all. Sharp money flows applies
min_grade, which defaults to B, so ungraded wallets are left out.
Both read large trades only, the same rows Whale trades returns. A fill below the large-trade size floor is in neither figure, whatever the wallet’s grade.
Discovery scores
Explore markets returnsdiscover_score and score_components on each market. discover_score is the ranking behind sort=hot. The same route also returns smart_score, smart_count, and smart_label, which can each be null; the explore smart_score is a different number from the sports edge smart_score below.
Sports edge scores
Pre-game sides uses three of its own scores.s_count, a_count, and b_count are how many S, A, and B wallets hold the heavier side, and sharp_pct is that side’s share of the graded dollars in the market.
The route lists markets that have not started yet, inside the kickoff window you ask for, where one side holds clearly more of the graded money. A market split close to 50/50 is left out.
min_grade takes S, A, or B, and C, D, or F answers 400.
Choose the field
What this does not tell you
- How the market resolves.
review_scorereads size, price, win rate, and age, and never the market’s current price or its result. - Whether a
YESflow is right.directionis the sum of fills already made, not a forecast. - How skilled the wallet is.
review_scorescores one trade; skill and copyability are in Quant metrics.