Clearly defined rules.
We turn your idea into clear rules covering markets, signals, entries, exits and position sizing. Every assumption is documented.
BESPOKE ALGORITHMS · QUANTITATIVE RESEARCH
Engineering
the Edge.
Turn your trading idea into a strategy you can test. We build bespoke trading algorithms and put every assumption through rigorous quantitative validation.
Bespoke research and development, delivered by StratEdgics. We do the research for you; we do not sell access to our research software.
STRATEDGICS / RESEARCH PRINCIPLE
An edge is not a promise.
It is put to the test.
FROM HYPOTHESIS TO EVIDENCE01 — EXPLORE ↓
MULTI-SCENARIO STRESS TESTING•MONTE CARLO•DEFLATED SHARPE RATIO•ZERO OVERFITTING•WALK-FORWARD•CPCV WITH EMBARGO•
01 / QUANTITATIVE ENGINEERING
Your strategy.
Our discipline.
From bespoke development to validation: a research service built around your trading logic, risk and constraints.
Statistical scrutiny.
We challenge results against held-out data, execution costs and adverse scenarios, within the agreed scope.
Evidence to guide your decision.
You receive a developed strategy, with a clear explanation of its strengths, weaknesses and the conditions it was designed for.
OUR CAPABILITIES / FROM IDEA TO MODEL
Imagine your strategy.
We turn it into trading rules.
A basic signal or a multi-source approach: we draw on an extensive research environment to formalise, develop and test your project. You bring the idea; we select the tools that serve it.
EXAMPLE PROJECT TO INVESTIGATE
“Enter after an opening-range breakout, only in the morning.”
01Price & time
02Opening range
03Entry condition
04Stop & exit
We define the session, range calculation, breakout threshold and exit rules. The intuition becomes logic that can be coded and tested.
To examine: data quality, false breakouts, costs and out-of-sample stability.
THE TOOLS BEHIND YOUR PROJECT
A complete research workflow.
Explore ten areas of expertise. We choose the tools that address your research question, then test the value each one adds.
01Strategy design & development
How do we turn intuition into rules?
Strategy design & development
How do we turn intuition into rules?- Visual strategy building & conditional logic
- Combine entries, exits, time filters and market conditions to describe your decision precisely.
- Custom indicators & formulas
- Combine price and flow indicators with quantitative functions; define a measure specific to your idea.
- Time-series analysis
- Use z-score normalisation, Kalman filters, the Hurst exponent, transfer entropy or frequency analysis, depending on the research question.
- Working with existing strategies
- Assess whether existing Python, MQL5, Pine Script, C++ or NinjaScript code can be imported. We review compatibility and any necessary changes before work begins.
FOR YOUR PROJECT
Explicit rules, an algorithm and traceable assumptions.
03Price, volume & microstructure
What is driving the price move?
Price, volume & microstructure
What is driving the price move?- Multi-asset historical data
- Study futures, equities, forex, crypto and metals, depending on the historical data available and the scope of your strategy.
- Order book & depth
- Observe spread, imbalances, displayed liquidity, walls and absorption to formalise entry or exclusion conditions.
- Aggressive order flow & block trades
- Analyse trades, aggressive order flow and block trades, then test how the market behaves after a signal is detected.
- Three-dimensional volume
- Read price, time and traded volume together to investigate activity beyond candles alone.
- Detection assessment
- Check effective sample size and false positives, including false discovery rate controls, before incorporating a detected signal.
FOR YOUR PROJECT
Defined market signals and an assessment of their contribution.
05Alternative data & narratives
Does information beyond price add value?
Alternative data & narratives
Does information beyond price add value?- On-chain flows
- Compare observable blockchain flows and large-participant positioning with market behaviour.
- Timestamped speeches & announcements
- Investigate speeches, earnings releases and news, and measure price reactions around publication.
- Sentiment & crowd behaviour
- Turn a sentiment or public-positioning hypothesis into variables that can be compared with prices.
- Weather, satellite & real-world economic data
- Test exogenous series and forecast surprises when relevant to the asset and historically available.
