REKT Alpha

The Alpha Score (0–100) grades how strong a wallet's REALIZED trading record looks in the history REKT CHECK could measure. It is the mirror image of the REKT Score, which grades damage.

It is not advice, not a signal, not a prediction and not copy-trading. It describes what already happened. A high Alpha Score says nothing about the next trade.

• Produced by the current analysis engine (older snapshots are never scored).
• Demo results are never scored.
• At least 12 measurable closed positions.
• At least 25 classified trades.
• Historical price coverage of at least 85%.
• At least $500 of realized USD volume behind the record.

When a wallet misses any of these, we say so plainly instead of publishing a score built on thin data.

• Realized win rate (25 pts) — share of measurable closed positions that were profitable, scaled from 35% to 80%.
• Net realized profit (25 pts) — realized gains minus realized losses in USD, log-scaled between $500 and $250,000.
• Sample size (15 pts) — more measurable closed positions means a more trustworthy record.
• Profit distribution (15 pts) — penalises records where one lucky trade produced nearly all of the gains.
• Loss control (12 pts) — realized gains versus realized losses (profit factor).
• History depth (8 pts) — how much classified history backs the record.

• 🦈ALPHA PREDATOR · 90+
• 🧠SMART MONEY · 80+
• 🔥CONSISTENT COOK · 70+
• 🧾HAS RECEIPTS · 60+

Only wallets scoring 60+ that explicitly opted into public listing can appear in the COMMUNITY OPT-INS list.

The Smart Wallets board is not a list of wallets that came to us. A background engine samples liquid, actively traded Solana tokens, collects the wallets that publicly swapped them, and then puts each candidate through the same REKT CHECK analysis engine that scores you. Third-party profit figures are never published: if we ever hold one, it is kept only to sanity-check our own number.

• Automation excluded — wallets whose trade frequency, age or holding behaviour looks like a bot, MEV or infrastructure account.
• Non-comparable inventory excluded — wallets whose gains came largely from tokens they never measurably bought. That is a provenance statement, not an accusation.
• One-hit wonders excluded — a wallet cannot rank when most of its realized gains came from a single position.
• Recent record required — ranking uses the selected window (7D, 30D or 90D, default 30D), with minimum realized profit, win rate, profitable tokens and sample size in that window.

Discovery runs on a schedule, server-side, at a deliberately gentle pace. Opening the board calls no provider and re-scans nobody, and we never lower the thresholds just to fill the board.

Confidence describes OUR data, never the wallet. It drops when historical pricing coverage is weak, when the transaction history was capped, when the recent sample is small, when profit is concentrated, or when an external reference disagrees materially with our own figure. Only HIGH-confidence wallets are published to the board.

Community listings show a shortened address only. Nobody is listed from their own analysis without opting in, and an admin can hide any listing.

Discovered wallets are found from PUBLIC on-chain market activity, so the public address is shown. No identity, owner, name or off-chain data is attached, inferred or stored.

Your watchlist is private to your browser: we store a random key (hashed on the server), never an account, email, signature or wallet connection. Clearing browser data clears the list.

Alpha pages read SAVED analyses only. Opening REKT ALPHA never scans a wallet, so figures update only when a wallet is analyzed again. "Last analyzed" means exactly that — it is not live chain activity.

The same saved analysis always produces the same Alpha Score. Scoring method: alpha-v1. When the method changes, the version changes with it.