Building a Trading System with 13 Capital Gates
A trading system that scales from $200 to $10M through 13 progressive gates. Each gate unlocks more capital only after the previous one proves it won't lose money.
I built a trading system that scales from $200 to $10M through 13 progressive gates. Each gate unlocks more capital only after the previous one proves it won’t lose money.
Most trading systems are designed around a single question: what should I buy? Mine is designed around a different question: how much should I risk?
The answer changes depending on how much evidence you’ve accumulated that your strategy actually works. Not backtested evidence. Not simulated evidence. Real money, real markets, real drawdowns survived.
The Problem with Most Trading Systems
Every retail trader I’ve watched blow up follows the same pattern. They find a strategy that backtests well, allocate real capital on day one, hit a drawdown they didn’t expect, panic sell, and conclude the strategy was broken.
The strategy wasn’t broken. Their position sizing was wrong for their experience level with that strategy.
Backtesting tells you what a strategy did in historical conditions. It tells you nothing about whether you can actually execute it when your account is down 8% and every financial news outlet is screaming about a crash. That gap between knowing and doing is where accounts go to die.
The fix isn’t better backtesting. It’s progressive exposure — proving at small scale before you earn the right to operate at large scale.
Two Intellectual Debts
This system stands on two bodies of work that most people treat as incompatible.
Gary Antonacci’s Dual Momentum says you should hold assets showing both absolute momentum (positive returns over a lookback period) and relative momentum (outperforming alternatives). When neither condition holds, you go to cash. The strategy is boring. It trades infrequently. It underperforms in raging bull markets. And it has crushed buy-and-hold on a risk-adjusted basis for decades.
Nassim Taleb’s Antifragility says you should build systems that gain from disorder. Not systems that resist shocks — systems that get stronger when shocks happen. In trading, this means position sizing that profits from volatility rather than being destroyed by it.
Most people pick one camp. Momentum traders optimize for trend-following. Antifragility practitioners optimize for tail events. My system uses momentum for asset selection and antifragility for capital allocation. The momentum engine picks what to hold. The gate system decides how much.
The 13 Gates
Each gate is a checkpoint. You start at G0 with $200 and paper trading. You graduate to the next gate only when specific conditions are met. No exceptions. No “I feel confident” overrides. Numbers or nothing.
G0 — Paper Trading ($200 notional) You prove you can follow the rules for 90 days without deviation. No real money at risk. The only thing being tested is discipline. Most people fail here because they can’t resist overriding the system when they “know” the market is about to move.
G1 — Micro Live ($200 real) First real money. The amount is intentionally trivial. You’re not trying to make returns — you’re testing execution mechanics. Does your broker fill orders the way your backtest assumed? What’s the actual slippage? How do you handle a real loss versus a simulated one?
G2 — Small Live ($500) First meaningful position sizing. You start tracking your actual Sharpe ratio against the backtest Sharpe ratio. If the gap is more than 0.3, you stay at G2 until you close it.
G3-G5 — Scaling ($1K, $2.5K, $5K) Each gate doubles or triples exposure. Graduation requires 90 days of positive expectancy at the current gate. The system tracks whether your behavior changes as dollar amounts increase. A lot of people are disciplined with $500 and erratic with $5,000.
G6-G8 — Intermediate ($10K, $25K, $50K) Tax implications become real. You need to start tracking wash sales, short-term versus long-term gains, and quarterly estimated payments. The system flags when a trade would create a wash sale violation.
G9-G10 — Advanced ($100K, $250K) Portfolio-level risk management kicks in. Sector concentration limits. Correlation monitoring across holdings. The system refuses to enter a position if it would push any single sector above 30% of portfolio value.
G11-G12 — Scale ($500K, $1M+) Institutional-grade concerns. Liquidity analysis before entry — can you exit this position in a day without moving the market? Counterparty risk monitoring. Multiple broker relationships for execution redundancy.
The numbers at each gate aren’t arbitrary. They’re calibrated so that a maximum drawdown at any gate doesn’t cascade into the next gate’s capital. If you lose 20% at G5 ($5K), you’ve lost $1,000. Painful but survivable. If you lose 20% at G12 ($1M), you’ve lost $200,000. That’s only acceptable if you’ve proven through G0-G11 that your maximum drawdown is actually closer to 8%.
Six Circuit Breakers
The gate system handles strategic risk — how much capital you deploy over time. Circuit breakers handle tactical risk — what happens when things go wrong right now.
Daily Loss Limit: -2% If the portfolio drops 2% in a single day, all new orders are cancelled and no new positions can be opened until the next trading day. This isn’t about the market being wrong. It’s about preventing revenge trading — the instinct to “make it back” that turns a 2% loss into a 10% loss.
