Learn Trading

Multiple studies conducted in 2024 and 2025 reveal the staggering extent of retail trading failures in cryptocurrency markets.

The Statistical Reality:
- 84% of retail traders lose money in their first year
- 58% of traders lose nearly all their capital
- 80% of prop trading challenge participants fail
- 78% of liquidations occur in accounts with less than $10,000

Root Causes:
1. Emotional Decision-Making: Retail traders make 47 trades per week on average, driven by FOMO and fear
2. Lack of Systematic Risk Management: Ad-hoc risk management leads to catastrophic losses
3. Overtrading: Excessive trading without statistical edge
4. Absence of Quantitative Validation: No backtesting or validation before trading

Institutional Approach:
- 2-4 trades per week on average
- Statistical edge-based decision making
- Systematic risk management
- 84-94% probability of profit over 5-year horizons

Smooth Brains AI addresses these fundamental shortcomings by providing retail traders access to institutional-grade algorithmic trading strategies previously reserved for elite hedge funds.

Smooth Brains provides system-managed, non-custodial execution on Hyperliquid under trade-only permissions.

What we publish:
- Artifact-backed metrics and trade logs on /performance (historical, not guarantees).
- Business and custody model in the whitepaper.

What we do not publish:
- Proprietary parameters, sizing formulas, or internal pipeline design.

Your primary controls are deposit size, trading ON/OFF, and pause/stop at any time.

Risk is system-managed so execution is consistent and auditable.

What you can rely on:
- Gating: entries can be blocked when conditions are hostile.
- Sizing: position sizing is risk-adjusted and bounded by hard caps.
- Controls: you can pause/disable trading at any time.

What we do not do:
- We do not publish “step-count” sizing checklists or signal tier marketing.

If you want the canonical record of behavior, use the published artifacts on /performance (metrics + trade logs). Past performance is historical and not guaranteed.

Markets move through different regimes (risk-on, risk-off, calm, elevated volatility). A strategy that ignores regimes tends to overtrade and take risk when the odds are structurally bad.

Why regimes matter:
- Volatility changes execution quality (slippage and stop distance).
- Liquidity changes fill probability and tail risk.
- Trend vs chop regimes change the expectancy of the same signal.

How Smooth Brains uses regimes:
- The system applies macro gates (fail-closed) that can block new entries when conditions are hostile.
- Execution uses system-controlled sizing and exposure caps so risk is bounded by design.

If you want to audit how these choices behave historically, use /performance for canonical artifacts (historical, not guarantees).

Live position sizing is deterministic and risk-based. It is a rules engine, not an AI model. Every live trade is sized by the same fixed math.

The Core Calculation:
1. Risk Budget: A fixed amount of risk is allocated to the trade.
2. Stop Distance: The system measures the distance from entry to the protective stop.
3. Base Size: Risk budget ÷ stop distance gives the base position size — tighter stops allow larger size for the same risk, wider stops smaller.
4. Exposure Cap: The base size is then capped by system-enforced exposure limits.

Regime-Aware Risk Controls:
The system reads the current market regime (trend and volatility) and applies rules-based controls that raise or lower how much risk it is willing to take. These are deterministic — fixed rules, not predictions.

What This Means:
- Risk per trade is bounded by design, not by a model's confidence.
- Sizing is fully explainable from the inputs above.
- The same framework applies to every user under system control.

On AI and Machine Learning:
Any neural-network or machine-learning work is exploratory research that runs separately from the live system. It is not used to size or decide live trades. The only AI used in production is a chat assistant (powered by Anthropic Claude) that answers customer questions — it does not make or influence any trade.

Past performance does not guarantee future results.

Monte Carlo simulations are a statistical method used to understand the range of possible outcomes for our trading strategies.

What Are Monte Carlo Simulations?
Monte Carlo simulations run thousands of random scenarios based on historical data to understand:
- Probability of profit: What % of scenarios are profitable?
- Expected returns: Average returns across all scenarios
- Worst-case scenarios: Maximum drawdowns and losses
- Best-case scenarios: Maximum gains

Our Process:
1. 1,000+ Simulations: We run 1,000+ Monte Carlo simulations across multiple market regimes
2. Random Sampling: Each simulation randomly samples from historical market conditions
3. Statistical Analysis: Results are analyzed to determine probabilities
4. Validation: 34 walk-forward validation windows test robustness

What We Learn:
- Win Probability: Distribution of outcomes across thousands of scenarios
- Expected Returns: Range of outcomes across different market regimes (net after fees)
- Risk Metrics: Maximum drawdowns, Sharpe ratios, profit factors
- Confidence Intervals: Range of likely outcomes

Why It Matters:
Backtesting shows what happened in the past. Monte Carlo simulations show what could happen in the future across thousands of scenarios, giving us confidence in strategy robustness.

Result:
Our strategies are validated with rolling historical backtests, plus Monte Carlo and walk-forward analysis where applicable.