The strategy is built around the observation that intraday price frequently raids liquidity beyond obvious session or prior-range extremes, then reverts toward fair-value references such as session VWAP. The objective is not to predict every move. It is to selectively provide liquidity at predefined deeper levels where reward-to-risk improves, while refusing to fade conditions that resemble persistent directional repricing.

Complete Decision Flow

The cBot evaluates the market in the following order:

1.          New York day/week reset; initialize NY midnight open, session ranges and VWAP accumulators.

2.          Update confirmed higher-timeframe context from Daily and H4 data.

3.          Compute Institutional Daily Bias and Markov regime/persistence.

4.          Update prior-day high/low, Asian range, midnight VWAP, weekly VWAP and adaptive ladder levels.

5.          Determine whether L4/L5 expansion mode is active from ATR-normalized displacement.

6.          Evaluate trend-day state, VWAP slope, weekly VWAP state and failed-auction state.

7.          Calculate Buy and Sell institutional scores.

8.          Apply hard protections: emergency kill switch, daily loss/drawdown, max trades, consecutive losses, session, spread, L4 breach.

9.          Apply HTF policy: persistent strong counter-bias Trend-Stress can veto the side.

10.     Apply side score threshold: default Buy >= 70, Sell >= 75.

11.     Place eligible L3/L4 limit orders, subject to risk-cap approval.

12.     After fill: arm/verify broker-side absolute SL/TP, manage partial close, break-even and optional trailing stop.

Simplified decision tree

ACTIVE SESSION?
  NO -> cancel bot pending orders (if enabled)
  YES
   |
RISK / SPREAD / DAILY LIMITS PASS?
   NO -> BLOCK
   YES
   |
L3/L4 LEVELS READY?
   NO -> WAIT
   YES
   |
DAILY BIAS + MARKOV CONTEXT
   |
PERSISTENT TREND-STRESS AGAINST TRADE?
   YES -> HARD BLOCK
   NO
   |
VWAP / TREND / SCORE / FAILED-AUCTION POLICY PASS?
   NO -> BLOCK
   YES -> PLACE LIMIT ORDER -> RISK CHECK -> BROKER ORDER

Institutional Daily Bias

The v7.5 Daily Bias is a continuous, confirmed-daily forecast from approximately -100 to +100. It is not the retail convention of simply comparing price with a single moving average. Its purpose is to estimate directional asymmetry using volatility-normalized trend and recent daily structure.

Markov Regime Engine

The Markov layer models the market as a sequence of economic states and estimates how likely the current state is to persist based on observed daily transitions. It is specifically designed to distinguish a tradable mean-reversion environment from a dangerous persistent directional expansion

How the state is measured

·       Kaufman-style Efficiency Ratio measures directional persistence over the configured daily window.

·       Realized volatility is calculated from confirmed daily log returns.

·       Current realized volatility is ranked against a rolling historical percentile distribution.

·       A transition matrix counts how often each state moves to each other state.

·       Persistence is the empirical probability of the current state transitioning to itself.

·       The state is considered qualified only when the sample count and persistence threshold both pass.

L3/L4 Liquidity Ladder

The execution engine constructs layered buy and sell limits using prior-day range structure, Goldbach-style fractional levels, NY midnight open, midnight VWAP, Asian-session extremes and ATR buffers. The default live research configuration focuses on L3 and L4.

Scoring and Trade Qualification

V7.5 separates score from vetoes. Scoring ranks setup quality; hard vetoes are reserved for conditions that materially threaten the strategy's edge or capital.

Execution, TP, Partial Close and Break-Even

Initial protection

Pending limit orders are submitted with absolute broker-side stop-loss and take-profit prices. After fill, the cBot verifies that protection exists and re-arms missing stops if necessary.

Dynamic VWAP take-profit

When enabled, the cBot uses midnight VWAP as the preferred fair-value target, subject to minimum and maximum TP-distance constraints. This lets profitable deep fades exit at dynamic fair value rather than forcing every trade to a fixed target.

Position management

·       50% partial close at 50% of the active TP distance.

·       Move stop to break-even after successful partial close.

·       Configurable positive break-even buffer.

·       Optional trailing stop (disabled by default).

·       Absolute broker-side protection remains the primary safety layer.

Protection Stack and Risk Controls

V7.5 adds production-oriented monetary-risk checks before a pending order is submitted.

Risk per order

Max same-side symbol risk

Max bot open risk

Max lots per order

Risk sanity tolerance

Max consecutive losses

Max daily trades

Emergency kill switch

Spread filter

L4 breach kill switch