Three bots. One mission.
Each bot has a distinct role. Together, they address the main market regimes.
Macro Bot
60%Manages the core exposure across market regimes, with confirmed entries and adaptive protection.
DCA Bot
10%Accumulates Bitcoin progressively and recycles profitable lots only.
Adaptive Shield
30%Adds separate upside exposure in rising markets and capped protection in falling markets.
How the strategy works.
Five risk rules before automating crypto
API keys, capital, manual intervention, drawdowns and backtests: the principles to understand before any automation.
Layer 1 — Regime and reinforced bull confirmation
The Macro Bot continuously analyzes long-term trend and momentum to classify the market. A BULL regime no longer triggers an entry on its own: an independent market-strength confirmation must also be present. This gate targets bull traps created when a falling moving average catches up with a ranging BTC price. Exact parameters remain internal.
Layer 2 — Overextension de-risking
When BTC moves too far away from its long-term trend, the bot progressively reduces exposure instead of staying fully invested in a statistically fragile zone. Exact thresholds remain internal: the public goal is simple to understand, de-risk when the market becomes excessive and preserve capital for later phases.
Layer 3 — Reserve and partial hedge
After a long opens, 15% of the sleeve remains in stablecoin (USDT or USDC depending on the venue) to fund an isolated 3x futures hedge capped at 40% of the long after a confirmed drawdown. In parallel, the protective exit follows the cycle high and tightens in stages as gains grow. The reserve belongs to the relevant bot and never uses DCA or another bot's capital.
Layer 4 — Bear market rally shorts
In BEAR regime, the bot detects failed rallies and shorts them with a calibrated sleeve, filtered by RSI, SMA200 and drop confirmation. This layer generates returns even in declining markets.
How the DCA Bot works.
Layer 1 — Regime-aware adaptive accumulation
The DCA Bot doesn't just buy at regular intervals. It adapts its purchases to the macro regime detected by the Macro Bot. In stable BULL, it accumulates moderately. In TRANSITION or BEAR, it accelerates buys as drawdown from the last peak deepens. An "ATH guard" protection prevents buying at the top: the bot abstains if the price is too close to the last peak, with a threshold adjusted to the regime (more permissive in BULL, stricter in BEAR).
Layer 2 — Intelligent capital recycling
When recycling conditions are met, the bot does not mechanically sell the whole DCA BTC inventory. It reviews the inventory lot by lot and only allows recycling on lots whose buy price is strictly below the current price. Non-profitable lots remain held. This rule frees capital when part of the inventory is profitable, without intentionally selling a lot at a loss.
Layer 3 — Operational safety (crash-recovery)
The DCA Bot includes a post-crash recovery state machine. If an abnormal restart corrupts internal state (for example, an inconsistent cost basis relative to held positions), the bot automatically detects the anomaly, enters protected mode, refuses to trade, and immediately alerts the administrator. Rather than continuing to trade on potentially faulty data, it fails loudly and preserves client capital. This logic is inspired by Macro Bot's robustness patterns.
How the Adaptive Shield works.
Layer 1 — Synchronized BULL extension
In a BULL regime, Shield enters only when the Macro long is actually open. It uses its own 30% sleeve, invests 85% and keeps 15% in reserve. It independently applies the same active chain as Macro: isolated hedge capped at 40%, safety exit, adaptive cycle-peak protection and progressive de-risking. If the Macro long closes, Shield closes its legs before any new exposure. Exposure and accounting remain separate.
Layer 2 — Adaptive BEAR protection
In BEAR, the bot analyzes completed daily closes only. It may stay in cash or open a capped short position sized from trend and volatility. A stress layer can reinforce protection during a confirmed rapid decline. The decision executes only on the following cycle.
Layer 3 — Capped and monitored futures risk
In BEAR, only a capped fraction of the Shield sleeve may be exposed through futures. Volatility sizing automatically reduces the target when markets become more unstable. The engine monitors margin, position consistency and runtime state on every supported exchange. These protections reduce liquidation risk but cannot eliminate it.
Fixed allocation & complementarity
Why 60/10/30?
The split is fixed only at inception: 60% Macro, 10% DCA and 30% Shield. Each bot then compounds its own settled net results. Macro carries the primary cycle, DCA accumulates in lots, and Shield adds separate BULL participation or capped BEAR protection.
Validation methodology
Decisions use completed data only and execute in the following period. The conservative publication doubles current exchange taker fees and adds 0.06% slippage to Shield exposure changes. Historical data was used for development; live tracking provides prospective validation.
Why no tight mechanical exits?
The strategy's asymmetric profile requires room for positions to develop. The engine uses bot-specific conditional exits and risk protections instead of a generic trailing stop that could cut favorable moves too early. Exact parameters remain internal.
⚠️ All figures above are from historical backtests from March 15, 2018 to July 27, 2026, using the conservative weakest result across all 8 supported exchanges. Optional lending is excluded. Past performance does not guarantee future results. Invested capital may be partially or totally lost.