Analyze strategies.
Automate systems.
Trade with discipline.
For more than 10 years I have been building automated trading systems for MetaTrader 5, Interactive Brokers and ICM — together with a network of specialized partners when needed. From the first idea and the backtest setup all the way to live trading. Equities, futures, derivatives. Disciplined, documented, data-driven.
Stuttgart, Germany / Mallorca, Spain · remote · English and German
For companies: AI workshops & training — more than 50 held.
Four pillars. One goal.
Turning strategies from the financial markets into robust, automated systems — and taking companies to the next level with AI, in implementation as well as in training. Personally accountable, with a network of specialized partners when needed.
Trading Automation
From a discretionary approach to a rules-based machine: Expert Advisors for MT5, Python bots for Interactive Brokers, FIX connectivity via ICM.
- MQL5 Expert Advisors & indicators
- Backtesting & walk-forward analysis
- Risk management & position sizing
- Live deployment & monitoring
Systematic Investing
Software and rule sets for your own long-term investment decisions: equities, ETFs, futures, derivatives. Diversified. Disciplined. Documented.
- Rule set for allocation & rebalancing
- Trend & momentum models
- Rules for options strategies
- Documentation for your tax advisor
AI Consulting for SMEs
AI as a real lever in day-to-day business — from the AI audit and the proof of concept through to productive implementation. Proven in trading, applied to your company. Together with Crowd Solution GmbH.
- Automating processes: invoices, emails, reports
- Your own AI assistants & RAG systems (with a data protection plan)
- AI audit: where AI really pays off
- Strategy workshops & coaching for executives
AI Training & Workshops
Bringing AI to your team in a way everyone understands — from a compact impulse workshop and a multi-day onboarding for SMEs to 1:1 coaching for decision-makers. No jargon, built around your own workflows.
- Impulse workshops & hands-on training
- AI onboarding for entire teams & SMEs
- Prompt engineering in practice
- Executive coaching for leaders
Tell me about your project.
Whether it's automating a strategy, a reliable backtest or AI in trading or in your company — briefly describe what you need. The first call is free and without obligation.
Get in touch, no obligation →From an idea to a working system.
Every strategy starts with a hypothesis. I translate it into code, test it on historical data, harden it against overfitting — and move it cleanly into live trading. Projects run in English or German. Process, services & FAQ →
Platforms I work with
-
MetaTrader 5 · MQL5 Expert Advisors, custom indicators, Strategy Tester with multi-symbol & multi-timeframe testing.
-
Interactive Brokers · IBKR API Python bots via ib_async (the successor of ib_insync) and the TWS API. Stocks, options and futures worldwide.
-
ICM Capital · FIX connectivity Direct market access for institutional setups and multi-account connections.
-
Python ecosystem pandas, NumPy, vectorbt, backtrader, scikit-learn, PyTorch — the full quant stack.
// Trend breakout · adaptive ATR stops input int LookbackBars = 55; input double RiskPercent = 0.75; input double AtrMultiplier = 2.4; int atrHandle; int OnInit() { atrHandle = iATR(_Symbol, PERIOD_H1, 14); return INIT_SUCCEEDED; } void OnTick() { double hi = iHigh(_Symbol, PERIOD_H1, iHighest(_Symbol, PERIOD_H1, MODE_HIGH, LookbackBars, 1)); double atr[1]; CopyBuffer(atrHandle, 0, 0, 1, atr); double bid = SymbolInfoDouble(_Symbol, SYMBOL_BID); if (bid > hi && !PositionOpen()) { double lots = CalcLots(RiskPercent, atr[0]); OpenBuy(lots, bid - atr[0] * AtrMultiplier); } }
Four phases. Clearly defined.
Analyze the strategy
First, I pick your idea apart with you: edge, market conditions, statistical significance. What can be expressed as rules? Where do bias risks arise? Which instrument fits?
Set up the backtest
A clean backtest with real spreads, slippage and commissions — no wishful thinking. Walk-forward optimization instead of curve fitting. Robustness checks across symbols and market regimes.
Automate
Implementation as an Expert Advisor (MT5) or Python bot (IBKR/ICM). Risk management, logging, fail-safe logic and clean error handling are all included.