FOR YOUR PROJECT
Time-aligned series and an assessment of their usefulness.
07Backtests & statistical validation
Does the result withstand scrutiny?
Backtests & statistical validation
Does the result withstand scrutiny?- Historical & trade analysis
- Connect results with trades, rules and calculation assumptions to understand strategy behaviour.
- Out-of-sample & walk-forward
- Separate development from evaluation, then repeat tests over time on held-out periods.
- Purged cross-validation
- Use combinatorial purged cross-validation (CPCV), where appropriate, to limit contamination between training and testing.
- Overfitting & multiple testing
- Examine the Deflated Sharpe Ratio, probability of backtest overfitting and the number of trials leading to selection.
- Monte Carlo & sensitivity
- Resample returns and vary parameters, assets, horizons and costs to identify weaknesses.
- Costs & decision
- Include execution costs in the protocol and conclude: proceed, observe in simulation, rework or reject the candidate.
FOR YOUR PROJECT
A robustness report with criteria, conclusions and limitations.
09Execution & monitoring
How can the model be put into practice with appropriate controls?
Execution & monitoring
How can the model be put into practice with appropriate controls?- Exports & integration
- Prepare export formats and protection for the target environment. Broker, platform and compatibility are confirmed during scoping.
- Execution limits
- Define the risk framework and the conditions that allow or block a position.
- Actual vs expected performance
- Compare actual behaviour with the research baseline and set alerts for deviations when ongoing monitoring is part of the project.
- Research workflow & review checkpoints
- Coordinate data collection, research, validation and reporting, with progress tracking and human review at agreed decision points.
FOR YOUR PROJECT
Integration and monitoring arrangements agreed before delivery.
02Research & optimisation
Which variants warrant further investigation?
Research & optimisation
Which variants warrant further investigation?- Hypothesis generation
- Explore combinations of signals and parameters from an idea, an asset list or a defined objective.
- Multi-asset & multi-horizon research
- Compare configurations across instruments and timeframes, then examine stability between contexts.
- Optimisation & confluence
- Evaluate parameters and agreement between signals; measure whether a filter adds information.
- Goal-driven research
- Plan successive rounds of research within an agreed budget, track every trial, document selection decisions and reserve data for validation.
FOR YOUR PROJECT
Ranked candidates and a record of the trials performed.
04Options & macro context
Under what conditions is the signal useful?
Options & macro context
Under what conditions is the signal useful?- Options-dealer exposure
- Analyse gamma exposure by strike, sign changes and put/call concentrations to contextualise price moves.
- Options research & quoting models
- Investigate options and, for relevant projects, quoting and inventory models such as Avellaneda–Stoikov.
- Regimes, factors & positioning
- Place signals in macroeconomic and positioning context, including Commitments of Traders reports.
- Seasonality
- Examine recurring time patterns and their stability before turning them into strategy rules.
FOR YOUR PROJECT
Context filters and testable regime hypotheses.
06Signal fusion & learning
How can different sources of information work together?
Signal fusion & learning
How can different sources of information work together?- Weighted consensus
- Combine the outputs of several models, weighting each contribution by its measured reliability.
- Machine-learning research
- Use machine-learning and screening tools suited to your project. We select the models, variables and testing process to match the research question.
- Change-point detection
- Track signal deterioration, including CUSUM-style cumulative controls, to trigger reassessment.
- Signal longevity
- Study configuration survival, including Kaplan–Meier curves, and exclude engines whose validity deteriorates.
FOR YOUR PROJECT
An explainable decision process and deactivation criteria.
08Risk & prop firm challenges
What risk and budget does your objective require?
Risk & prop firm challenges
What risk and budget does your objective require?- Position sizing & expectancy
- Study expectancy and sizing approaches, including Kelly, with drawdown constraints and model limitations in view.
- Challenge convexity
- Seek a strategy compatible with the selected programme’s targets, loss limits and rules, while measuring failure costs.
- From purchase to payout
- Simulate pass rates, repeat attempts, the cost of reaching a funded account and the likelihood of a first payout. Explore the process in the prop firm section.
