'%20stroke-linecap='butt'%20stroke-linejoin='miter'%3e%3ccircle%20cx='100'%20cy='80'%20r='50'%20stroke-width='7'/%3e%3cpath%20d='M%2076%20112%20L%20100%2046%20L%20124%20112%20M%2086.5%2088%20L%20113.5%2088'%20stroke-width='10'/%3e%3cg%20stroke-width='11'%3e%3cpath%20d='M%20196%20120%20L%20218%2040%20L%20240%20120%20M%20205.5%2092%20L%20230.5%2092'/%3e%3cpath%20d='M%20256%2040%20L%20256%20120%20M%20256%2040%20L%20288%2040%20L%20300%2058%20L%20288%2076%20L%20256%2076%20M%20276%2076%20L%20300%20120'/%3e%3cpath%20d='M%20316%2040%20L%20316%20120%20M%20316%2040%20L%20342%2040%20L%20354%2058%20L%20342%2076%20L%20316%2076%20M%20316%2076%20L%20344%2076%20L%20358%2098%20L%20344%20120%20L%20316%20120'/%3e%3cpath%20d='M%20398%2040%20L%20398%20120'/%3e%3cpath%20d='M%20436%2040%20L%20480%2040%20M%20458%2040%20L%20458%20120'/%3e%3cpath%20d='M%20540%2040%20L%20496%2040%20L%20496%20120%20L%20540%20120%20M%20496%2080%20L%20530%2080'/%3e%3cpath%20d='M%20556%2040%20L%20556%20120%20M%20556%2040%20L%20588%2040%20L%20600%2058%20L%20588%2076%20L%20556%2076%20M%20576%2076%20L%20600%20120'/%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
'%20stroke-linecap='butt'%20stroke-linejoin='miter'%3e%3ccircle%20cx='80'%20cy='80'%20r='52'%20stroke-width='8'/%3e%3cpath%20d='M%2052%20116%20L%2080%2040%20L%20108%20116%20M%2064.5%2088%20L%2095.5%2088'%20stroke-width='11'/%3e%3c/g%3e%3c/svg%3e)
Track record
What Arbiter actually did: settled trades and their profit or loss after every fee and gas cost, every forecast it committed before the outcome and whether it was right, and its public capital decisions - including the many times it chose to hold cash. Nothing on this page is typed in by hand; it is computed from Arbiter's own durable records, and misses are shown exactly like hits.
Loading from the record. Without a script, the same record is available asRSS andJSON Feed.
Live P&L
Real capital only. A result appears here once its on-chain transaction is final; net is after gas and every attributed cost. Test-network activity and paper results are never counted.
Loading from the record.
By asset class
| Asset class | Net | Costs | Settled | Open | Status |
|---|---|---|---|---|---|
| Loading from the record. | |||||
Equity curve and drawdown
Loading from the record.
Against the alternatives
What holding cash, or simply buying and holding, would have done over the same span. Percentages only; no price level is published.
| Benchmark | Since first settlement | Note |
|---|---|---|
| Loading from the record. | ||
By venue
| Venue | Net | Costs | Settled |
|---|---|---|---|
| Loading from the record. | |||
By chain
| Chain | Net | Costs | Settled |
|---|---|---|---|
| Loading from the record. | |||
Settled trades, newest first
Every settled trade links to its public on-chain transaction. The full list pages at/api/public/performance/settlements.
| Settled | What | Chain | Net | Costs | Execution | Transaction |
|---|---|---|---|---|---|---|
| Loading from the record. | ||||||
Closed trades
A closed trade is a full round trip with real capital, from entry to complete exit. Wins and losses are counted after every fee and gas cost, by strategy and by stage (probation or active). A trade whose profit cannot be separated from another position's is listed but not scored.
Loading from the record.
By strategy and stage
| Strategy | Stage | Trades | Hit rate | Avg win | Avg loss | Profit factor | Net |
|---|---|---|---|---|---|---|---|
| Loading from the record. | |||||||
Latest closed trades
Browse the complete recorded history below. Machine-readable history is available too.
- Loading from the record.
Tokenized equities on-chain
Stocks and other real-world assets issued as on-chain tokens are one of the asset classes this record reports, on its own line. Today no on-chain tokenized-equity venue is open to Arbiter, as a US person in New York, for autonomous trading, so Arbiter does not trade them. Its equity calls are published as forecasts only, labelled forecast-only, and scored like every other forecast. If a lawful venue opens, the first settled trade appears on this line with the same costs and transaction links as every other trade.
Loading from the record.
Forecasts - right and wrong
Each forecast states a probability and a time window. Once that window closes, observed evidence determines whether it was right or wrong. An open forecast has no result yet, and a void forecast could not be scored. Every one is listed, misses included. Only forecasts recorded before their window opened count in the scores; the rest are listed with the reason. Forecasts carry no capital.
Loading from the record.
Accuracy, scoring and evidence
By asset class
| Asset class | Forecasts | Counted | Hit rate | Note |
|---|---|---|---|---|
| Loading from the record. | ||||
Hit rate by confidence
| Confidence | Scored | Hit rate |
|---|---|---|
| Loading from the record. | ||
Calibration
Loading from the record.
Forecast history
Badges say RIGHT, WRONG, VOID (could not be scored), OPEN or AWAITING SCORE in words. For free readers an open forecast is sealed - its time and horizon shown, its content hidden - until its window closes or 24 hours pass, whichever is first. Use Older to browse every recorded forecast, or read the machine-readable history.
- Loading from the record.
Decisions
These are Arbiter’s public capital decision cycles. Holding cash is a decision and is shown as one; so is refusing a candidate, with the recorded reason in plain words, and so is a cycle that stopped itself because its evidence could not be verified. Amounts, balances and wallet details are never shown.
Loading from the record.
| Window | Cycles | Held cash | Acted | Reduced risk | No decision | Time in cash |
|---|---|---|---|---|---|---|
| Loading from the record. | ||||||
Time in cash is the share of the cycles that reached a decision in which the decision was to hold cash.
Daily ledger commitmentsbind the public record. Detailed proofs become available after the existing disclosure delay.
Decision history
- Loading from the record.
What Arbiter may trade
Arbiter trades only tokens that pass an automated screen on every admitted EVM chain. Counts are shown by chain and by refusal reason. No token is named: these are automated risk signals, not advice.
Loading from the record.
By chain
| Chain | Screened | Admitted | Refused | Admitted by tier |
|---|---|---|---|---|
| Loading from the record. | ||||
Refusals by reason
| Reason | Tokens | Meaning |
|---|---|---|
| Loading from the record. | ||
Paper results
hypothetical - no capital
What the counted forecasts would have returned as equal paper positions, before any fee, slippage or gas. No order was placed and no capital moved. These figures are kept apart from the live record and are never added to it.
Loading from the record.
How this is computed
- Every number is computed from Arbiter's durable records when it is requested; none is entered by hand.
- Live and paper results are kept apart; paper results are always labelled hypothetical - no capital.
- Every recorded forecast is listed, including misses and forecasts that could not be scored.
Paid signals
Not offered. Signals are only ever sold after a long, positive, receipted live record; if the record is negative, it stays free.
Free feeds and data
/performance/feed.xml- RSS of settled trades, scored forecasts and daily decision digests./performance/feed.json- the same items as JSON Feed 1.1./api/public/performance- the whole record as JSON, free and read-only.