Run it live
Deployment on a VPS, monitoring, performance reports, continuous re-evaluation. If the market changes, I propose adjustments — they are only implemented with your approval.
Broad coverage, focused execution.
Stocks
Individual stocks and indices, worldwide. Trend and mean-reversion approaches, earnings plays, pair trading.
Futures
Index, interest rate, energy and agricultural futures. Spreads, trend following, seasonality.
Derivatives & Options
Structures for income, volatility and hedging. Greeks-aware execution.
FX & Crypto
Major and minor pairs on MT5. BTC/ETH for 24/7 strategies.
No black box. A system that belongs to you.
What you end up with is not a mysterious bot that only I understand — but a documented, transparent system in your hands.
A working system
A ready-to-run Expert Advisor (MT5) or Python bot (IBKR/ICM) — installed, tested and signed off on your demo account. You trigger the live start.
Commented source code
You get the code, not just a black box. Fully commented and documented — yours to use permanently for your own purposes.
An honest backtest report
With real spreads, slippage and commissions — including drawdown, Sharpe ratio and trade statistics. No embellished equity curves.
Risk & money management
Position sizes, stops and fail-safe logic — configured to your risk profile, not to grand promises of returns.
Documentation & onboarding
How the system works, when it steps in, what needs monitoring — plus a walk-through together until you feel confident.
Support after go-live
A monitoring setup and a point of contact who stays — even when the market turns and the system needs fine-tuning. I only implement changes after your approval.
What I don't do.
I develop software to your specification. I sell no winning strategies or signals, make no promises of returns, manage no client money and do not trade on your account. Trading decisions and the live start are your responsibility.
Risk warning: Trading securities, derivatives, futures, foreign exchange and cryptocurrencies involves substantial risk of loss, up to and including total loss. Backtests and past performance are not a reliable indicator of future results. Not investment advice.
Send a no-obligation project inquiry →Build wealth — instead of hoping.
Long-term investing doesn't take predictions. It takes rules you'll stick to even in bad years. I develop the tools for that: your rule set as a document and as software. You set the rules and make the investment decisions yourself.
Portfolio construction
Rules for your asset allocation that you set yourself and apply consistently — instead of gut feeling. Equities, bonds, commodities and alternatives in transparent weightings.
Rebalancing & rule set
Clear rules for when to buy, sell and reallocate — automated where it makes sense. Discipline beats the disposition effect.
Trend & momentum
Systematic trend following on indices and sectors. Risk-off rules designed to help limit losses in a crisis — without bailing out at every wobble.
Options for income
Cash-secured puts, covered calls, wheel strategies. Rules for premium strategies on existing holdings with a clear risk framework, cleanly documented for your tax advisor and custodian bank.
Note: I don't provide investment advice, portfolio management within the meaning of the German KWG/WpIG, or tax advice, and I don't make recommendations on individual financial instruments. I develop tools for your own investment decisions — you make the decisions yourself.
Three principles that hold it all together.
Rules over gut feeling.
Every decision follows a documented rule set. That makes performance traceable, drawdowns bearable and improvements measurable. Rules make decisions verifiable — even in bad years.
Diversification that works.
Real diversification doesn't come from holding many positions, but from uncorrelated ones. Equities + bonds + commodities + trend — four sources, four characteristics, one portfolio whose risks you aim to estimate and limit in advance.
Factor in costs and taxes.
Trading costs and frequently realized gains eat into returns. That's why the rule set accounts for gross and net from the start. Tax advice itself stays with your tax advisor.
Who benefits from a systematic approach.
The active investor who wants a system
You've been investing for years, you've tried quite a few things, and you've realized: without a system, you stay stuck in a loop of hope and regret. You want a clear rule to follow — even when the markets are moving against you.
The business owner with cash to invest
You want to invest your company's cash according to your own, traceable rules — instead of being offered the same funds over and over in the classic bank advisory setup. You want a transparent setup of your own that you decide on yourself.
The family with a multi-generation horizon
Preserving wealth across generations is a different task from short-term performance. I build rule sets and tools that don't depend on any single person and are documented so that they are still understandable twenty years from now.
