gavindiaz/hermes_weatherbot
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs: update README highlighting Hermes Agent autonomous learning

Browse Source
This commit is contained in:
John Doe
2026-04-18 18:00:31 +08:00
parent 5d949e491e
commit d9a73618b2
1 changed files with 108 additions and 142 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+108 -142
View File
@@ -1,6 +1,6 @@
# 🌦️ WeatherBet — Polymarket Weather Trading Bot
# 🌦️ WeatherBet — Powered by Hermes Agent
> Autonomous trading bot that exploits weather forecast errors to find mispriced Polymarket prediction markets — and self-improves over time.
> **完全自治的预测市场交易机器人** — 基于 ECMWF 天气预报数据在全链上自动寻找错误定价的 Polymarket 市场进行投注，并借助 **Hermes Agent** 框架实现全自动进化学习。
[![Python 3.13](https://img.shields.io/badge/Python-3.13-blue.svg)](https://www.python.org/downloads/)
[![Polygon](https://img.shields.io/badge/Chain-Polygon%20137-9B59B6.svg)](https://polygon.technology/)
@@ -8,41 +8,44 @@
---
## 🎯 What It Does
## 🤖 为什么选择 Hermes Agent
The bot monitors **6 US cities** (NYC, Chicago, Miami, Dallas, Seattle, Atlanta) and bets on Polymarket's temperature prediction markets using **real ECMWF weather forecasts** as its edge. When the forecast predicts a temperature bucket, but the market price implies a different probability, the bot calculates the Expected Value (EV) and places a trade if EV > threshold.
本项目展示了 **Hermes Agent** 框架在自动化交易领域的强大能力：
| Hermes Agent 特性 | 在本项目中的应用 |
|---|---|
| **自主学习进化 (Self-Learning)** | 机器人从历史交易中自动调整 Kelly 分数和 EV 阈值参数 |
| **全自动化执行 (Autonomous Execution)** | 60 分钟循环扫描市场 → 计算信号 → 自动下单 → 链上结算，全程无需人工干预 |
| **多消息平台接入 (Multi-Platform Gateway)** | 通过 Telegram 实时推送交易通知，手机随时掌控全局 |
| **长期记忆 (Persistent Memory)** | 交易日志 + 学习模型持久化存储，跨会话保留学习成果 |
| **模型无关 (Model Agnostic)** | 可自由切换任意 LLM 提供者进行决策推理 |
| **工具编排 (Tool Orchestration)** | 整合天气预报 API + 链上 CLOB 交易 + Telegram 通知 |
---
## 💡 Why It Makes Money
## 🎯 这个项目做什么
**The edge is weather forecast accuracy.**
机器人监控 **6 个美国城市**（纽约、芝加哥、迈阿密、达拉斯、西雅图、亚特兰大）的天气预报，在 **Polymarket 温度预测市场** 中寻找错误定价机会。
Polymarket traders rely on gut feel and consensus. This bot uses **ECMWF** — the world's most accurate weather model — to calculate the true probability of each temperature bucket, then compares it to the market price.
```
True Probability (from ECMWF) vs. Market Price (from Polymarket)
```
When `Market Price < True Probability`, the market is **underpriced** → BUY.
**核心逻辑：** 当天气预报预测某温度区间的概率与市场隐含概率不一致时 → 计算期望值 (EV) → EV > 阈值则自动投注。
---
## 🧮 The Math
## 🧠 核心技术：Gaussian Bucket Model
### Step 1 — True Probability (Gaussian Bucket Model)
### 第一步 — 从 ECMWF 获取真实概率
```python
import math
def norm_cdf(x):
"""Cumulative distribution function of standard normal — uses math.erf, no scipy needed."""
"""标准正态分布累计分布函数"""
return 0.5 * (1.0 + math.erf(x / math.sqrt(2.0)))
def bucket_prob(forecast_temp, t_low, t_high, sigma=2.0):
"""
The forecast says 72°F ± 2σ.
What's the probability the actual high falls in the 70-75°F bucket?
天气预报给出 72°F ± 2σ
计算实际高温落在 70-75°F 区间的真实概率
P(t_low ≤ X ≤ t_high) = CDF(z_high) - CDF(z_low)
"""
z_low = (t_low - forecast_temp) / sigma
@@ -50,104 +53,98 @@ def bucket_prob(forecast_temp, t_low, t_high, sigma=2.0):
return norm_cdf(z_high) - norm_cdf(z_low)
```
### Step 2 — Expected Value
### 第二步 — 计算期望值 (Expected Value)
```python
def calc_ev(true_prob, market_price):
"""
EV = P(win) × payoff - P(lose) × cost
If EV > 0, the market underprices this outcome.