FOR YOUR PROJECT
A risk framework and quantified scenarios for your objective.
10Research tools & documentation
How do we make the research clear and reusable?
Research tools & documentation
How do we make the research clear and reusable?- Terminal, signals & indicators
- Use terminal, signal-generation and indicator-tracking utilities to organise analysis work.
- Python tools
- Use a Python toolkit to extend export analysis and structure work specific to your project.
- Methodological resources
- Technical learning paths, exercises and references support our research environment. Your engagement covers development, analysis and agreed deliverables.
FOR YOUR PROJECT
Clear reports and reusable research materials, as agreed in the project scope.
EXTENSIVE TOOLS. A TAILORED ENGAGEMENT.
We use these capabilities to deliver your project. Specialised data may require dedicated access; availability, selected tests and deployment environment are confirmed before work begins.
Ambitious ideas need clearly defined models.
We work with simple ideas and complex projects alike. Feasibility depends on whether the rules can be clearly defined, the data is available and the target platform can support them.02 / OUR METHODS IN ACTION
A backtest is not enough.
Put the idea to the test.
Explore our quantitative research process. Each view addresses a practical question: which ideas deserve further work, how to test them, what could go wrong and when to reassess.
INTERACTIVE RESEARCH DEMONSTRATION
Selection, walk-forward testing, Monte Carlo and monitoring.
Illustrative simulations using fictional data, not performance forecasts.
WHAT WE LOOK FOR
Robust results beyond a single best setting.
A strategy that works only at one precise setting may be fragile. We test nearby settings too: if a small adjustment wipes out the result, that is a warning sign. Consistent results across a range of settings suggest stability, which we then check on held-out periods.
Realistic data and assumptions
We check historical data quality, what was known at each point in time, commissions and slippage so that tests reflect real market constraints.
Cross-validation
Depending on the project, time-based splits and combinatorial purged cross-validation (CPCV) help limit information leakage between training and testing.
Multiple-testing correction
Testing many variants raises the chance of a lucky result. The Deflated Sharpe Ratio helps put apparent performance into perspective.
Risk & sensitivity
Monte Carlo resampling, parameter variation, regime changes and rising costs: we look for breaking points.
The protocol is tailored to each engagement. A statistical test provides evidence, never a performance guarantee.
WHAT THE RESEARCH TELLS US
A clear next step. Even when it means stopping.
We explain what the tests mean and recommend a next step. Strong historical performance alone is not enough to justify live trading.
Proceed
Agreed criteria are met. Assess implementation and its controls in the target environment.
Monitor in simulation
Monitor the strategy on new data without risking real capital, while recognising the limits of simulated execution.
Rework
An identifiable weakness calls for revised rules or risk settings, followed by fresh validation.
Reject
The evidence does not justify proceeding. The reasons for rejection form part of the deliverable.
Criteria, tests and any simulation or monitoring stages are defined in the engagement scope. No conclusion guarantees future results.
03 / PROPFIRM ENGINEERING
Build for a funded account.
Simulate through to payout.
We develop strategies and algorithms around the asymmetric risk and reward of prop firm challenges: pursuing a funded account while accounting for the cost of failed attempts, challenge rules and payout conditions.
THE OBJECTIVE
Maximise the probability of securing a funded account, within a defined purchase budget and risk constraints.
An optimisation objective, never a promise of qualification.INTERACTIVE RESEARCH DEMONSTRATION
Explore the path from challenge to payout.
Illustrative simulations using fictional data, not performance forecasts.
Optimise the journey, not just returns.
Entry costs, retries and conditional access to payouts create an asymmetry. We study sizing, losing sequences and breach risk to seek the strategy best suited to this objective.
Know the numbers before you enter.
Compare strategies, estimated pass rates, the number of attempts needed, the cost of reaching a funded account and the likelihood of a payout. The report also covers failure scenarios and the limits of the analysis.
04 / OUR SERVICES
What do you need?
Let’s start where you are.
An idea to structure, a strategy to validate or an algorithm to build: choose the service that fits your project. These needs can be combined within one engagement.
Turning an idea into trading rules
You have a trading idea.
