# Predictive Extrema, Unprofitable Policies: An AI-Assisted Audit of Candle-Based Binance Spot Timing Models

arXiv: https://arxiv.org/abs/2607.19453 · HTML: https://mlsanity.com/paper/2607.19453/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 32 references that are in the map.
- First posted on arXiv 21 Jul 2026.

## Builds on (most cited first, 7 of 7)

- [Selective Classification for Deep Neural Networks](https://mlsanity.com/paper/1705.08500/) (2017)
- [Smart "Predict, then Optimize"](https://mlsanity.com/paper/1710.08005/) (2017)
- [Evaluating time series forecasting models: An empirical study on performance estimation methods](https://mlsanity.com/paper/1905.11744/) (2019)
- [Anticipating cryptocurrency prices using machine learning](https://mlsanity.com/paper/1805.08550/) (2018)
- [Agentic Trading: When LLM Agents Meet Financial Markets](https://mlsanity.com/paper/2605.19337/) (2026)
- [Deep Reinforcement Learning for Cryptocurrency Trading: Practical Approach to Address Backtest Overfitting](https://mlsanity.com/paper/2209.05559/) (2022)
- [Machine Learning-Based Bitcoin Trading Under Transaction Costs: Evidence From Walk-Forward Forecasting](https://mlsanity.com/paper/2606.00060/) (2026)

## Abstract (arXiv)

We audit whether candle-based machine-learning models can turn predictions of cryptocurrency extrema or short-horizon outcomes into positive Binance Spot paper policies after assumed costs. Numerical results come from scripted fixed-seed model runs and deterministic simulators; human-supervised AI agents supported the July 20 evidence-integrity revision through literature retrieval, separately tasked critique, artifact reconciliation, documentation, and source packaging, not trading decisions. The strongest later-period evidence, conditional on extensive predecessor search, is negative: an unchanged ten-pair mandatory-daily selector lost 6.72% over 19 July cycles at an assumed 31-bps completed-cycle cost, with 3 wins and 16 losses. In short model-specific July evaluations, the validation-selected local-minimum policy returned -1.79%, while the local-maximum sell-to-cash/re-entry policy underperformed continuous holding by 2.80%; their gross mean advantages of 11.11 and 12.21 bps were below even the 21-bps stress. A Gurgul-inspired, OHLCV-only daily adaptation attained minimum/maximum ROC AUC of 0.874/0.896 but average precision of only 0.134/0.116 and lost 44.30% over seven cycles, versus -41.20% for buy-and-hold. A forensic audit also downgraded an earlier One4All "30-day holdout": its dates had influenced prior architecture work, its four-hour outcome horizon was not purged at split boundaries, it used same-close entry, and its raw result directories were absent. Across the tested, mostly exploratory protocols, event-ranking performance did not establish positive executable policy value. Every operational decision remains NOTRADE.