Peregrine Kapmere trading desk displaying live market data and predictive model output

Systematic trading infrastructure

Backtested logic, deployed into live execution

Peregrine Kapmere monitors global markets around the clock and acts on signals generated by models validated against historical data, without emotional interference.

Get Started View Sample Strategy

Systematic Edge

Models built on structured and alternative data

Peregrine Kapmere ingests pricing data, order book depth, and alternative data sources, then applies historical pattern matching to generate trade candidates. Every signal passes through a multi-factor model before it reaches your execution queue.

Multi-factor models weigh momentum, volatility, and liquidity conditions before ranking candidates.
Low-latency data pipelines reduce the gap between signal generation and order placement.
Risk-parity position sizing adjusts exposure based on asset volatility, not fixed capital splits.
Pattern matching references comparable historical regimes rather than static thresholds.

Technical specification

Data ingestion frequency: Continuous, tick-level
Signal refresh interval: Sub-second
Model type: Multi-factor, ensemble
Position sizing: Risk-parity weighted
Supported asset classes: Equities, FX, futures
Execution latency target: Sub-100ms venue dependent

Validation methodology

Walk-forward testing, not curve-fitted backtests

Each model is trained on a historical window, then tested on a subsequent unseen period before any parameter is finalised. The window rolls forward repeatedly, so the model is never evaluated on the same data it was tuned against. This reduces the risk of curve-fitting, though it does not remove it entirely.

Training window

12 months

Out-of-sample window

3 months

Re-optimisation cycle

Rolling, monthly

Model approval

Manual review required

Past performance, whether simulated or actual, is not a reliable indicator of future results. All trading involves risk of loss, and backtested results are subject to assumptions about slippage, fill rates, and market impact that may not hold in live conditions. Peregrine Kapmere does not guarantee any specific outcome.

Workflow

From screening to order fill, without manual steps

Automated screening

The platform scans configured markets continuously. It filters candidates against your model's criteria, discarding anything outside defined risk or liquidity bounds before a signal is ever shown to you.

Signal verification

Every candidate is cross-checked against current volatility and correlation data. Signals that no longer match the conditions they were generated under are rejected automatically, cutting stale or conflicting trades.

Direct execution

Approved orders route through your connected brokerage API. There is no copy-paste step and no re-entry of order parameters, which removes a common source of manual error.

Common questions

Technical and operational detail

How much lag exists between a data update and a signal?

Data refreshes continuously and signal generation runs on a sub-second cycle. Total lag depends on your venue's own feed latency and your network connection to Peregrine Kapmere's infrastructure, which we do not control.

What happens with slippage on live orders?

Backtests apply a configurable slippage assumption, but live fills can differ, particularly in low-liquidity instruments. We log realised slippage against the backtest assumption so you can audit the gap directly.

What hardware or connection is required?

Peregrine Kapmere runs server-side. A standard broadband connection is sufficient for monitoring and configuration; no local GPU or dedicated trading rig is required.

How are subscription tiers structured?

Tiers differ by the number of concurrent models and asset classes you can run, and by access to historical data depth for backtesting. Full pricing is provided during onboarding, based on your intended usage.

Which brokerage APIs are supported?

Integration depends on the broker's available API and your account's execution permissions. We confirm compatibility with your specific brokerage during setup before any live connection is made.