Strategy Arena
Build crypto strategies node by node, import Pine Script v6, backtest with fees and paper trade on live order books.
Explore Strategy ArenaAPEX OS / QUANT LABS + COMPUTE FABRIC
Crypto quant research
and backtesting, built in.
Quant Labs run 47 WebAssembly kernels on the market data Apex OS records: volatility and regime analysis, Monte Carlo simulation, backtests with overfitting statistics, portfolio construction and factor research. No notebooks to maintain and no data to download.
- Kernels
- 47 WebAssembly quant kernels
- Included from
- Starter plan, $129 per month
- Matrix size
- Up to 120 markets × 1,000 bars
- Backtests
- Out-of-sample, deflated Sharpe
- Simulation
- Monte Carlo Studio
- GPU research
- Neural Strategy Forge on request
What are Quant Labs?
Quant Labs are the research apps of Apex OS. They run on Compute Fabric, a WebAssembly quant core with worker threads, inside your browser, on the candles and order books Apex OS records from your connected exchanges.
Each lab answers one question a quant desk would ask before it risks money: how volatile is this market and in which regime, how likely is this target before that stop, does this backtest survive out-of-sample, and how do these positions behave together under stress?
Volatility, regimes and market structure
- Quant Lab: volatility, persistence, cycles, GARCH and hidden regimes of one market
- Forecast Lab: GPU-trained volatility and direction forecasts, scored against an EWMA baseline
- Market Topology: correlation networks, clusters, systemic concentration and lead-lag
- Market Chronicle: observed changes and private server-captured checkpoints
Backtest Foundry: tests that admit overfitting
Try enough parameter combinations and one will look brilliant by chance. Backtest Foundry runs parameter sweeps and reports what most backtesters leave out: out-of-sample performance and overfitting statistics, including the deflated Sharpe ratio, which discounts a result for the number of variants you tried.
Monte Carlo Studio
Monte Carlo Studio simulates thousands of price paths to estimate bracket probabilities (how likely a take-profit is hit before a stop-loss), draws fan charts of possible outcomes and simulates execution costs, so a trade plan starts from odds instead of hope.
Stress before size
- Basket Lab: hypothetical baskets, concentration and after-cost stress scenarios
- Liquidity Stress Lab: order-size sweeps, partial fills and depth contraction
- Execution Lab: depth, fees and estimated execution costs
Portfolio and factor research
- Portfolio Lab: shrinkage covariance, six allocators, the efficient frontier and bootstrap risk
- Factor Lab: cross-sectional rank IC, quantile spreads and adjusted t-statistics
- Correlation Lab: aligned correlation matrices from recorded returns
- Pair Observatory: aligned closed returns, beta and causal divergence
Neural Strategy Forge: GPU strategy evolution
For research at scale, Neural Strategy Forge evolves strategy candidates on GPU, replays them in a server-side queue, ranks them and imports the best as inactive drafts into Blueprint Studio, where they face the same Shadow and Paper tests as any other strategy. Strategy Forge is available on request.
Compute Fabric limits per plan
Compute Fabric runs the quant kernels on matrices of markets and bars. Larger plans run larger matrices.
| Plan | Markets per matrix | Bars per market |
|---|---|---|
| Starter | 40 | 500 |
| Trader | 60 | 1,000 |
| Premium | 120 | 1,000 |
Plans with Quant Labs
Prices per month in USD, excl. VAT. Quarterly billing saves 10%, yearly billing 20%. The pricing page also shows EUR at the live exchange rate.
Starter$129 / monthQuant Lab, Forecast Lab, Market Topology, Market Chronicle and Compute Fabric at 40 markets × 500 bars.
Trader$249 / monthMonte Carlo Studio, Backtest Foundry, Basket Lab, Liquidity Stress Lab, Execution Lab and 60 markets × 1,000 bars.
Premium$495 / monthPortfolio Lab, Factor Lab, Correlation Lab, Pair Observatory and 120 markets × 1,000 bars.
Frequently asked questions
What is a deflated Sharpe ratio?
A Sharpe ratio corrected for how many strategy variants were tried and for non-normal returns. It lowers the score of results that are likely to be luck from testing many combinations, which makes it a better filter for backtests.
Do I need Python or my own market data?
No. The labs run in your browser on WebAssembly and use the market data Apex OS records from your connected exchanges.
Which plan includes Monte Carlo simulation and backtesting?
The Trader plan, from $249 per month: Monte Carlo Studio, Backtest Foundry, Basket Lab and Liquidity Stress Lab. Strategy replays with fees are part of Strategy Arena on the same plan.
Can I use my own GPU?
Yes. Forecast Lab and Neural Grid use models on a paired GPU PC. Neural Strategy Forge, for GPU strategy evolution, is available on request.
Research like a quant desk.
Request access and we show you the labs on the markets you trade.
Trading crypto-assets involves a high risk of loss. Apex is software, not investment advice.