Risk Management Philosophy

Defense first: any single trade can fail without meaningfully damaging the account.

1. Modest leverage — 3x as a ceiling, not a target

At 3x leverage, a futures position only faces liquidation after roughly a 33% adverse move. Crypto altcoins swing 5–15% intraday, but 30%+ single-day moves against you are rare. Because every position carries a 3–5% stop-loss, the SL does the risk work long before leverage becomes dangerous — leverage amplifies profits within controlled moves, it doesn't determine survival. Higher leverage would shrink the safety buffer while the take-profit stayed the same: asymmetric temptation with no compensating upside.

2. Daily caps — bounding the bleed rate

The daily caps (default $500/day for futures, and the grid daily cap) answer a different question than position sizing: what is the worst a single bad day can do? Even if every signal fails, every stop-loss hits, and every grid breaks out of range simultaneously, the day's damage is bounded and known in advance. Drawdowns stay linear, not exponential — a bad day can cost yesterday's profits, never the account.

3. Small, uniform orders — no single point of failure

Orders start at a $100 baseline for both grids and futures. One blown trade costs a bounded amount, not the book. The compounding rule is the heart of position sizing:

• Take-profit → next order grows +25% — press the edge while it's working.
• Stop-loss → next order shrinks −25% — the market is telling you the signals are wrong right now.

Losing streaks automatically de-risk; winning streaks automatically scale — with a $10 floor and manual override so it never fully stops or runs away.

4. Concentration limits — diversify across coins and time

A maximum of 3 concurrent futures positions means no single coin (or correlated cluster) can dominate the book. Auto-opening at most 1 position per scan run also spreads entries across time — the day's capital is never deployed on one signal at one instant.

5. Every position is pre-armed with an exit

Futures carry a 3% take-profit and 3–5% stop-loss set at entry. Grids have a locked 3% take-profit, a default 3% stop-loss, and a defined range — the bot's lifecycle is decided before money is committed. No position is ever "held and hoped."

6. Regime gating — trade only where you have an edge

The deepest layer is structural: grid bots run only in sideways regimes; futures run only in upward/downward regimes. A grid in a trending market is a slow bleed; a momentum long in a chop market is a chop-shop. Gating each strategy to its favorable regime removes most tail risk before position sizing even matters.

7. Self-correction — the system learns from its mistakes

The learned-weights feedback loop shifts score weight toward indicators that actually predicted winners. Fewer bad candidates are selected at all — fewer bad entries mean less reliance on the downstream defenses.

8. Transparency — no silent losses

Every open and close is reported (performance pages, equity curve, Telegram alerts), history is archived, and the platform is non-custodial. The worst failure modes — silent drawdowns, unexplained losses — can't happen unnoticed.