01
Out-of-sample data
Part of the history is set aside before anything is fitted, and used once, at the end.
Home / Trading
Fig. 03 · For traders and investors
BrightDragon Trading
Trading and investment tools · Automations · Strategy testing
Software and calculators for traders and investors, and strategy testing and automation for people who bring their own rules.
Important
We do not give financial advice, and we make no promises about returns. Our tools and calculators do not recommend trades, give signals or tell you what to buy or sell. They work only from the figures and rules you give them. We test and automate the strategy you bring us. Whether to trade it, and whatever it does in the market, is your decision and your risk.
BrightDragon Systems Ltd provides software development and testing services only. Nothing on this page is investment advice, a recommendation, or an invitation to trade. We are not authorised or regulated by the Financial Conduct Authority and we do not provide regulated financial services.
Trading carries risk and you can lose money. Any decision to trade, and any strategy the software implements, is entirely yours. No software can guarantee an outcome, and past testing results are not a reliable indicator of future results.
Shop
Tools that do your own arithmetic and give no advice. Each one goes on sale here, with its price, when it is ready.
Also in planning: MetaTrader utilities for the routine parts of placing and managing a trade, which act only on your settings, and Python tools for people who trade shares outside MetaTrader. See every product.
We describe what each tool does. We never publish performance figures, win rates or backtest results for any of them.
Strategy testing and automation
A paid service, taken on by enquiry. We turn your trading rules into code, test them against historical data under realistic conditions, and automate them if you decide to go further. We work in MQL5 for MetaTrader, and in Python for shares and other markets.
What we do
- Turn your strategy into an unambiguous written rule set
- Code those rules and test them against historical data
- Separate in-sample from out-of-sample, and run walk-forward analysis
- Model spread, commission and slippage the way they actually land
- Test how sensitive the result is to the parameters and to the order of trades
- Report what the data showed, with every assumption and limitation written down
- Automate a validated rule set in MQL5 for MetaTrader, or Python elsewhere
- Implement your risk controls, logging and safe-failure behaviour
- Fix, refactor and document code somebody else wrote
- Set up deployment and monitoring so it runs unattended
What we do not
- Advise on whether a strategy is any good, or suggest one
- Tell you whether something is worth trading
- Provide signals, tips or market opinions
- Make any claim about performance, profitability or returns
- Sell or licence a ready-made "winning" system
- Publish anyone's results, ours or a client's
- Manage accounts, handle client money, or trade on anyone's behalf
- Guarantee an outcome from any piece of software
How we test
Eight checks applied to every test. A result that has not been through them tells you very little.
02
Walk-forward analysis
The rules are fitted on one window, tested on the period after it, then rolled forward and repeated.
03
Realistic costs
Spread as it was at the time, your actual commission schedule, and a stated slippage assumption.
04
Parameter sensitivity
Results are checked across a range of settings. Rules that only work at one exact value have been fitted to the past.
05
Monte Carlo analysis
Trade order and price paths are resampled separately, to show how much of the result depended on luck.
06
Number of variants recorded
Test enough variations and one will look good by chance, so we report how many were tried.
07
Data integrity
No information that wasn't available at the time. Delisted instruments stay in. Gaps and bad ticks are handled explicitly.
08
Stated assumptions
The report lists what the test assumed, what it could not account for, and where the data was thin.
What you receive
Everything needed for you, or anyone you choose, to re-run the test and check it.
01
Written report
What the data showed and under which assumptions, with the limitations stated, including when the idea does not hold up.
02
Rule set
Your strategy written out precisely enough to code from, with the edge cases resolved.
03
Source code
Full, commented source, so the test can be re-run, audited or extended by another developer.
04
Test conditions
Data source and period, cost assumptions, parameter ranges and the number of variants tested.
Building and automation
If the rules hold up and you want them running automatically, that is a separate piece of work. The platform depends on what you trade.
MQL5
MetaTrader 4 and 5
For FX, CFDs, indices and anything your broker offers through a MetaTrader terminal.
- Expert Advisors coded to the written specification
- Custom indicators, panels and utility scripts
- Position sizing, exposure caps, spread and slippage filters
- MQL4 to MQL5 migration, quoted against the source
- VPS deployment, logging, alerting and monitoring
Python
Shares and other markets
For stocks, ETFs, futures, crypto and multi-asset work, where MetaTrader is not an option or not the right tool.
- Strategy research and testing across large instrument universes
- Screeners, scanners and scheduled research jobs
- Data pipelines, cleaning and point-in-time handling
- Execution through your broker's or data provider's API
- Deployment and monitoring, with defined behaviour when something fails
Safe failure
Reconnection handling, defined behaviour on bad or missing data, and a system that stops rather than carries on when something is wrong.
You get the source
The implementation of your rules is delivered as full commented source, with the specification it was written from. Generic components stay with us, and you hold a permanent right to run and modify them as part of your system.
Confidentiality
Written into our standard terms from the first conversation, with a separate non-disclosure agreement available before you send anything, at no cost.
Support afterwards
A walkthrough of the report and code on delivery, and support afterwards if your broker, market or requirements change.
Pricing
Fixed price against a written specification. The rules are pinned down first, and the quote follows.