EV = P(赢) × 收益 - P(输) × 损失
EV > 0 表示市场定价偏低 → 买入信号
"""
win = true_prob * (1 / market_price - 1) # profit if we win
lose = (1 - true_prob) * 1 # we lose our stake
win = true_prob * (1 / market_price - 1)
lose = (1 - true_prob) * 1
return win - lose
```
**Example:**
- Forecast: 72°F → 75% chance of 70-75°F bucket
- Market price: $0.30 (implies 30% probability)
- `EV = 0.75 × (1/0.30 - 1) - 0.25 = +1.25` → **Strong BUY**
**示例：**
- 天气预报：72°F → 70-75°F 区间概率 **75%**
- 市场价格：$0.30（隐含 30% 概率）
- `EV = 0.75 × (1/0.30 - 1) - 0.25 = +1.25` → **强烈买入信号** 📈
### Step 3 — Kelly Criterion (Optimal Bet Size)
### 第三步 — Kelly Criterion 最优投注
```python
def calc_kelly(p, price):
"""
Kelly % = (bp - q) / b
where b = 1/price - 1, p = true_prob, q = 1-p
"""
"""Kelly % = (bp - q) / b，使用 1/4 Kelly 保守分数"""
b = 1.0 / price - 1.0
f = (p * b - (1.0 - p)) / b
return round(min(max(f, 0.0) * KELLY_FRAC, 1.0), 4)
def bet_size(kelly, balance):
return round(min(kelly * balance, MAX_BET), 2)
```
- Uses **1/4 Kelly** (conservative fraction) to survive variance
- Caps bet at `$2.00` per trade
- **Only trades when EV ≥ 10%** (adaptive floor, self-improving)
### Summary: Why This Strategy Wins
| Component | Detail |
|---|---|
| **Edge** | ECMWF weather model is more accurate than consensus |
| **Signal** | Mispriced markets when `Market Price < True Probability` |
| **Sizing** | Kelly Criterion — mathematically optimal bet sizing |
| **Filter** | EV ≥ 10% (adaptive), volume > $500, spread < 3% |
| **Execution** | Real Polymarket CLOB on Polygon (not simulation) |
| **Learning** | Self-tuning Kelly fraction + EV floor from trade history |
---
## 🧠 Self-Learning System
## 🌀 自动进化学习系统
After each trade, the bot records the outcome and adjusts its strategy:
这是 Hermes Agent 框架的核心优势之一 — 机器人**从实战中学习，自动调参**：
```
data/learning/
├── trade_log.json # All trades: city, bucket, cost, outcome, pnl
└── model.json # Learned parameters per city/bucket
├── trade_log.json # 所有交易记录：城市、区间、成本、结果、盈亏
└── model.json # 每个城市/区间的学习参数
```
**Adaptation rules:**
- Winrate < 45% → Kelly fraction ×0.8, EV floor +10%
- Winrate > 55% + PnL > $2 → Kelly fraction ×1.1, EV floor −5%
- Per-city winrate tracking adjusts confidence in each market
- Starts conservative (25% Kelly) → converges to optimal as data accumulates
**自适应规则：**
-胜率 < 45% → Kelly 分数 ×0.8，EV 阈值 +10%
-胜率 > 55% + 盈利 > $2 → Kelly 分数 ×1.1，EV 阈值 −5%
-按城市追踪胜率，调整各市场置信度
-初始保守（25% Kelly）→ 随着数据积累自动收敛到最优
---
## ⚙️ Setup
## 📊 系统架构
### Requirements
```
ECMWF 天气预报 API
↓
Hermes Agent（自主决策引擎）
├── 高斯桶模型 → 真实概率
├── calc_ev() → 期望值计算
├── calc_kelly() → 最优投注
└── 自适应学习 → 自动调参
↓
Polymarket CLOB（Polygon 链上执行）
↓
Telegram（实时推送通知）
```
---
## ⚙️ 快速开始
### 环境要求
- Python 3.13+
- Polygon wallet with USDC.e (on chain 137)
- Polymarket CLOB approval
- Polymarket API credentials
### Installation
- Polygon 钱包 + USDC.e
- Polymarket CLOB 授权
- Polymarket API 凭证
### 安装
```bash
git clone https://github.com/yourhandle/weatherbot.git
cd weatherbot
git clone https://github.com/nicolastinkl/hermes_weatherbot.git
cd hermes_weatherbot
python3.13 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
```
### Configuration
Create `.env`:
### 配置
创建 `.env`：
```env
PK=your_polygon_private_key
WALLET=your_polygon_address
SIG_TYPE=0
```
Edit `config.json`:
编辑 `config.json`：
```json
{
"balance": 0,
