> For the complete documentation index, see [llms.txt](https://docs.trex.trade/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.trex.trade/lab-journals/prehistoric-vaults-revertx.md).

# Prehistoric Vaults — RevertX

📓 **Lab Journal Entry #03**

**Title:** *RevertX: Mean Reversion Under Observation* \
**Author:** *Byteceratops, Vault Systems Engineer* \
**Recovered From:** *Strategy Terminal Δ-9, Signal Behavior Unit*

***

### Purpose

This log outlines the trading behavior and mechanics of the RevertX Strategy. This vault is designed to hunt for mean-reversion between $ETH and $BTC using a dynamic, signal-driven trading model.

***

### Overview

RevertX was developed to hunt price symmetry between $ETH and $BTC — a behavior we’ve observed frequently in the wild. When the relationship between the two assets drifts beyond a defined threshold, the vault opens long/short positions to capitalize on the expected snap-back.

* Target Assets: $ETH / $BTC
* Style: Pairs Trading (Mean Reversion)
* Core Signal: Deviation from regression-based fair value
* Trade Entry Rule: Signal > 1.5× standard deviation

***

### Signal Mechanics

The strategy opens a long or short position when $ETH's price meaningfully diverges from its expected value relative to $BTC. This expected value is calculated via a regression that updates daily using the past 20 days of price data.

> **Signal Formula** 
>
> Signal = ETH Price − (β₁ × BTC Price + β₀) 
>
> Where β₁ ≈ 0.077 and β₀ captures longer-term drift.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXc_0ACijoBBnh_kDX938tRbf-i7M_e8xdauEO0TpEg6hHBMvulh_2GYdK_azXxeKrdQ_BTq57YW290dYVhTCjOHS5QX0mQHCr92TcjbDo1WusWNcM30JWbwwL_lQieHqsxZHOnW?key=9a4vBn7CNDz6E99JbQ8asA" alt=""><figcaption><p><em>Figure 1: Trading signal fluctuates around the fair value (black line). Trades are triggered when the signal breaches the red (long) or green (short) thresholds.</em></p></figcaption></figure>

The RevertX-trained Rexes are conditioned to strike only when this signal breaches ±1.5σ — filtering noise, conserving energy, and maximizing trade precision.

***

### Backtest Results

To observe RevertX in the wild, we deployed it across historical terrain spanning late 2017 to early 2024. Using real market data and a consistent $1,000 trade size, the system completed 372 full excursions — each representing a signal-triggered round-trip trade.

| Metric                 | Result                     |
| ---------------------- | -------------------------- |
| Annualized Return      | \~53%                      |
| Max Drawdown           | $112 (11.2% of trade size) |
| Average Trade Duration | \~5 days                   |
| Win Rate               | 33.6%                      |

Despite a low win rate, the strategy remained net-profitable due to its strong reward-to-risk dynamic, favoring frequent small losses and infrequent large gains.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdDkEzHxT9c-VCgmx5rViQ4Yg_PfvWpxvIH7qc5Wj16Qh_MEy0ATSWfxxIAR4_axK4TMTU5CUHU5lqXJCkLiyq9TCX5kSTleX5O8I0qea41mNmy0TuvJ559sSVourW0KJVizjJA4A?key=9a4vBn7CNDz6E99JbQ8asA" alt=""><figcaption><p><em>Figure 2: RevertX performance over time. Red line shows cumulative PnL. Signal entries are aligned with ETH/BTC divergences.</em></p></figcaption></figure>

***

### Notable Behavior

* RevertX prefers frequent small losses to wait for occasional large wins
* All trades are stat-tested for signal stationarity
* Optimized for short trade cycles and capital rotation

***

### 🧪 REX Technician’s Log

We fed the RevertX-trained Rexes a steady diet of $ETH-$BTC residuals. When the signal roared above threshold, they struck with precision. Most excursions resolved in under 5 days — unless the jungle changed its rules.

*End of Strategy Log*

*Signed: Byteceratops — authenticated* \
*Reviewed by: Dr. Morales, Risk Oversight Unit*

<br>