Drawdown Kill Switch: -10% If the portfolio drops 10% from its high-water mark, everything liquidates to cash. Full stop. The system enters a 30-day cooling period before any new positions can be opened. This has triggered exactly once in testing. It saved the portfolio from what would have been a 23% drawdown.
Volatility Spike Breaker When the VIX exceeds 2 standard deviations above its 90-day mean, position sizes are cut in half automatically. High volatility means the normal distribution assumptions baked into position sizing are wrong. Cutting size is the honest response.
Correlation Breaker When cross-asset correlations spike above 0.8 (everything moving together), the system treats the portfolio as a single concentrated bet and sizes accordingly. Diversification is a myth when correlations converge.
Liquidity Breaker If average daily volume for any holding drops below 10x the position size, the system flags it for exit at the next rebalancing window. Illiquid positions are positions you can’t leave.
Execution Deviation Breaker If actual fill prices deviate from expected prices by more than 0.5% on three consecutive trades, the system pauses and alerts for manual review. Something is wrong with execution — maybe the broker, maybe the market microstructure, maybe a fat finger.
Key Rotation and Security
Every API key, broker credential, and encryption key rotates on a 90-day cycle. The system generates new keys, tests them against the broker’s sandbox environment, swaps them into production, and revokes the old ones. No manual steps.
This matters because trading systems are high-value targets. A compromised API key with withdrawal permissions can empty an account in seconds. The 90-day rotation limits the window of exposure if a key leaks.
Keys are stored in an encrypted vault, never in environment variables or config files. The vault itself requires two-factor authentication to unlock, and it only unlocks for the duration of a trading session.
Why Gates Beat Backtesting
Backtesting answers the question “would this have worked?” Gates answer the question “can I actually execute this?”
The difference matters because execution is where most strategies die. A strategy with a 1.5 Sharpe ratio in backtesting and a 0.3 Sharpe ratio in live trading isn’t a good strategy with bad execution. It’s a strategy whose backtest was lying about the difficulty of execution.
Gates expose this gap early, when the capital at risk is small. By the time you reach G8, you have 2+ years of live trading data at progressively larger sizes. Your live Sharpe ratio is a real number, not a fantasy extrapolated from historical data.
The Antifragility Layer
Here’s where Taleb’s thinking changes the architecture. Most trading systems try to minimize volatility. Mine tries to profit from it.
At each gate level, 5% of the portfolio is allocated to asymmetric bets — positions where the maximum loss is the 5% allocation but the potential gain is 10-50x. These are typically deep out-of-the-money options on assets the momentum engine has identified as having extreme relative strength or weakness.
In calm markets, this 5% allocation bleeds slowly. It’s a cost. In violent markets — the ones that destroy traditional portfolios — these positions explode in value. The portfolio gets stronger when things get weird.
This is antifragility in practice: a system that has more upside than downside from disorder. The gate system ensures you never bet more than you can afford to lose at your current level. The antifragility layer ensures that the worst market conditions are your best profit opportunities.
What I Learned Building This
Building a trading system taught me things about system design that apply far beyond finance.
Progressive trust is better than binary trust. Don’t give a system full permissions on day one. Let it earn permissions by proving competence at each level. This applies to AI agents, employee access controls, and deployment pipelines.
Circuit breakers are cheaper than recovery. Every circuit breaker in this system was inspired by a real failure mode I either experienced or observed. The 30 minutes it took to implement each breaker has saved hundreds of hours of recovery work.
The system should make the right thing easy and the wrong thing hard. When the daily loss limit triggers, I can’t override it without physically changing code, recompiling, and redeploying. That’s intentional. Future-me under stress should not have the power to undo a decision past-me made while calm.
Measure behavior, not just outcomes. A profitable month where you broke your rules three times is worse than an unprofitable month where you followed them perfectly. The gate system tracks rule compliance separately from returns. You can graduate with mediocre returns and perfect compliance. You cannot graduate with great returns and poor compliance.
Current Status
The system is at 45% completion. Paper trading is live. Gates G0-G3 are implemented and tested. The circuit breaker framework is complete. Remaining work is the upper gates (G4-G12), the antifragility allocation engine, and the tax tracking module.
I’m not rushing it. A trading system with bugs loses money. A trading system deployed before it’s ready loses money faster.
If you’re building systems that handle risk — whether financial, operational, or reputational — the gate architecture might be useful to you. The principle is universal: don’t grant capability until competence is proven, and always have a way to step back down.
Building systems that handle real consequences? I work with teams designing risk-aware architectures for trading, AI governance, and operational infrastructure. Book a call to discuss your architecture.