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

Update README.md

Browse Source
This commit is contained in:
alteregoeth-ai
2026-03-09 02:26:29 -05:00
committed by GitHub
parent 60cfeaa81c
commit 3fd76b3f03
1 changed files with 78 additions and 52 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+78 -52
View File
@@ -1,6 +1,6 @@
# 🌤 Weather Trading Bot — Polymarket # 🌤 Weather Trading Bot — Polymarket
Automated weather market trading bot for Polymarket. Finds mispriced temperature outcomes using Open-Meteo forecasts. Automated weather market trading bot for Polymarket. Finds mispriced temperature outcomes using real station data from NWS and Visual Crossing.
No SDK. No black box. Pure Python. No SDK. No black box. Pure Python.
@@ -9,36 +9,66 @@ No SDK. No black box. Pure Python.
## Versions ## Versions
### `bot_v1.py` — Base Bot ### `bot_v1.py` — Base Bot
The foundation. Scans 6 cities, fetches forecasts from Open-Meteo, finds matching temperature buckets on Polymarket, and enters trades when the market price is below the entry threshold.
The foundation. Scans 6 US cities, fetches forecasts from NWS, finds matching temperature buckets on Polymarket, and enters trades when the market price is below the entry threshold.
No math, no complexity. Just the core logic — good for understanding how the system works. No math, no complexity. Just the core logic — good for understanding how the system works.
### `bot_v2.py` — Kelly + EV Edition *(current)* ### `bot_v2.py` — Kelly + EV Edition
Everything in v1, plus: Everything in v1, plus:
- **Expected Value** — skips trades where the math doesn't work - **Expected Value** — skips trades where the math doesn't work
- **Kelly Criterion** — sizes positions based on edge strength, not a flat % - **Kelly Criterion** — sizes positions based on edge strength, not a flat %
- **Live monitor** — updates `simulation.json` every 10s so the dashboard stays current
- **Auto-exit** — closes positions when price hits the exit threshold - **Auto-exit** — closes positions when price hits the exit threshold
- **Live dashboard** — updates `simulation.json` so the dashboard stays current
~400 lines of pure Python. ~400 lines of pure Python.
### `bot_v3.py` — Coming Soon ### `bot_v3.py` — Auto-Cycle + Forecast Monitoring (current)
- Auto-cycle every hour — continuously scans and opens new positions
- Forecast monitoring every 10 minutes — closes positions if the forecast changes and EV goes negative Everything in v2, plus:
- Historical calibration — uses 10 years of weather data to validate probabilities
- **Auto-cycle** — scans every hour, synchronized to the clock (:00, :01, :02...)
- **Forecast monitor** — checks every 60 seconds, closes positions if forecast changes or EV goes negative
- **Real station data** — NWS hourly observations + Visual Crossing for today's actual readings
- **Correct airport coordinates** — each city mapped to the exact station Polymarket resolves on
- **Liquidity-aware sizing** — reduces position size for low-volume markets, hard cap at $20 for markets under $1k volume
- **Slippage simulation** — entry price reflects real market impact
- **16 cities** — 6 US cities via NWS, 10 international via Open-Meteo
--- ---
## How It Works ## How It Works
Polymarket runs markets like *"Will the highest temperature in NYC be between 40–41°F on March 4?"* These markets are often mispriced — the forecast says 78% likely, but the market is trading at 8 cents. Polymarket runs markets like "Will the highest temperature in Chicago be between 46–47°F on March 7?" These markets are often mispriced — the forecast says 78% likely but the market is trading at 8 cents.
The bot: The bot:
1. Fetches 4-day forecasts from **Open-Meteo** (free, no API key)
2. Finds the matching temperature bucket on Polymarket 1. Fetches forecasts from NWS (US cities) and Open-Meteo (international) using airport coordinates
3. Calculates Expected Value — skips the trade if EV is negative 2. Combines real station observations with hourly forecast to get the true daily maximum
4. Calculates Kelly Criterion — sizes the position based on edge strength 3. Finds the matching temperature bucket on Polymarket
5. Runs a full $1,000 simulation against real market prices before you risk anything 4. Calculates Expected Value — skips the trade if EV is below threshold
5. Calculates Kelly Criterion — sizes position based on edge strength
6. Adjusts position size for market liquidity and simulates slippage
7. Monitors all open positions every 60 seconds — closes if forecast shifts
---
## Why Airport Coordinates Matter
Most bots use city center coordinates. That's wrong.
Every Polymarket weather market resolves on a specific airport station. NYC resolves on LaGuardia (KLGA), Dallas on Love Field (KDAL) — not DFW. The difference between city center and airport can be 3–8°F. On markets with 1–2°F buckets, that's the difference between the right trade and a guaranteed loss.
| City | Station | Airport |
|------|---------|---------|
| NYC | KLGA | LaGuardia |
| Chicago | KORD | O'Hare |
| Miami | KMIA | Miami Intl |
| Dallas | KDAL | Love Field |
| Seattle | KSEA | Sea-Tac |
| Atlanta | KATL | Hartsfield |
--- ---
@@ -56,7 +86,7 @@ EV = (our_probability × net_payout) − (1 − our_probability)
Kelly % = (p × b − q) / b Kelly % = (p × b − q) / b
``` ```
We use fractional Kelly (25%) and cap each position at 10% of balance. We use fractional Kelly (25%) and cap each position at 5% of balance.