@@ -160,96 +157,65 @@ Edit `config.json`:
}
```
### Run
### 运行
```bash
# One-shot scan
# 单次扫描
python bot_v3.py scan
# Continuous trading loop
# 持续交易循环
python bot_v3.py run
# Check status
# 查看状态
python bot_v3.py status
```
---
## 📊 Architecture
## 🛡️ 风险管理
```
bot_v3.py
│
├── Weather Data
│ ├── ECMWF API — 10-day temperature forecast (primary signal)
│ └── METAR — current obs for D+0 override
│
├── Signal Evaluation
│ ├── bucket_prob() — Gaussian model → true probability
│ ├── calc_ev() — expected value vs market price
│ ├── calc_kelly() — optimal bet fraction
│ └── Adaptive floor — self-learning EV threshold
│
├── Execution
│ ├── py_clob_client — Polymarket CLOB on Polygon
│ ├── place_buy_order — market order with 10s timeout
│ └── on-chain settlement
│
├── Monitoring
│ ├── Telegram — real-time trade alerts
│ ├── Self-learning — trade_log + model.json
│ └── 60-min loop — continuous scan
│
└── Market Resolution
└── Outcome check — PnL update when market resolves
```
---
## 🔐 Trading Flow
```
1. Fetch ECMWF forecast for each city (D+0 to D+3)
2. Query Polymarket for temperature bucket markets
3. Calculate true probability (Gaussian model, σ=2°F)
4. Compare to market price → calc EV
5. If EV ≥ adaptive threshold → calculate Kelly bet size
6. Execute market order on Polymarket CLOB (Polygon)
7. Record trade → update self-learning model
8. Send Telegram notification
9. Repeat every 60 minutes
```
---
## ⚠️ Risk Management
| Parameter | Value | Purpose |
| 参数 | 值 | 说明 |
|---|---|---|
| Max bet | $2.00 | Cap per-trade exposure |
| Kelly fraction | 25% | Survive variance (1/4 Kelly) |
| Min EV | 10%+ | Only trade positive EV |
| Min volume | $500 | Avoid illiquid markets |
| Max spread | 3% | Avoid high-slippage markets |
| Adaptive floor | 10-20% | Self-tuning from performance |
| 最大投注 | $2.00 | 单笔交易上限 |
| Kelly 分数 | 25% | 1/4 Kelly 保守策略 |
| 最低 EV | 10%+ | 只交易正期望值 |
| 最低成交量 | $500 | 避免低流动性市场 |
| 最大价差 | 3% | 避免高滑点 |
| 自适应阈值 | 10-20% | 根据表现自动调整 |
---
## 📦 Tech Stack
## 🔐 交易流程（全自动化）
- **Language:** Python 3.13
- **Trading:** [py_clob_client](https://github.com/polymarket/py-clob-client) — Polymarket CLOB
- **Weather:** ECMWF OpenMETAR / Open-Meteo API
- **Chain:** Polygon (Chain ID 137) — USDC.e stablecoin
- **Notifications:** Telegram Bot API
- **Self-learning:** Pure Python JSON persistence (no DB needed)
```
1. 拉取 ECMWF 天气预报（D+0 ~ D+3）
2. 查询 Polymarket 温度区间市场
3. Gaussian 模型计算真实概率（σ=2°F）
4. 与市场价格对比 → 计算 EV
5. EV ≥ 自适应阈值 → 计算 Kelly 投注大小
6. Polymarket CLOB 链上下单（Polygon）
7. 记录交易 → 更新学习模型
8. Telegram 实时通知
9. 每 60 分钟循环
```
---
## 📝 Disclaimer
## 💡 技术栈
This bot trades real markets with real money. Past performance does not guarantee future results. Trade at your own risk. The bot is provided as-is for educational and research purposes.
- **框架：** Hermes Agent（自主学习 + 多消息平台）
- **语言：** Python 3.13
- **交易：** [py_clob_client](https://github.com/polymarket/py-clob-client) — Polymarket CLOB
- **天气：** ECMWF OpenMETAR / Open-Meteo API
- **链：** Polygon（Chain ID 137）— USDC.e 稳定币
- **通知：** Telegram Bot API
- **学习：** 纯 Python JSON 持久化（零依赖数据库）
---
*Built with 🐍 on Polygon — autonomous weather prediction trading.*
## ⚠️ 免责声明
本机器人使用真实资金交易真实市场。过去表现不代表未来结果。风险自担。本项目仅供教育和研究目的。
---
*Built with 🐍 + Hermes Agent on Polygon — Autonomous Weather Prediction Trading.*