The trader who wants to diversify
Active trading + long-term investing are two different games. If you do both, you need a clear separation of accounts, strategies and mental modes. I help draw that line cleanly on the technical side.
In the end, you hold a rule set in your hands.
No power of attorney, no access to your account — just a system you can follow yourself. Here's what you actually get:
A written rule set
Your strategy as a clear, traceable document — when to buy, sell and reallocate. No more gut decisions.
Asset allocation
An allocation rule across equities, bonds, commodities and alternatives that you set based on your own goals — transparently documented.
A rebalancing plan
Fixed rules for when the portfolio is brought back into balance — automated where it makes sense. Discipline instead of the disposition effect.
Tracking & reporting
A simple way to follow the system and track its progress — without having to stare at your account every day.
Clean documentation for tax purposes
Documentation your tax advisor understands. Gross and net considered from the start — tax advice itself stays with your tax advisor.
You stay in control
I build the system — you make the decisions. No dependency, full transparency, adjustable at any time.
Risk warning: Investing in securities, bonds, commodities and other financial instruments involves risk of loss, up to and including total loss. Backtests and past performance are not a reliable indicator of future results. Not investment advice.
AI as a tool. Not a buzzword.
I help small and midsize companies use AI where it measurably saves time and money — not as a prestige project, but as a concrete tool in day-to-day business. From the AI audit to the proof of concept to production rollout. Personally accountable, with Crowd Solution GmbH as a partner when a project needs more capacity.
My edge comes from first-hand experience: for years I have been building AI systems that run in production in trading — where models have to work not just in a pilot, but under real pressure and with real money. That hard-won practical experience goes into every company engagement.
AI in Trading
- ML-based signal models
- News & sentiment analysis with LLMs
- Regime detection
- Feature engineering from market data
AI in Business
- AI audit: where does AI really pay off?
- Process automation
- Custom assistants & RAG systems
- Implementation planned with data protection in mind
Coaching & Workshops
- AI fundamentals for executives
- Prompt engineering in practice
- Choosing tools: ChatGPT, Claude, Gemini
- A sparring partner for in-house builds
Where AI has an immediate impact in SMEs
AI doesn't have to be a big project. The biggest levers are often in everyday, recurring processes. A selection of typical use cases — by business area. All details on AI consulting (German) →
Administration & Accounting
- Capture incoming invoices automatically & pre-code them for bookkeeping
- Triage inboxes & draft replies
- Reports & analyses at the push of a button
Sales & Marketing
- Qualify & prioritize leads automatically
- Quotes & sales brochures in minutes instead of hours
- Generate content for web, social & newsletters
Customer Service
- 24/7 assistant for recurring inquiries
- Classify & route tickets automatically
- A knowledge base that answers questions instead of making you search
Production & Engineering
- Predictive maintenance: anticipate failures
- Image-based quality control
- Data-driven process & energy optimization
Knowledge & Documents
- RAG system over your company's own documents
- Internal chatbot instead of endless file searches
- With a data protection plan, on your own infrastructure if you prefer
Management
- Data analysis & decision dashboards
- AI strategy & roadmap with a clear ROI focus
- Sparring on equal terms — in confidence
For in-depth AI consulting for SMEs: Crowd Solution.
For company projects involving financing, strategy and AI implementation, I work closely with Crowd Solution GmbH, a consulting firm focused on small and midsize businesses.
How AI projects really work.
Most AI initiatives don't fail because of the technology, but because of the setup. This is how I approach them — whether in a trading context or at a midsize company.
Use case audit
I work with you to identify the three areas in your company (or your trading setup) where AI delivers measurable value — and rule out those where it would only mean effort without return. That often saves months of expensive trial and error.
Proof of concept
One concrete use case gets implemented — lean, in two to four weeks. The goal isn't perfection but insight: does it work? Where are the data gaps? Who on the team needs to be involved? How do users react?
Scaling
If the PoC works, the production rollout follows. With clear processes, monitoring, a data protection plan and a roadmap for the next twelve months. If it doesn't work, you know early — and save yourself an expensive large-scale project.
Enablement
An AI system that only one person in the company understands is a single point of failure. I build up knowledge in your team — workshops, documentation, sparring. So that 12 months from now, you depend on me no more than you have to.