--- ---
@@ -65,7 +95,7 @@ We use fractional Kelly (25%) and cap each position at 10% of balance.
```bash ```bash
git clone https://github.com/alteregoeth-ai/weatherbot git clone https://github.com/alteregoeth-ai/weatherbot
cd weatherbot cd weatherbot
pip install requests pip install requests pytz
``` ```
Add your settings to `config.json`: Add your settings to `config.json`:
@@ -74,7 +104,8 @@ Add your settings to `config.json`:
{ {
"entry_threshold": 0.15, "entry_threshold": 0.15,
"exit_threshold": 0.45, "exit_threshold": 0.45,
"locations": "NYC,Chicago,Seattle,Atlanta,Dallas,Miami", "max_position_pct": 0.05,
"locations": "nyc,chicago,miami,dallas,seattle,atlanta",
"max_trades_per_run": 5, "max_trades_per_run": 5,
"min_hours_to_resolution": 2 "min_hours_to_resolution": 2
} }
@@ -85,27 +116,19 @@ Add your settings to `config.json`:
## Usage ## Usage
### bot_v1.py ### bot_v1.py
```bash ```bash
# Scan markets and show signals python bot_v1.py # paper mode — shows signals, no trades
python bot_v1.py python bot_v1.py --live # simulates trades with $1,000 balance
python bot_v1.py --reset # reset balance back to $1,000
``` ```
### bot_v2.py ### bot_v3.py
```bash ```bash
# Paper mode — shows signals + Kelly/EV analysis, no trades python bot_v3.py --live # run bot (entry scanner + forecast monitor)
python bot_v2.py python bot_v3.py --positions # show open positions and PnL
python bot_v3.py --reset # reset simulation back to $1,000
# Simulation mode — executes trades, updates virtual $1,000 balance
python bot_v2.py --live
# Live monitor — updates dashboard every 10s
python bot_v2.py --monitor
# Show open positions and PnL
python bot_v2.py --positions
# Reset simulation back to $1,000
python bot_v2.py --reset
``` ```
--- ---
@@ -118,12 +141,12 @@ Run a local server in the bot folder:
python -m http.server 8000 python -m http.server 8000
``` ```
Then open `http://localhost:8000/sim_dashboard.html` in your browser. Then open `http://localhost:8000/sim_dashboard_repost.html` in your browser.
- Balance chart with floating +/- labels on each trade - Balance chart with history
- Open positions with Kelly %, EV, and price progress bar - Open positions with Kelly %, EV, and current PnL
- Full trade history with W/L tracking - Full trade history with ENTRY/SELL labels
- Refreshes every 10 seconds automatically - Refreshes automatically every 10 seconds
--- ---
@@ -133,12 +156,25 @@ Then open `http://localhost:8000/sim_dashboard.html` in your browser.
|-----------|---------|-------------| |-----------|---------|-------------|
| `entry_threshold` | `0.15` | Buy below this price | | `entry_threshold` | `0.15` | Buy below this price |
| `exit_threshold` | `0.45` | Sell above this price | | `exit_threshold` | `0.45` | Sell above this price |
| `locations` | `NYC,...` | Cities to scan | | `max_position_pct` | `0.05` | Max position size as % of balance |
| `max_trades_per_run` | `5` | Max trades per run | | `locations` | `nyc,...` | Cities to scan (comma separated) |
| `max_trades_per_run` | `5` | Max new trades per scan |
| `min_hours_to_resolution` | `2` | Skip if resolves too soon | | `min_hours_to_resolution` | `2` | Skip if resolves too soon |
--- ---
## APIs Used
| API | Auth | Purpose |
|-----|------|---------|
| NWS (api.weather.gov) | None | US city forecasts + station observations |
| Open-Meteo | None | International city forecasts |
| Visual Crossing | Free key | Today's actual hourly readings by station |
| Polymarket Gamma | None | Market data |
| Polymarket CLOB | Wallet key | Live trading (optional) |
---
## Live Trading ## Live Trading
The bot runs in simulation mode by default. To execute real trades, add Polymarket CLOB integration: The bot runs in simulation mode by default. To execute real trades, add Polymarket CLOB integration:
@@ -147,17 +183,7 @@ The bot runs in simulation mode by default. To execute real trades, add Polymark
pip install py-clob-client pip install py-clob-client
``` ```
Then replace the paper mode block in `bot_v2.py` with your CLOB buy function. Full guide in the article linked below. Then replace the paper mode block in `bot_v3.py` with your CLOB buy function. Full guide in the article linked below.
---
## APIs Used
| API | Auth | Purpose |
|-----|------|---------|
| Open-Meteo | None | Weather forecasts |
| Polymarket Gamma | None | Market data |
| Polymarket CLOB | Wallet key | Live trading (optional) |
--- ---