—
No chart data available.
Watchlist
Symbol
Rank
Last
Chg
Chg%
Simulated on daily bars at a $10,000 position per signal, net of trading costs — a commission-free broker, 5bps of adverse slippage on every fill, and the SEC/FINRA fees charged on sales. Financing and borrow are not modelled on long positions. Scored tickers are the S&P 500 as it stands today, so names that left the index — or failed — are absent, and every ticker measured is one that lasted; this flatters results across the universe. The rules were also selected using this same history, so it is not an out-of-sample test. Past results are not indicative of future returns. Informational only — not investment advice.
Tell us what went wrong, or what you would like changed. It goes straight to the ALGORA1 team.
A deployed portfolio places orders on its own when its rules fire, inside the limits you set — no conversation and no credit. It is checked once each trading day after the close, and every order it places is written to the activity log. Nothing trades until you deploy it, and you can undeploy at any time.
Whatever authorises the connection is encrypted before it is stored and is never sent to your browser. Disconnecting deletes it, and a sign-in can also be revoked from your broker. Activity is recorded in a log you cannot edit — open it from the agent window.
Short selling is never permitted by these guardrails. Limits are enforced by the terminal against your live account, not by the model — an order that breaks one cannot be sent, whatever the agent proposes.
News and leaderboard coverage are saved to an account — create a free account to choose them.
A narrow focus can lower the odds a strategy passes and is published.
Each round, the research model writes this many different strategies for the ticker, and every one is backtested on its own.
More per round gives each round more chances of finding one that passes. It costs more, though not in proportion — most of a round is the model reading the market, not writing each candidate.
Every candidate faces the same tests, and the bar rises a little as more are tested, so doubling the candidates does not double the chance of a pass.
Every message to the algorithm builder and the brokerage agent is charged on what it actually cost to run, so the amount varies. Your plan tops this balance up each period and unused credit rolls over.
Payments and invoices are handled by our payment processor; we never see or store your card details. Cancelling stops the next renewal and keeps your access until the end of the period you have paid for — your account, watchlist, saved algorithms and remaining credit are kept. See Legal & Privacy §3.
On the free month you can still select any published strategy from the research desk and chart, backtest and deploy it — no AI needed.
Every message is charged on what it actually cost to run — including one that only asks a question, or that produces a strategy the engine can't express. Running your strategies costs nothing. Credit pays for conversation with the builder and the agent; once a strategy is saved, the terminal scoring it, deploying it and acting on it every day draws no credit at all. Beyond what your plan includes there is no free allowance for messages. Your remaining balance is shown beside the composer at all times.
Your plan tops your balance up every month. A credit pack is a one-time purchase on top of that; both roll over and do not expire while your account is active. Credit has no cash value and is not refundable or transferable once spent. Building an algorithm sends the text you submit to our AI provider — see Legal & Privacy §5.5.
Each file is one sleeve, trades only its own ticker, and carries the strategy selected for that ticker — sleeves need not run the same one. Hand several files to the assistant together to layer them — the weights below are what each sleeve may draw from your base capital — the account equity recorded once, before the first order. Any figures shown here are hypothetical results computed from past data, not a forecast, and exclude commissions, slippage and taxes. Automated trading carries risk; you are responsible for anything run on your live account.
These limits belong to the ACCOUNT, not to one allocation — saving replaces whatever your guardrails say now, and the terminal enforces them against your live account on every order the agent proposes, whichever strategy it came from. Both are off by default. With no halt set the agent keeps placing orders through a drawdown of any size. A halt is reduce-only: for the rest of that day it refuses anything that opens or adds to a position, while closing orders still go through — it will not close a position for you. The stop loss is the one that does, measured from each position's own entry price. A gap through the level fills at the next open, so a stopped trade can still lose more than the number you set. Note the difference in how the two are applied: the halt is enforced at the gate, so an order that breaches it cannot be sent; the stop is carried out by the agent on each run, because closing a position means placing an order and a guardrail can only refuse them. The suggested value is derived from the weights on screen and from price history only; it is not a forecast, and no limit can prevent a loss.