We turn your reasoning into precise rules: markets, signals, entries, exits and risk management. We identify the data required and questions to resolve before coding.
WHAT YOU RECEIVE
Explore a sample deliverable This is what I need Documented trading rules and a specification to prepare for development.
Backtesting and validation
You want to put a strategy to the test.
We assess your strategy on historical data while accounting for execution costs. Held-out testing, parameter sensitivity and risk scenarios are selected to suit your project.
WHAT YOU RECEIVE
Explore a sample deliverable This is what I need A test report, identified weaknesses and an evidence-based recommendation for next steps.
Algorithm development
Your rules are ready to be coded.
We develop your strategy from defined rules or existing code. Signals, position sizing and execution constraints are implemented for the agreed platform and scope.
WHAT YOU RECEIVE
Explore a sample deliverable This is what I need An algorithm built to the specification, functional checks and documentation.
WHAT DOES YOUR DELIVERABLE LOOK LIKE?
See the detail before committing.
Explore three fictional examples. The content and delivery formats are agreed with you for your engagement.
SE / SPEC-001 · V1.0
Specification & formalised rules
Example scoping document · M15 breakout strategy
- Objective
- Specify a long breakout strategy that can be tested before deployment.
- Market & data
- BTC/USD · M15 candles · bid/ask data timestamped in UTC. Data history and quality to be validated.
- Trading window
- Entries from 08:00 to 16:00 UTC. One open position at a time. Mandatory exit at 16:45 UTC.
- Entry signal
- At candle close, price exceeds the highest high of the previous 20 candles. The signal candle is excluded from the calculation.
- Execution
- Buy at the first available ask after the signal. Reject entry if spread exceeds 15 USD or data is more than 5 seconds stale.
- Exits & sizing
- Stop 200 USD below entry, target 300 USD above. Theoretical risk of 0.5% of equity at the stop, before fees and slippage. Round BTC quantity down to the platform quantity step; skip orders if the minimum size exceeds the intended risk.
- Safeguards
- Suspend new entries at a realised daily loss of 1.5% of opening equity. Resume the next day. Log every rejection and reconnection.
Acceptance criteria
- No signal accesses future data.
- Two simultaneous signals do not create duplicate orders.
- Stops, sizing and trading hours follow the rules.
- A reconnection reconciles positions before resuming.
To agree before development
Broker and platform, data source, account currency and risk conversion, actual costs, delivery formats and code ownership.
This example covers specification. Development, validation and monitoring are scoped separately.
Have a different need or still working it out? Tell us about your situation and we’ll define the scope together.
Something else05 / FREQUENTLY ASKED QUESTIONS
Before we
begin.
Do I need to buy or learn software?
No. We handle the research, development and validation for you, then explain the findings. We do not sell access to our research software. We agree any data access and trading platform requirements with you before work begins.
What do I actually receive?
Development and quantitative analysis tailored to your project, including clearly defined rules, an algorithm, a test report and documentation. We agree the scope, delivery formats, code ownership and any ongoing monitoring before work begins.
Can we start from my own strategy?
Yes. We can start from your idea, existing rules or existing code. The initial discussion establishes feasibility, required data and the target platform.
Do the animations show real results?
No. These charts are educational demonstrations using fictional data. They illustrate methods and represent neither client performance, StratEdgics results nor a live market feed.
Does every method apply to my project?
The protocol depends on your strategy, data frequency, available history and execution constraints. Selected tests and their limitations are specified in the project scope.
Does validation guarantee profitability?
No. It helps identify weaknesses and quantify certain uncertainties. Markets change; a backtest or simulation never guarantees future results.
What are the timelines and fees?
We provide a timeline and quote once we understand your project. These depend on the complexity of your strategy, the data required and the agreed deliverables. We also confirm which markets and platforms we can support.
06 / YOUR NEXT STEP
It starts
with a hypothesis.
An idea to formalise, a strategy to engineer or results to challenge? Describe your project.
Prefer to talk? Choose your time slot in 2 minutes.
Book a callTell us about your strategy, your objectives and your constraints.