The tools I use myself.
Language models
Claude (Anthropic) for deeper analytical work, ChatGPT (OpenAI) for day-to-day tasks, Gemini for long-context tasks. In trading: the Claude API for parsing earnings calls and news sentiment.
ML stack
Python with XGBoost and LightGBM for tabular data, scikit-learn for classic models, PyTorch for deep learning, SHAP for explainability. Standard in professional quant work.
Automation
n8n for AI workflows in companies (self-hosted if you prefer), Make for quick SaaS integrations, custom Python pipelines for trading-specific data flows.
Data
Norgate for clean stock histories (including survivorship-bias-free data), LSEG (formerly Refinitiv) for institutional setups, Massive (formerly Polygon.io) and Alpha Vantage for API pipelines, IBKR for live data.
Development
Cursor as an AI IDE, GitHub for version control, Jupyter for research, VPS setups for live strategies. Documentation in Notion or right in the code.
RAG & knowledge systems
For company engagements: custom RAG systems over your company's own documents — with a data protection plan and, if you prefer, on your own infrastructure, without sharing data with external SaaS providers.
What companies ask first.
Is AI even worth it for my company?
That is what the AI audit clarifies, before any money is spent. The biggest levers are often in everyday, recurring processes — not in expensive large-scale projects.
Will my data stay secure?
I build data protection into the plan from the start. If you prefer, the solution runs on your own infrastructure, so no data flows to external SaaS providers. The data protection assessment in each individual case remains with your data protection officer.
How quickly will I see results?
A first proof of concept is usually ready in two to four weeks. The goal is early clarity on whether the use case pays off — not an expensive year-long project.
Do I need a technical background?
No. I explain everything without jargon and build up knowledge in your team — so you don't stay dependent on me or on any one tool.
Want more detail? Read the detailed AI consulting page (German) →
AI your team will actually use.
Most AI training gets stuck in theory. Mine doesn't. I show your team, using its own processes, where AI already saves time today — clearly, without jargon, with tools that are ready to use the next business day. From skilled-trades businesses to midsize companies to executive leadership, on-site or remote, in English or German. I have run more than 50 workshops this way. All details on the AI workshops →
AI Impulse Workshop
The compact introduction for teams that want to finally make AI tangible.
- The basics without jargon — what AI can do today (and what it can't)
- Live demos of the key tools using real examples
- Quick wins you can use right away in your daily work
- First use cases straight from your company
Outcome: Your team understands which tools are worth using for which task — and gets started right away.
Request a workshop →AI Onboarding for SMEs
The complete program when AI is meant to reach your whole company in a structured way.
- A tailored AI roadmap for your company
- Process analysis: where does AI pay off, and where doesn't it?
- Hands-on training & change management within the team
- Data protection, EU AI Act & clear ground rules
Outcome: A clear roadmap with priorities — and a team that can put it into practice.
Request onboarding →Executive Coaching
Confidential sparring for decision-makers who want to use AI strategically.
- Clarity instead of overload: what AI means for you
- A concrete strategy with clear priorities & an eye on ROI
- Sparring on equal terms — honest feedback
- Competitive advantage through measurable automation
Outcome: You decide on AI spending with the full picture, not on gut feeling.
Request coaching →Don't see the right format? Every training is tailored to your industry, prior knowledge and goals — from a 2-hour impulse session at a team event to a multi-part training series. Tell me what you need.
What you will actually learn.
The topics are building blocks — for your training, I put together exactly the ones that fit your daily work.
AI basics, made clear
How AI & language models work, what they can do today — and where their limits and risks lie. No technical background required.
The right tools
ChatGPT, Claude, Gemini, Copilot & co. — which tool is right for which task and how they actually differ.
Prompt engineering
How to get reliable, usable results from AI instead of hit-or-miss output. The technique behind it — practiced hands-on.
AI in everyday work
Emails, documents, research, meeting minutes, analyses, quotes — practiced on real tasks from your company.
Data protection & the law
GDPR, EU AI Act, works council agreements: what is permitted in your company and how to use AI with clear rules.
Your own use cases
Together with you, I identify the processes in your business where AI saves measurable time right away — and prioritize them.