Testing and validation
Testing your rules
Priced once the rules are unambiguous.
- Feasibility read
Your strategy read, the data checked, and a written answer on what can be tested and what it would take. Credited in full against the test if you proceed£195 - Single strategy test
One instrument, one timeframe. Coded, out-of-sample data, costs modelled, written reportFrom £450 - Full validation
Walk-forward analysis, parameter sensitivity, both Monte Carlo tests, variants recorded, and a full robustness reportFrom £1,150 - Multi-instrument or portfolio
The same rules across several instruments, or several rule sets tested togetherFrom £1,950 - Historical data
Where the job needs data we do not already hold. Passed through at cost and agreed before it is boughtAt cost
Building and automation
Putting it into software
MQL5 or Python. Quoted once the rules are written down, whether or not we did the testing.
- Indicator or script
Single-purpose and self-containedFrom £180 - Automating a rule set
Single strategy, single instrument, coded to your specificationFrom £550 - Advanced build
Multiple instruments, advanced risk handling, state that survives a restart£1,200 – £2,200 - Code diagnostic
Your existing code read and run, with a written report on what is wrong and what it would cost to fix. Credited in full against the work£95 - Review, debug and harden
Your existing system tested against its failure modes and documented. Quoted from the diagnosticStarting fee £250 - MT4 to MT5 migration
Ported, tested and documented. Quoted against the MQL4 sourceFrom £400 - Support
Changes and help after handoverOn request
The rules are pinned down before we quote. We agree in writing what counts as a signal and what happens on a gap, a requote, a partial fill or a disconnection. You then receive a fixed quote, with no obligation to accept it. We do not take work under £180, and we do not reproduce a system from a screenshot or a results page. These are founding rates, held below where they will settle, for a limited number of engagements. Prices shown are the total payable; no VAT is added.
Free tools
For traders and investors. They run entirely in your browser. Nothing you enter is stored or sent anywhere.
Position size calculator
How big to make a trade so a stop-out loses only what you set. Forex, shares, spread bets and CFDs.
OpenForex pip value calculator
What one pip is worth on your position, in your account currency. Any pair, gold, silver or a CFD.
Open
On the way: compound growth, profit and loss on a share trade, averaging down, and risk-reward with the break-even win rate. All free tools for traders & investors.
Questions
When can I buy the position sizing and risk workbook?
It is being made now. It goes on sale on its own page, with its price, when it is ready. Like our free calculators, it does the arithmetic on your own figures and gives no advice.
Will a strategy make money?
Nobody can tell you that in advance, and we don't try. Our work is to measure how your rules would have behaved on historical data and to build software that does exactly what the specification says. Whether the strategy is sound, and whether to trade it, is your decision and your risk.
What if the test shows my idea does not work?
Then the report says so, with the reason: which assumption it failed on, and whether it failed across the whole period or only in particular conditions. The price is the same whatever the result.
What data do you use?
It depends on the instrument and how precise the answer needs to be. FX work can often use freely available tick data. Equities usually need licensed data that includes delisted companies and prices as they were reported at the time. We will tell you what the job needs and what it costs. Data is passed through at cost and agreed before it is bought.
Can you test a strategy I did not write myself?
Yes, provided the rules can be written down, for example from a written method, a course, a set of conditions or existing code. We will not reconstruct a system from a screenshot of a results page or from marketing claims with no rules behind them.
Do I have to share my strategy with you?
Yes, in enough detail to code it. A mutual non-disclosure agreement is signed before you send anything. We do not trade client strategies, resell them or reuse them in anyone else's work.
Who owns the code?
The implementation of your rules is yours, in MQL5 or Python, delivered as full source with a perpetual licence to run and modify it. We keep ownership of generic components such as the risk module, logging and testing harness. Outright ownership of everything is available, priced separately and agreed before work starts.
Can you optimise my system for better backtest results?
We can set up optimisation properly and show you how to run it. A backtest tuned until it looks good is usually fitted to the past, so we would rather give you an honest test than a better-looking result.
Will you run it on a live account for me?
No. We set up the deployment and monitoring, but the account stays yours and under your control. We do not manage accounts, hold client funds or place trades on anyone's behalf.
What this is, and what it is not
We do not give financial advice, and we make no promises about returns.
A backtest tells you how a defined set of rules would have behaved on historical data, under stated costs and stated assumptions. It is not a prediction, it is not a recommendation, and it is not a guarantee of any result. We do not tell you whether a strategy is worth trading, we do not give strategy advice, and we do not sell signals or manage accounts. You bring the rules; we build them, test them honestly, and report what the data actually shows, including when it shows the idea does not work.
The same applies to our tools and calculators, free or paid. They carry out the calculations and instructions you give them. They do not choose trades, give signals or recommend any investment, and their output is not advice.
BrightDragon Systems Ltd provides software development and testing services only. Nothing on this page is investment advice, a recommendation, or an invitation to trade. We are not authorised or regulated by the Financial Conduct Authority and we do not provide regulated financial services.
Trading carries risk and you can lose money. Any decision to trade, and any strategy the software implements, is entirely yours. No software can guarantee an outcome, and past testing results are not a reliable indicator of future results.