How to use Algo Arena

Algo Arena runs a House of trading bots that trade a shared, simulated USD book on live market data, around the clock. Each bot follows a different strategy, and a live leaderboard ranks them by hypothetical risk-adjusted return.

You can watch the House compete, build a fund from its strategies, design your own, and replay ideas over real market history — all without risking a cent. It is built for learning how automated trading behaves, not for handling real money.

• —Hypothetical only — no real money is ever traded, deposited or withdrawn.
• —All values are shown in USD, for illustration.
• —Nothing here is financial, investment or tax advice.

Create an account to take the controls. On first sign-in you confirm a short set of disclosures — that Algo Arena is a simulation and not financial advice — and the dashboard unlocks.

Your data lives inside a workspace, and workspaces can have several members. Switch between them from the top bar.

• —Roles: viewer (watch), operator (run controls), admin (manage members).
• —You can export or erase your personal data any time from Account & data.

The dashboard opens on the Arena Overview — total equity, average return, active bots and trade count across the House, with a performance chart and the live leaderboard underneath.

• —Timeframes: 1 hour, 24 hours, 7 days, 30 days and all-time.
• —Jump to Compare, export a CSV, or open a printable report from here.

The Arena pits funds against the House in tournaments. The House is the benchmark to beat — a smart portfolio that actively selects the strongest strategies, drops the laggards, caps how much any single one can hold, and rebalances as the market moves, all on the same simulated capital as everyone else, so the comparison is fair.

Standings update continuously as the market moves. Watch which approaches pull ahead and how they hold up when conditions change.

• —Anyone can watch the arena and its live standings without signing in — you only need an account to enter your own fund.

Trust in the standings starts with knowing the rules of the simulation. Every combatant — House, dark horse and player fund alike — trades under the same fill model: each simulated trade pays a taker fee, suffers adverse slippage, and crosses a synthetic bid/ask spread that widens with market volatility, so hypothetical returns are not flattered by free, frictionless fills.

The exact figures below are published live from the engine's own configuration — the same constants the simulation runs on — so this page can never drift from what actually happens in the arena.

• —Sharpe ratios are shown only once a fund has at least 24 hours of history, and are always labeled with the window they cover.
• —The composite Score percentile-ranks each fund against the field on four weighted measures: risk-adjusted return, return, drawdown and volatility.

Beyond the House and the player funds, the leaderboard carries a third force: a dark horse called RONIN. It trades its own single strategy on the same simulated capital and competes for rank like everyone else, yet belongs to neither the House portfolio nor the player pool.

Watch for the dark-horse row on the standings. It is there to keep both the House and your own fund honest — another independent yardstick to measure a strategy against.

Alongside the House and RONIN, the leaderboard pins a fourth marker: ATLAS. It holds an equal slice of every strategy, never drops a laggard, and never rebalances — the simplest possible 'own it all' portfolio, on the same simulated capital.

ATLAS is the passive yardstick. If the House's active selection is earning its keep, it should out-perform ATLAS; and a fund that cannot beat even the naive equal-weight index has real work to do.

Your fund is your entry into the competition: pick the strategies you want to back, set how capital is split between them, and let it trade the live simulation alongside the House.

• —Everyone starts from the same simulated capital — your placing reflects strategy and allocation choices, not who deposited more.

Seasons give the competition a rhythm — each one is a defined window with its own leaderboard, so newcomers start on a level footing rather than facing an all-time scoreboard.

Track the current season's status and standings, and review how previous seasons played out.

The builder lets you design a custom strategy by combining signals with entry and exit rules, then save it as a bot you can run and rank against the House.

• —Start from a simple idea — a moving-average cross, a momentum filter — and iterate.
• —Save versions, and archive the ones you have moved on from.

The library is the catalogue of available strategies — trend-followers, mean-reversion, momentum and AI learning bots — each with a plain-language description of how it thinks.

Use it to understand a strategy before you add it to your fund, or to find inspiration for the builder.

The leaderboard ranks bots by hypothetical risk-adjusted performance, updating in real time as the simulation trades.

• —Columns: return, win rate, maximum drawdown, Sharpe ratio, trade count, a health badge and a recent-trend sparkline.
• —A signed-out version is visible on the landing page, so anyone can watch the arena before joining.

Compare overlays several bots on one chart and lines up their key metrics, so you can see which strategy suits which conditions.

It is the quickest way to decide what belongs in your fund and what to cut.

The regime radar classifies current conditions — trend, chop and volatility — from live data, so you can see which strategies are likely to shine right now and which may struggle.

• —The app surfaces an alert when the regime flips, so you are not caught out.
• —The leaderboard highlights bots that suit the current regime.

The allocator shows a target split of capital across strategies, blending each one's track record with the live regime and health signals into a single read-model.

History is reconstructed from point-in-time factors, so what you see for a past moment reflects what was known then — not today's hindsight.

The backtest lab runs a strategy against real market history and reports how it would have performed, so you can pressure-test an idea before committing it to your fund.

The standing caveat applies: simulated and historical results never predict future outcomes. Real markets add fees, slippage and latency a backtest cannot fully reproduce.

Open any bot to see its full history: equity curve, trades, win rate and risk metrics. It is the best place to understand why a bot sits where it does on the leaderboard.

For AI learning bots, the detail view also tells the story of the Deep Q-Network behind it — how its decisions have evolved.

From a bot's detail page or the leaderboard you can open a clean, printable report: one page laying out the headline metrics, equity curve and standings as of the moment you open it.

Each report is generated on demand and stamped with the time it was produced, so a shared or printed copy is never ambiguous — still hypothetical figures, still illustrative USD.

The risk centre models the controls a professional trading desk would use: per-bot and arena-wide loss caps, a kill switch that halts trading, and realistic execution with fees and slippage.

• —Safeguards are latched and durable — a tripped loss cap or kill switch stays in force across a restart until it is deliberately re-enabled.

Readiness summarises governance state: approvals and gates that would matter before any real-money path could ever be considered. It is purely informational.

Nothing on this page enables real trading. Algo Arena does not hold a financial services licence and never moves real funds.

The House can take cues from broad, aggregated signals across the field to inform its decisions, summarised transparently rather than copying any one participant.

Under strict real-data mode the bots only trade and learn on real market data; pairs without a healthy live feed are paused instead of guessed.

The alerts bell in the top bar collects notable events — a regime flip, a risk control tripping, notable moves — so you can catch them without watching the screen all day.

Unread counts are kept per browser, so the badge survives a reload.

Operators and admins can review an audit trail of sensitive actions — who changed what, and when — with before-and-after snapshots.

The trail is append-only and retained for an accountability window, even after personal data is erased (the actor is anonymised).

From Account & data you can export the personal data Algo Arena holds, or request erasure of it. Erasure removes your profile, memberships, devices and invitations.

For the full picture, read the Privacy Policy, Terms of Use and Risk Disclosure — they are linked in the footer of every page.