RSI Strategy with Auto Tuner (PF)

# RSI Auto‑Tuner Strategy — How To Use

This document explains **how to use** the RSI Auto‑Tuner strategy. It intentionally avoids math and implementation details. Follow this as an operating guide.

---

## 1. What This Tool Is For

This strategy helps you:

* Discover **which RSI length works best** on a given ticker and timeframe
* Measure performance using **Profit Factor (PF)**
* Improve RSI performance on noisy markets by **transforming price first**

The auto‑tuner is a **research tool**, not a live trading signal generator.

---

## 2. Two Modes You Must Treat Differently

### Research Mode

Used to explore and discover parameters.

* Auto‑Tune: **ON**
* Parameters are allowed to change
* Results may look very good
* Overfitting risk is real

### Trading Mode

Used for forward testing or live trading.

* Auto‑Tune: **OFF**
* Parameters are fixed
* Behavior is stable and repeatable
* This is the only acceptable mode for live use

**Never trade live with Auto‑Tune enabled.**

---

## 3. Manual Mode (Trading Mode)

Use this after parameters are finalized.

Steps:

1. Set **Auto‑Tune = OFF**
2. Choose:

* Source (raw price or transformed price)
* RSI Length (manual, default 14)
* Oversold / Overbought levels
3. The strategy will:

* Enter long when RSI crosses up through Oversold
* Enter short when RSI crosses down through Overbought
* Flip positions on opposite signals

This mode is predictable and safe for forward testing.

---

## 4. Auto‑Tune Mode (Research Mode)

Use this to find optimal RSI lengths.

Steps:

1. Set **Auto‑Tune = ON**
2. Configure the search range:

* Minimum Length (default 5)
* Maximum Length (default 14)
* Step Size (default 1)
3. The strategy will:

* Internally simulate trades for each RSI length
* Track gross profit, gross loss, and trades
* Select the length with the highest Profit Factor
4. The best length is applied automatically

Auto‑Tune evaluates historical data only.

---

## 5. Using a Transform on Price (Critical)

RSI does **not** have to run on raw price.

You can significantly improve results by:

* Applying a **price transform** first
* Feeding the transformed series into the RSI Source input

Examples of transforms:

* Moving averages
* Low‑pass filters
* Butterworth filters
* Any smoother or denoiser

Why this works:

* Busy, wicky markets cause RSI to whipsaw
* Transforms remove micro‑noise
* RSI responds to structure instead of chaos
* Profit Factor often increases dramatically

Best practice:

* Auto‑tune on raw price
* Auto‑tune on transformed price
* Compare PF, trade count, and stability

---

## 6. Reading the Status Label

At the last bar, the on‑chart label shows:

* Whether Auto‑Tune is ON or OFF
* Whether candidates were built successfully
* Number of RSI lengths tested
* Best RSI length found
* Profit Factor and trade count

If Auto‑Tune is OFF, the label shows the manual length.

---

## 7. Recommended Workflow

1. Choose ticker and timeframe
2. Enable Auto‑Tune on **raw price**
3. Record best RSI length and PF
4. Enable Auto‑Tune on **transformed price**
5. Compare results
6. Lock parameters
7. Disable Auto‑Tune
8. Forward test

---

## 8. Warnings and Discipline

* High PF with few trades is unreliable
* Transforms can hide execution costs
* Always validate on a different period
* Auto‑Tune is a **lens**, not an edge

Treat this tool as a research microscope, not an autopilot.