Practice, not slide-deck theory.
I'm Marcel Gautsche, a quant developer from Stuttgart. For more than ten years I have been building automation systems that have to work under real pressure, since 2022/23 with AI as well — in algorithmic trading, where a model has to hold up with real money, not just in a pilot project.
I bring this practical experience to every training: not abstract theory, but tools your team will use the very next day. I explain AI without jargon and always with your own processes as examples — from skilled-trades businesses to midsize companies to executive leadership. Where companies need to go deeper, I work with Crowd Solution GmbH and other specialized partners.
My aim: you and your team leave the training with routines that stick, not just notes.
-
10+
Years of systematic work, with AI since 2022/23 AI systems in production in algorithmic trading — under real pressure, with real money.
-
Quant developer, not a theorist Python, ML stack, LLMs — built and run in production myself, not just explained.
-
Crowd Solution partner Partner network for larger SME projects.
-
Stuttgart & across Germany On-site training anywhere in Germany, or conveniently remote.
From inquiry to routine.
Free preliminary call
In a 20–30 minute call, I clarify goals, prior knowledge, industry and scope with you. No obligation — so I know what your team really needs.
Tailored concept
You get an agenda tailored to your daily work — not off the shelf. Modules, examples and pace fit your team, not the other way around.
The training
Interactive, with live demos and exercises on real examples from your company. No lecturing — your team works hands-on from the very first hour.
Follow-up
A cheat sheet, a tool list and the most important prompts to take with you. If you like, a follow-up session after a few weeks to answer questions that come up in practice.
Time to bring AI to your team.
Tell me briefly about your company and what you want to achieve — I'll get back to you with a suitable training proposal. The preliminary call is free and without obligation.
Request training →Partner area.
For clients and partners: live performance from real trading — complete, with good and bad months, imported automatically from Interactive Brokers every day.
Performance in numbers
Preview: the figures are only visible after signing in.
Inside the protected area
Real numbers. Straight from the broker.
No screenshots, no hand-picked favorite trades: the data comes in automatically through the Interactive Brokers Flex Report and is updated every day.
Performance as an index
Time-weighted return (TWR): deposits and withdrawals are factored out. The curve starts at 100 — without absolute amounts.
Drawdowns
Every decline from the last peak is visible — because how a system gets through rough patches says more than any return.
Monthly returns
Every month since inception in one table — the bad ones just as much as the good ones.
Key metrics
Volatility, Sharpe ratio, maximum drawdown, best and worst month, share of positive months.
Automatically up to date
Pulled daily straight from the broker account, read-only. No manual upkeep, no selection after the fact.
Personal access
Every partner gets their own username and their own password. I issue the logins personally.
Interested in the partner area?
Clients and partners receive personal login credentials. A short message is all it takes.
Request access →The figures shown reflect trading with my own capital only. No offer, no investment advice, no asset management, no acceptance of third-party capital. Past performance is not a reliable indicator of future results.
Thoughts on trading bots, systems & AI.
Notes from practice — what works, what doesn't, and why. No trading tips, just the craft: broker APIs for Interactive Brokers, MetaTrader 5 Expert Advisors, TradingView webhooks, platform comparisons and the honest state of AI in trading and in companies.
TradingView to IBKR: Webhook Bridges vs Your Own Receiver
TradingView cannot send orders to Interactive Brokers on its own. How TradersPost, SignalStack and PickMyTrade connect, and a tested Python webhook receiver that queues alerts without placing orders.
Read articleProp Firm EA Rules in Code: Daily Loss, Max Drawdown and News Windows
How to enforce daily loss, static or trailing drawdown, trading days and news windows in an MT5 Expert Advisor, with FTMO, FundedNext and The5ers rules from their own pages.
Read articleExpert Advisor Not Placing Trades in MT5: A Diagnostic Checklist
Expert advisor not placing trades in MT5? An eight-step diagnostic order, from logs and permissions to return codes, filling mode, stop levels and the EA's own entry logic.
Read articleIBKR Web API vs TWS API: Which One Your Trading Bot Needs
TWS API or Web API for an Interactive Brokers bot? Login, IB Gateway, rate limits and sessions compared, with a decision table and a tested Python pacing check.
