In this area
Three entry points, one standard.
Every project starts from a concrete job — not from a platform. Prices are starting points for well-scoped work; the exact scope is fixed in the free first call.
| DAX | FAIL |
| FTSE | FAIL |
| Dow | FAIL |
| NQ | PASS |
Strategy validation
from$250
- ·Real M1 data, costs included
- ·Out-of-sample + walk-forward
- ·Report with an honest verdict — even "no edge"
The cheapest deliverable is a no before development.
Start a project →
Indicator development
from$150
- ·TradingView Pine v6 or MT5
- ·Alerts, multi-timeframe, no repainting
- ·Source code + guide
Ten public instruments show the standard.
Start a project →
Expert advisor / full system
from$500
- ·Entries, trailing, risk sizing, session logic
- ·Backtest report + documentation
- ·30 days of support
Larger systems quoted by scope.
Start a project →Smaller jobs (from $35) run through Fiverr — same code standard, smaller scope, with the platform's buyer protection. The prices above are direct projects: the full five-step process including written spec, validation gate and 30 days of support.
ProofTestimonial, numbers
Inspect the work before you pay.
Everything below is public — the same code standard and methodology every project gets.
MT5 systems at work
EAs and tools with risk sizing, session logic and clean logging. Here: the lot calculator from the free catalog, live on DAX and FTSE.
Free tools in the catalog →Risk is enforced, not hoped for
Our own drawdown guardian tracks daily and overall limits in real time and throttles position size before a limit breaks. Published on the MQL5 Market as PropGuardian; next to it PropMirror, the snapshot copier between your own accounts, which stops copying while the guardian holds a risk lock.
PropGuardian on the Market ↗PropMirror on the Market ↗
Custom Suites
Standalone applications from daily operations, as desktop program or web app: the analog-day analysis Studio, market context and alerts in MarketPulse. For you I build the same around your workflow: your data, your rules, one interface you use every day.
See Studio →See MarketPulse →"Really cleanly solved! Congratulations!"
PlatformsMT5 · TradingView · cTrader · NinjaTrader · Python
Five platforms, built natively.
No lowest common denominator — full use of each platform API. Always delivered with source code and documentation. NinjaTrader is in preparation — requests for it are already open.
MT5 Expert Advisors
.ex5 + source + docs
Pine Script Strategies
Pine code + guide
cTrader cBots
cBot + source code
NinjaTrader Add-ons
Add-on + source code
Coming soon
Python & Research
Scripts + documentation
Methodology6 principles
The approach — and why the result holds.
Most trading systems fail not on the idea but on weak validation. A backtest on coarse data with optimistic fills almost always produces a pretty curve — the very one that falls apart in a real account. The approach flips that: before a strategy counts as good, every effort is made to break it.
You can read exactly how that works: one edge from the first measurement to a running EA, documented in five phases — including the points where it nearly failed.
The series: from edge to EA · Full methodology
- 01Refute first, then trustEvery strategy is deliberately attacked — across data, parameters and market regimes. What survives may go live.
- 02Costs from the startSpread and slippage are in every metric. What only works gross does not work.
- 03Real dataM1 data as the base, quote origin checked. Where the order of events inside one minute decides the result, it is resolved at tick level: no fill that never existed.
- 04Out-of-sample & walk-forwardOptimized on one window, tested on the next unseen one — rolling over years.
- 05Monte Carlo & regimesTrade order, parameter sensitivity, crash, bear and chop — not just the one good year.
- 06Lean logic, clean code, hard riskClear rules over optimized indicator stacks; logging, configurable parameters, position sizing, stop & drawdown limits built in.
- 07Live-forward is the only truthEvery backtest comes from known data — and edges decay. This is stated openly instead of hidden.
Process5 steps
From request to delivery.
From first request to delivery — five clear steps with timelines. The methodology above lives in step 4.
- 01
Initial consultation & analysis
1–2 days
Clarify strategy, requirements and fit for automation — free and honest. If an idea will not hold, that is said up front.
- 02
Concept & specification
2–3 days
Precise entry/exit/risk rules in writing — approved before a single line of code.
- 03
Development
1–4 weeks
Clean, maintainable code with logging and configurable parameters, with regular progress updates.
- 04
Validation
3–5 days
The strategy runs the full validation protocol (see Methodology above) — what fails here does not go live.
- 05
Delivery & support
30 days support
Finished algorithm + source code, documentation, install help and 30 days of support.
The questions everyone asks.
Answered briefly and without sales tone. Question missing? The first call is free.
Can you make money with automated trading?
Yes — but not by buying a bot. An algorithm makes a strategy consistently executable and measurable; it does not create an edge that is not in the rule set. That is why every project here starts with a measurement, and a negative verdict is delivered as readily as a positive one.
What does it cost to have an Expert Advisor programmed?
Full systems from $500, indicators from $150, a strategy validation from $250. Smaller jobs from $35 run through Fiverr. The exact price depends on scope and is fixed in the free first call — before any money changes hands.
What happens if my idea shows no edge?
Then the report says so, with sample size, benchmark and cost assumption, and the project ends there instead of running into an implementation that would not change the finding. That is exactly why validation is the cheapest step: a no before development costs the validation price once, not the price of a full system. The measurement often also shows under which condition the rule holds and under which it does not; the no then turns into a narrower version of the idea. What such a no looks like is public in the research: the rejected hypotheses stand there with their numbers next to the confirmed ones.
I have a strategy but no code — is that enough?
Yes. Precise rules for entry, exit and risk are enough. Step 1 is always a written specification you approve before a single line of code is written. Even semi-discretionary rules can often be made measurable — we clarify that in the first call.
Which platforms are supported?
MetaTrader 5 (MQL5), TradingView (Pine Script v6), cTrader (C#) and Python — NinjaTrader is in preparation. Delivery always includes source code and documentation; nothing stays a black box.
How do I know the backtest is not embellished?
Real M1/tick data, spread and slippage included, out-of-sample and walk-forward testing — the full methodology is openly documented. The strongest evidence: the research section also lists the strategies that failed.
I built my strategy with ChatGPT or Claude. Can you check it?
Yes, and it has become the most common starting point. The code usually compiles; the mistakes are elsewhere: information not yet known at entry time, fills that do not exist for real, missing costs, session hours without daylight-saving logic, the point scale of the broker. Validation checks exactly these classes before money is attached. The full list with the numbers from our own tests is in the research section under "Building a trading strategy with AI". Anyone who keeps building on their own has the same checklist in hand.
Do you sell signals or ready-made profitable bots?
No. This site offers measurement tools, risk utilities and custom development for your own strategy. The systems I trade live myself are not for sale.
How long does a project take?
From first call to delivery typically two to six weeks depending on scope: specification in days, development one to four weeks, validation three to five days. Followed by 30 days of support.
First call free.
Analyze requirements, get an honest assessment — whether and how your strategy can be automated. If an idea visibly will not hold, we say so in the first call, before any money changes hands.
Get in touch