Probabilistic ICT Order Blocks with a Supervised Price-Context Model

Summary

This indicator identifies bullish and bearish order blocks after market structure breaks, then attaches a probability derived from a supervised classifier trained on recent chart data. Its inputs are RSI, money-flow index, relative ATR, and a volume z-score; its target is whether price will be higher after five bars. The model retrains periodically. The probability describes the broader momentum, volatility, and volume context when a block forms, rather than the likelihood that the specific zone will hold.

Blocks are tracked through a first test and later resolution, with wins and losses displayed in a dashboard. The document warns that the reported zone win rate is structurally asymmetric: a win requires a close beyond one edge, while a loss requires a close beyond the full zone. It therefore measures zone survival, not trade profitability, and cannot directly support an expectancy calculation. The suggested validation is to compare resolved outcomes across probability deciles; the text provides no independent performance results for that test.

Key ideas

• Order blocks are triggered by a close crossing a recent swing high or low, with the zone selected from a preceding candle.\nA periodically retrained classifier uses RSI, money flow, relative volatility, and unusual volume to estimate whether price will rise over its target horizon.\nThe model probability describes conditions when a block forms and is not the probability that the block will hold.\nA block is tested on a return to its zone and resolved by a close beyond a defined edge.\nThe dashboard win rate is not a trade hit rate because its win and loss thresholds are asymmetric.\nComparing outcomes across probability groups is proposed as a way to test whether the model adds useful filtering.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.