Read articleIBKR MCP With Claude or ChatGPT: What Works and What Does Not
Claude, ChatGPT or Claude Code with Interactive Brokers via MCP: what the official server can do, the Review & Submit step, read-only setup, GitHub servers and security risks.
Read articleIB Gateway Docker Setup After IBC: Headless, 2FA and Restarts
IBC is archived. How to run IB Gateway in Docker as of November 2026: what IBKR supports, which 2FA methods work, a Compose example with secrets and a tested watchdog for your bot.
Read articleib_async vs ib_insync: Migrating an IBKR Python Bot in 2026
ib_async vs ib_insync in 2026: what changed in 2.x, stub-tested migration code for qualifyContracts and reconnects, error 200, and when the official TWS API fits better.
Read articleHow Much Does a Trading Bot Cost? Hire a Developer or Build It With AI
How much a trading bot costs depends on cost drivers, not a price list: what drives the build effort, which running costs stay either way and honest criteria for hiring or building it with AI.
Read articleInteractive Brokers API Developer: Hiring Guide and Checklist
How to hire an Interactive Brokers API developer: the specification, the IBKR steps you complete yourself, an acceptance checklist for the paper account and who owns the code.
Read articleAI Literacy Under EU AI Act Article 4: Rules After the Omnibus
What Article 4 of the EU AI Act requires since the Digital Omnibus of July 2026, who is covered, how Germany enforces it, a checklist and an example AI literacy record.
Read articleBuilt a Trading Bot With Claude Code or ChatGPT? A Pre-Live Checklist
Claude Code and ChatGPT can write a trading bot. Before it trades live: 24 verifiable checks, from look-ahead bias and reconciliation to kill switch, paper phase and secrets.
Read articleThe full library of 986 in-depth articles is in German, sorted into 18 topic hubs:
No articles found for .
Is a topic relevant to your project?
Whether it's a trading system or AI in your company: briefly describe your plans — you'll get an honest first assessment within 48 hours on business days.
Get in touch, no obligation →Calculators & Simulators
25 tools and games from the everyday work of systematic traders and investors — calculators, simulators and interactive mini-games, right in your browser, no sign-up. Change the parameters and instantly see what happens.
All tools are for illustration only and do not constitute investment advice. Calculations are simplified.
Position size calculator
How many units can you trade without risking more than your set limit?
Kelly criterion calculator
The bet size that maximizes long-run growth in theory, with a growth simulation that shows why professionals risk only a fraction of Kelly.
Account growth simulator
Compound returns plus regular contributions (at the end of each month) over time, made visible.
Monte Carlo drawdown simulator
Same strategy, hundreds of possible futures. Shows how widely results spread from chance alone.
*Share of runs that at some point lost more than 50% of starting capital. Results in % of starting capital (100% = break-even). Histogram: distribution of all ending values.
Equity curve backtest playground
Moving average crossover on synthetic sample data (long only, no costs). Move the sliders and the strategy recalculates instantly.
Black-Scholes option calculator
Theoretical option price and Greeks for European options without dividends, with an interactive profit and loss chart for a long option at expiration.
Sharpe lab
Return alone tells you little. Only in relation to risk can you see whether a strategy actually holds up.
Correlation lab
Why diversification is called the only “free lunch” in markets: change the correlation and watch the portfolio risk.
Recovery calculator
The harsh math of drawdowns: losses and the gain needed to recover are not symmetric, and the gap widens fast.
Dividend calculator
The dividend snowball: how reinvested payouts can turn into dividend income over the years.
Simplified model in annual steps: dividends and contributions are reinvested at year-end, all rates stay constant, before taxes and costs.
ETF Savings Plan Calculator
What is left at the end, after fund costs (TER) and inflation, before taxes and assuming a constant return.
CAGR / Return Calculator
The geometric, annualized growth rate between two values.
Risk-Reward Calculator
Does the trade add up? Risk-reward ratio, break-even win rate and expected value at a glance.
Leverage & Liquidation Calculator
How close is liquidation? With leveraged CFD or crypto positions, that is the key question.
🎮 Spike Trader
The price moves live. Go long or short and close at the right moment. 45 seconds: how much can you make?
