Changelog

Version history for the WAX Arb Scanner. Newest first.

v1.8.0 2026-08-14 Added

Cross-Market Spreads now models slippage. Previously it just compared raw spot prices with no notion of trade size, so the "spread %" was purely theoretical. It now solves a closed-form profit-maximizing size against the AMM's real reserves (both directions derived and verified against brute-force numerical search) and reports the realized profit at that size. The orderbook leg still assumes you can fill at the quoted price, since Alcor's API doesn't expose depth beyond the last trade — treat sizes as an upper bound.
Token contract now shown in parentheses everywhere a token appears — routes, pairs, and the new token picker — e.g. "GOLD (goldgoldgold)". This makes same-ticker collisions visible instead of hidden.
Pick any base token for the triangular scanner, not just WAX. A searchable field lets you choose any token seen in the currently loaded pools; the search stays a 3-hop cycle (token → hop → hop → token) regardless of which token you pick. An unrecognized entry falls back to WAX with a visible note rather than failing silently.
Verified all three changes end-to-end against the actual shipped code (not just the underlying formulas): non-WAX base token routing, contract-label rendering, budget-capping and the size-warning flag all re-tested together after the base-token generalization.

v1.7.0 2026-08-14 Fixed

Fixed the cause of "some routes are real arb, some aren't." Routes were matched purely by ticker symbol (e.g. "GOLD"), but WAX has thousands of permissionless token contracts and it's common for unrelated tokens to reuse the same ticker. A route through two different contracts that both happen to be called GOLD isn't a real trade at all — you'd end up holding a token the next hop can't actually use.
All matching (triangular graph edges and cross-market spread pairing) now keys on Alcor's actual unique token id (symbol-contract, e.g. wax-eosio.token) instead of the bare symbol. Verified with a test: a fabricated route through two different "GOLD" contracts now correctly produces zero results, while a route through the same contract's GOLD twice still works.
Display labels (route names, pair names) still show the human-readable symbol — only the underlying matching logic changed.

v1.6.0 2026-08-14 Added

Added an "Available WAX" input. Every route's profit is now computed at min(that route's own optimal size, your available WAX) rather than always assuming the unconstrained optimal — so a route requiring more capital than you have isn't shown as if you could fully capture it.
Routes are ranked by realized profit at your actual size, not by unconstrained optimal size. This matters: a deep-liquidity route capped well below its own optimal can still beat a shallow route's full optimal, because deeper pools have less slippage per unit even far from their peak. Verified against a synthetic small-budget-vs-large-budget test.
Added a "Best trade for your available WAX" summary panel above the table showing the top route in plain language, including whether it's capped by your balance and, if so, what the route's true unconstrained optimal would need.
"Optimal Input" column renamed to "Trade Size" and now shows a "capped" tag when your balance is the binding constraint rather than the route's own slippage curve.

v1.5.0 2026-08-14 Added

Triangular routes now show the actual profit-maximizing trade size instead of only a no-slippage marginal rate. Each hop's constant-product swap is represented as a Mobius transform; composing the three hops and solving where marginal output equals marginal input gives a closed-form optimal input size, verified against brute-force numerical search.
Added "Est. Profit (WAX)" and "Optimal Input (WAX)" columns; "Profit %" now reflects the realized return at that optimal size (always ≤ the old marginal-rate number, since slippage only ever makes returns worse as size grows).
Routes are now sorted by absolute profit (WAX) rather than percentage, since a huge percentage on a tiny trade is less actionable than a solid absolute return.
Flags routes (⚠) where the optimal size is large relative to the shallowest pool in the route — the model assumes plain full-range constant-product pools, and Alcor's concentrated-liquidity pools can behave differently once a trade pushes price outside the currently active range.

v1.4.0 2026-08-14 Fixed

Fixed the cause of absurd "profit" numbers (routes reporting quintillions of a percent). The code was reading a tvl field that doesn't exist on Alcor's pool payload (the real field is tvlUSD), so pool liquidity was always unknown and near-empty "dust" pools were treated the same as real, tradeable ones. A route through a pool holding a few cents produces a meaningless price ratio, and chaining three of those compounds into nonsense.
Added a "Min pool liquidity ($)" control (default $250) that excludes low-liquidity pools from both the triangular search and the cross-market comparison.
Added a "Weakest Pool ($)" column to the triangular results table showing the lowest-liquidity pool in each route, so results are self-verifying instead of requiring trust in the filter.
The "Pools parsed" indicator now also shows how many pools were excluded for low liquidity.

v1.3.0 2026-08-14 Fixed

Alcor's WAX API doesn't reliably send CORS headers for arbitrary browser origins, and the scanner's only fallback (api.allorigins.win) is an unmonitored free proxy that goes down or gets rate-limited without warning — when that happened, both data sources failed silently and the page just showed empty tables with no explanation.
Added a second fallback proxy (api.codetabs.com), tried automatically if the first one fails.
Fetch failures are now shown directly on the page in red text, not just logged to the console, so a data outage is obvious instead of looking like "there's no arbitrage right now."
Pointed the AMM pools request at the wax.alcor.exchange subdomain directly (matching the markets endpoint) instead of relying on a redirect from the bare alcor.exchange domain.

v1.2.0 2026-08-14 Added

v1.1.0 2026-08-14 Fixed

Fixed AMM pool reserve parsing — Alcor's real /api/v2/swap/pools response nests reserves inside tokenA.quantity / tokenB.quantity, not flat keys like reserveA. This was causing every pool to be silently dropped, so triangular routes and cross-market spreads never had real data to work with.
Added a "Markets parsed / Pools parsed" counter to the scanner UI so a data-parsing failure is visible instead of just showing an empty result table.

57c026a

v1.0.0 2026-08-14 Initial

Initial release: live WAX Arb Scanner comparing Alcor Exchange's orderbook markets against its AMM swap pools, plus a triangular-route search across AMM pools.
Runs entirely client-side against Alcor's public API, with a CORS-proxy fallback.

776824c