Risk fundamentals.

Sizing matters more than signal quality

The cruelest math in trading: a 70%-win-rate strategy with bad sizing loses money. A 40%-win-rate strategy with good sizing makes money. The signal is the easy part; the sizing is what separates traders who compound from traders who don't.

BABA gives you the signal. You decide what size to put behind it. This is the lever the engine cannot pull for you, because we don't know your account size, your risk tolerance, your tax situation, or how you'd react to a 20% drawdown. The default $3–$5 size is conservative on purpose — it's the right starting point until you've built a personal track record.

The 1% rule

The strongest single piece of risk discipline in trading: never risk more than 1% of your total account on a single position. "Risk" means the distance from entry to your stop-loss, not the notional size.

Example: your total trading capital across all venues is $1,000. The 1% rule allows you to risk $10 per trade. A signal arrives with entry $1.789 and SL $1.852 — that's a 3.5% stop. To risk only $10 on this trade, your position size should be $10 ÷ 3.5% ≈ $286 notional, with stop attached.

The reason the rule is 1%, not 5%: at 1% per trade, even a brutal 10-loss streak (a 99% probability event over hundreds of trades) only draws your account down 10%. At 5% per trade, the same losing streak takes 50% off. One is a bump; the other is a career-changer.

What drawdown actually feels like

Drawdown is the percentage decline from your account's peak to its current value. Every strategy has drawdowns — even ones with stellar long-run returns. The question isn't whether you'll have a 15% drawdown; it's what you'll do when you're in one.

The honest answer for most people: badly. They size up to "make it back," widen stops to "give it room," abandon the system that was working in favor of intuition that isn't. The losses compound. By the time they capitulate, the drawdown that started at 15% is at 40% and the strategy has just turned the corner without them.

The discipline: pre-commit to your behavior before the drawdown happens. Decide now, in calm conditions, what you'll do if your account drops 10%. Most professionals halt new trades, review the last 20 fills, and resume only when they understand why the drawdown happened. That's the playbook BABA's circuit breakers are designed to enforce automatically.

Correlation risk — 10 positions ≠ 10 independent bets

If you hold longs on BTC, ETH, SOL, and HYPE at the same time, you don't have four diversified bets. You have one bet on "crypto going up" expressed four ways. When the macro wind shifts — Fed hike, ETF outflow, geopolitical headline — all four positions move together, in the same direction, by similar magnitudes. Your effective risk is roughly 4x what you'd compute from each position in isolation.

BABA's MD AI explicitly tracks open-position correlation via the multi-venue, multi-pillar dispatcher. The engine won't fire a fifth crypto long while you already have four open — the concurrent-position cap is a correlation guard, not a venue cap. Respect it. If you've already got TON short, BCH short, and ETH short open, the engine will tell you a fourth alt short is "redundant exposure" rather than letting you stack risk you don't realize you're taking.

The stop-loss contract

Set the SL before you enter. Place it venue-side, not as a mental note. If the SL is in your head and not in the order book, the SL doesn't exist — it's just intention.

The engine handles this automatically: every BABA signal includes a venue-side conditional SL order placed at the same time as the entry. You don't have to remember; the system remembers. The discipline lesson is that this is the right behavior even when you're trading without the engine.

The other side: do not move the SL adversely once it's set. "Adversely" means: long position, SL moved lower; short position, SL moved higher. Both directions widen the loss. The temptation arrives whenever you're 80% to your stop and the price hasn't broken your thesis yet — "let me just give it a little more room." That's how 1% risk becomes 4% risk. Don't.

When NOT to trade

Some conditions defeat any strategy. BABA's circuit breakers detect most of them, but the operator (you) is the final line:

• Circuit breaker armed. If the engine has armed a CB (cumulative loss kill, venue outage, source-staleness alert) — do not manually override and trade through it. The CB is calibrated; you are not.
• Stale data. If a signal references a price more than 5 minutes old, or a funding rate from a venue that's not currently responding to API calls — Skip.
• FOMC, CPI, NFP days. Macro print volatility is the single largest source of "I had a stop in but it gapped through it" stories. The engine knows the calendar and pauses ahead of high-uncertainty events. If you're trading manually, do the same.
• You're emotionally activated. If you just lost money on a trade and the next one feels personal, that's revenge trading. The win rate of revenge trades is famously bad. Walk away, come back tomorrow, the engine will still be there.

Position sizing rules of thumb

1. Account < $500: Take the engine's default size ($3–$5). Don't size up until you have 20+ closed trades to evaluate.
2. Account $500–$5,000: Risk 0.5–1% per trade computed from the SL distance. Adjust the engine's default size up proportionally.
3. Account $5,000+: Risk 0.25–0.5% per trade. At this size, you can afford to be smaller, not larger, because the absolute dollar amounts are larger and emotionally heavier.
4. Account in drawdown > 10%: Halve your usual size until the drawdown is recovered. This is the single most reliable way to not blow up.