Tip: discipline beats overtrading. You don't have to trade every tick.
🔮 Bull vs. Bear
You see a chart. Will the next candle go up or down? Build your streak and notice how hard “reading” a chart really is.
Over many rounds almost everyone ends up near 50%. The prices here are pure chance: whatever you think you see in a chart has to be tested statistically first.
💀 Drawdown Survival
Choose your risk per trade and get through 40 trades with a slightly positive expected value without blowing up your account. Greed gets punished here.
The expected value is slightly positive (52% win rate, 1:1). Even so, oversized bets ruin many runs (at 25% risk, nearly half of them).
🧠 Pattern Quiz
Which classic chart pattern do you see? Train your eye for head and shoulders, double top, flag, triangle and more.
Tip: patterns are probabilities, not guarantees. A backtest shows what actually holds up.
🎯 Random or Trend?
Each chart is either pure chance (a random walk) or has real structure (trend/mean reversion), with equal odds every round. Can you tell the difference? It is much harder than it looks, and that is the lesson.
Lesson: the human eye sees patterns even where there is only noise. That is why you need statistical validation instead of gut feeling.
🔥 Tilt Test
You are in a losing streak. Every decision counts: follow the plan, take a break, or go for revenge? 8 rounds, one discipline score.
—
Revenge trading after losses wrecks many accounts. This game shows why “doing nothing” is often the best decision.
🫧 Bubble Timing
The price climbs parabolically: pure euphoria. Sell before the bubble bursts. The later you sell, the bigger the gain … until there is nothing left.
Lesson: “this time is different” rarely is. If you try to catch the exact top, you usually hold too long.
🧩 Portfolio Challenge
Split 100% across four asset classes. Then a random market scenario hits. How robust is your mix? The scenario returns are simplified game assumptions, not historical data and not a forecast.
Tip: a robust mix is not the one with the highest return in a single scenario, but one that holds up reasonably well across many.
📰 News Trader
A headline appears. Long or short? Decide before the market reacts. 10 rounds.
Ready?
News trading looks easy, but it rarely is. Often the reaction is already priced in.
🕯️ Candlestick Quiz
Which classic candlestick pattern do you see? Hammer, doji, engulfing and more: the basics of technical analysis. The candles appear without trend context; in practice, the same shape goes by a different name depending on the prior trend (hammer or hanging man).
Single candles are hints, not signals. Only context and testing turn them into a strategy.
⚠️ Margin Call
Pick your leverage and survive 30 volatile ticks without getting liquidated. More leverage means bigger potential gains, but also a much faster route to a total loss.
Leverage does not improve your odds; it shrinks the distance to liquidation. That is where many traders fail.
Want models like this for real — on your own data, with real backtests and a live connection?
Get in touch, no obligationMarcel Gautsche.
I'm Marcel Gautsche, a quant developer from Stuttgart. For more than ten years I have been focused on financial markets, strategy automation and — increasingly in recent years — the question of how AI becomes a real lever rather than just a marketing term.
I do the programming and testing myself — strategies, code reviews and backtests are done by me, not by third parties. The Insights articles on this site are created with AI assistance — exactly the way I teach companies to work with AI: AI handles research and the first draft; I set the topics, direction and practical angle and take responsibility for the content. If you spot a mistake, I will correct it. For scaling and specialist topics (AI infrastructure, regulatory questions, consulting for SMEs) I work with a network of experienced partners, including Crowd Solution GmbH. Personal responsibility plus a strong network — that keeps projects robust.
In recent years I have developed trading systems for MetaTrader 5, Interactive Brokers and ICM — from discretionary trader setups to fully automated multi-strategy portfolios. In parallel, I advise companies on implementing AI in their operational processes, together with Crowd Solution GmbH and its partner network. I have now run more than 50 AI workshops for teams and executive leadership.
What drives me: combining hard statistics with honest practice. Markets don't reward opinions — they reward discipline, testing and a system that holds up under stress.
What shaped me.
From trader to quant developer
I started out in the markets as a discretionary trader — reading charts, interpreting news, taking trades based on setups. What I learned: on good days I was a good trader, and on bad days a bad one. The only solution was to move the system out of my head and into code. Since then, the rule set decides, not how I happen to feel that day.
