Apex Lake Request a demo

Systematic trading platform

From backtest
to live trading.

Apex Lake gives quant teams one place to test strategies on market history, deploy the same code, and compare live fills with the backtest.

Explore the Research Lab

Engine benchmark

One year of market history

Nov 2025 – Oct 2026

Median backtest time
30 sec
Perpetual markets
4
One-second bars available
126 million

The platform

Research, deploy, reconcile.

Follow a strategy from its first out-of-sample test to the orders your venue actually fills.

Research lab

Explore variants, stress assumptions, and test your selections on unseen market periods.

Strategy runtime

Backtest and deploy the same strategy code through a shared event-driven trading API.

Live reconciliation

Replay a live window against history. Trace differences in entries, exits, and fills back to individual trades.

Inside the research lab

Which signal holds up after fees?

Use the lab’s coding agent to compare strategies on past data, then test your selection on the next unseen period.

  • Compare signals
  • Test across markets
  • Stress-test costs
  • Test allocations

Momentum scenario

Does a slower signal hold up?

20-bar baseline vs. 60-bar candidate over one year
Signal comparison
+4 bpAverage gross return rose from 9 to 13 bp per trade.
Market breadth
3 of 4BTC, ETH and SOL beat the baseline. XRP lagged.
Break-even fee
6.5 bp / fillAbove this fee on entry and exit, the candidate loses its edge.
Portfolio drawdown
4.1%A 50/50 mix cut the worst decline from the candidate’s 6.2%.

Coding agent

Build and test strategies with an agent.

Describe the change or the question in plain language. The agent works directly with your strategy code, experiments, and run history.

  • Write strategy code

    “Add a volatility filter to my entry rule. Keep position sizing fixed.”

    The agent edits the TypeScript strategy and runs its code checks.

  • Test a change

    “Backtest the filter against the original with the same fees.”

    The agent runs both versions and compares trade count, net return, and drawdown.

  • Run autoresearch

    “Try to reduce turnover. Stop after five experiments.”

    The agent chooses the next experiment from prior results and records what it kept or rejected.

  • Debug a live strategy

    “The backtest bought ETH here. Why is the live fill missing?”

    The agent traces the running release and compares the replay with the worker’s orders and venue responses.

Architecture

One strategy.
Two environments.

Run the same strategy on historical and live data. Compare simulated fills with the fills your venue reports.

Data & connectivity

Historical data · Binance, Bybit, OKX

Live execution · Binance, Interactive Brokers

Fig. 02  /  Strategy execution paths Compare trades

Live reconciliation

Where did live trading diverge?

Line up replay and live trades. See where fills matched and trace a missing trade back to the venue’s response.

ETH/USDTSession time · UTC

Matched fill.

Both bought 1 ETH at 3,100.00 USDT within the same second.

Replay fill
Live fill

No live match.

At 09:32, the venue rejected the buy order during maintenance.

Replay fill
1 ETH at 3,106.00 USDT
Live fill
No fill

Portfolio operations

Inside the live book.

Follow each live strategy back to its release, its portfolio, and the capital behind it.

Verify the live release
Compare the strategy revision and engine build you deployed with the versions running live.
Track portfolio capital
Trace each portfolio to its capital owners and wallets. See how their balances have changed.
Investigate a quiet strategy
Check when its worker last reported, whether it’s ready to trade, and whether its balance checks passed.
Portfolio overview

Portfolio equity

20,528.04 USDC

Initial allocation
20,076.98 USDC
Above allocation
+451.06 USDC

Service health

14:34:42
Trading worker
Reporting
Wallet exposure
Matched

Get in touch

Walk through a strategy with us.

See a research run, its deployed release, and the fills it produced.

Request a demo