$100,000 Capital - Automatic Trading Bot: First Week Performance (Python, API)
Short summary
**P/L:** **+$1,782** (≈ **1.78%**) in the first week.
**Context:** I’ve run this bot personally for 1 year with an average of **\~8% monthly**. I launched a public YouTube challenge and will run it for a full month to show statistical consistency. In my intro video I stated a target range of **8–16% monthly**; one week is too small a sample to judge.
How the system works
* **Tech stack:** Written in **Python**, connects via exchange **API**, state persisted in **JSON**; codebase ≈ **5,000 lines**.
* **Structure:** Runs **10 profiles** concurrently (can scale to 20+). Each profile has its own config: **timeframes** (1m, 3m, 5m, 15m), **RSI** thresholds, and **take-profit** levels. Some profiles use **partial laddered exits**, others full exits.
* **Adaptive RSI:** Instead of fixed 30/70, thresholds adapt based on **position size** and **market trend**.
* **Staged entries / position management:** The bot builds positions in small portions (scalping style). If earlier portions lag, the system gradually brings them closer via calculated new entries to improve average entry price and the chance of closing profitably. It respects a **global take-profit cap** and will not exceed it. The farther the distance, the faster older portions are brought closer.
* **Risk allocation:** Each profile uses a small percentage of the balance (e.g., **0.05%–0.1%** per profile) to prevent any single profile from blowing up the account.
* **Diversification:** Profiles differ in aggressiveness, timeframes, and directional bias (some favor longs, some shorts, some inverse). This sleeve-like diversification helps the system survive adverse events (e.g., October crash) because losing profiles are often offset by winners
It does not risk the entire capital at one price but gradually builds the position and minimizes the average entry price. the system works so that the bot trades in small portions according to its conditions. When older portions lag behind, mathematical calculations gradually bring them closer to the price to increase the chance of closing. The aim is to ensure, gradually and with mathematical guarantees, that the global take-profit percentage is not exceeded and losses are avoided by comparing old and new portions. It considers the global take-profit percentage and can only use up to 70% of it, also taking into account the size of the new portion. The greater the distance, the faster it brings the old portions closer to the price. The bot trades scalping, so it needs time to manage these portions.
Results analysis and perspective
* **Week 1:** \+$1,782 → **\~1.78%**. Emotionally it may feel small versus the 8–16% monthly target, but weekly returns are noisy.
* **Why this matters:** If the system can reproduce consistent weekly gains with controlled drawdowns, monthly and annualized returns compound meaningfully. The one-year personal track record averaging **\~8% monthly** is the stronger signal; the public challenge is to demonstrate that live.
* **Risk/return tradeoff:** The bot prioritizes steady, repeatable gains and drawdown control over chasing large one-off spikes.
Technical notes and limitations
* **Timeframes used:** 1m, 3m, 5m, 15m.
* **Profiles:** 10 independent strategy profiles in one account/pool.
* **Exits:** Mixed partial and full exits; exit logic checks balance and per-profile allocation before opening new trades.
* **Persistence and recovery:** All necessary state is stored in JSON so the bot can resume after restart.
* **Operational risks:** API outages, latency, slippage, and extreme market moves can still cause losses. Scalping and staged entries require reliable execution and monitoring.
* **Statistical caution:** One week is insufficient to validate the 8–16% monthly claim; at least one full month (preferably several) is needed for meaningful statistics.
Transparency and what I will / will not share
This is a technical, non-promotional post with P/L and methodology context. I will **not** post ready-made, working strategy code publicly here. I can explain algorithmic logic and key design choices at a high level (adaptive RSI concept, portion-management math, risk allocation rules) but I will not publish the full, deployable strategy.
If you want deeper details
If you want a focused explanation, tell me which of these you want more detail on and I’ll provide a high-level description (no copy-paste strategy code):
* **Adaptive RSI mechanics** and how thresholds shift with position size and trend.
* **Portion management math** for bringing old portions closer to price.
* **Partial exit logic** and how global take-profit is enforced.
* **Risk allocation** per profile and how it prevents blowups.
https://preview.redd.it/9e5rw43c22gg1.png?width=954&format=png&auto=webp&s=e11d10eb1f0310cfb719959c072795d341f03c96
https://preview.redd.it/fgbbdqqd22gg1.png?width=1189&format=png&auto=webp&s=16edb2154563fde0b1f7ee2adb766ba0ef080b80
https://preview.redd.it/uc48o6yg22gg1.png?width=1385&format=png&auto=webp&s=666b81ac1b5d6a445408b73fa372ade0901883bf