From strategy to platform
First Expert Advisors for MT4, then MT5. In parallel, I built a Python stack for equity and options strategies via Interactive Brokers. Then bridges between the two — Python computes, MT5 executes: the best of both worlds.
AI as the next stage
With ChatGPT's leap forward in 2022/23, it became clear: this changes not only office work, but also how market data is processed. Since then: ML models for signal generation, LLMs for sentiment analysis, and in parallel AI consulting for companies — together with Crowd Solution GmbH.
What remains
Markets don't interest me because they are spectacular, but because they are honest. A strategy works or it doesn't — no “actually”, no “in the long run, yes”. I look for the same clarity when working with clients: what works, what doesn't, and why.
How I work.
Honesty before business
If your idea doesn't work, I'll tell you. If a tool isn't the right fit, I'll turn the project down. Nothing is more expensive than a wrong recommendation that was only made because I wanted the job.
Code you keep
You receive everything I build for you — source code, documentation, models — yours to use permanently for your own purposes. No black boxes, no vendor lock-in. So you stay self-sufficient after the project ends.
Data protection is not an add-on
GDPR compliance is a must, not a nice-to-have. Self-hosted setups for company projects, local models for sensitive data, clear records of processing activities. Compliance gets built in from the start, not bolted on afterwards.
Start small, learn fast
Better a working PoC in four weeks than a requirements specification in six months. Markets and technologies change too fast for classic large-project methods. Iteration beats planning.
You remain the decision-maker
I build systems, you make the decisions — on trades, on investments, on business strategy. I have no portfolio management license, nor do I want one. You keep control and responsibility.
Long-term relationships
Markets change, models age, strategies need readjusting. The best setups come from years of working together — not from one-off consulting days. I prefer to work long term — but clearly defined individual projects are just as welcome.
What people often ask.
Are you an active trader yourself?
Yes. I trade my own strategies live — otherwise I wouldn't have the reality check I need for client projects. I show the live performance to clients and partners in the protected partner area.
Will I get “the” winning strategy?
No. I don't sell strategies — I implement yours: according to your rules, your time budget and your risk limits. A universal winning strategy would be worthless the moment it spread.
How quickly can I go live?
A simple MT5 strategy including backtest: 2–4 weeks. More complex Python setups with ML components: 6–12 weeks. Larger multi-strategy portfolios: 3–6 months. I'll go through the plan with you in the intro call.
What does it cost to develop a trading system?
That depends on the scope and complexity of the strategy, which is why I don't quote fixed prices: after a free, no-obligation intro call you receive an individual quote, as a fixed price, on a time-and-materials basis or as ongoing support.
Do you take responsibility for performance?
No. I take responsibility for a cleanly built, tested system — not for market movements. There are no performance guarantees in serious trading, and anyone who promises them is lying.
Do you also run AI workshops for companies?
Yes. I have already run more than 50 AI workshops — for teams, SMEs and executive leadership. Formats range from a half- or full-day AI impulse workshop (up to 12 people) and a two- to three-day AI onboarding program for SMEs to 1:1 coaching for executives. On site in Stuttgart and the surrounding region, on site anywhere in Germany, or remote.
How much does an AI workshop cost?
That depends on group size, duration and scope. After a free, no-obligation preliminary call you receive an individual quote.
How can I reach you?
By email at info@marcelgautsche.de or via the contact form. In-person meetings in Stuttgart and the surrounding region are possible; otherwise remote via video.
Tell me about your project.
Describe what you need in a few sentences — your goal, your starting point, your preferred timeframe. You'll receive a well-founded first assessment within 48 hours on business days. Free of charge and without obligation.
- info@marcelgautsche.de
- Location
- Stuttgart / Mallorca
- Response time
- Within 48 hours on business days
What I can help with
- Automating a strategy
- Analyzing & improving existing EAs
- Backtest or walk-forward setup
- Integrating AI into trading or your business
- AI training or a workshop for your team
- Coaching & sparring
Thank you for your inquiry!
Your message has arrived. I usually reply personally within 48 hours, often sooner. For anything urgent, you can reach me directly at info@marcelgautsche.de.