Project: BotFire
Project: BotFire
Vision
Automated Kalshi sports betting platform focused on finding and proving profitable bot strategies through rigorous trend analytics. Success = bots that trend upward over time with reliable upside and mitigated risk. Every strategy must prove itself with real P&L data before scaling.
Core principles:
- Data before amplification — don't scale a strategy until analytics prove it has a real dollar-positive edge, not just a win rate
- No bullshit numbers — report actual P&L vs cost basis, never oversell performance
- Test before deploy — verify code works end-to-end before claiming deployed
- Track everything — every bet must have full visibility: entry, exit, P&L, reason, bot, game
- Cut losers fast — if analytics show a bot is underwater, disable it
Current Focus: Live Edge Only
All bots disabled except Live Edge (2026-07-10, PR #10). Analytics proved none of the other bots were dollar-positive. Live Edge contracts doubled from 12 to 24.
Why Live Edge: 85-89c entry band shows 61.5% WR (above 46.6% breakeven), P&L trending upward in second half of trades (+$1.41). Only bot with a positive trend.
Disabled Bots
| Bot | WR | P&L | Why Disabled |
|---|---|---|---|
| Lucas Bot | 17.9% | -$13.95 | Proven loser, no late-game filter |
| Lucas Late Game | 38.5% | -$20.16 | Catastrophic settlement losses at 90c+ entries |
| Consensus Bot | 44.4% | -$5.14 | One -$5.40 settlement loss wiped all gains |
| Portfolio Grouper | 40% | -$2.70 | One settlement loss, too few trades |
| Volatility Bot | — | $0 | 0 fires after rework |
| Marcus Bot | 0% | -$0.42 | Paper mode only |
Architecture
See arch-bots for service inventory, key files, order patterns, and infrastructure.
Rails 8.1 + SQLite, 7 Docker Compose services. IOC orders only,
fill_count > 0 invariant, average_fill_price for all entry/exit tracking.Board
board-bots (BotFire Sprint)User Stories
Board items on
board-bots trace to these stories via story:<name> labels.| Story Label | User Story |
|---|---|
| <code>story:bot-strategy</code> | As a bot operator, I want to identify and exploit profitable trading conditions so that my automated bots generate consistent positive P&L. This covers all Live Edge optimization work — entry price tuning, time-of-day filtering, exit strategy improvements, and new strategy development based on analytics. |
Status
Balance: ~$322. All-time P&L: ~-$44 (mostly from disabled bots). Only Live Edge active at 24 contracts.
Milestones
- M1: Analytics dashboard — DONE (PR #6)
- M1.5: Focus on Live Edge, disable losers — DONE (PR #10)
- M2: Confirm Live Edge edge over 100+ trades
- M3: Build alternative strategies based on Live Edge's success patterns
Repos
| Repo | Platform | Role | Status |
|---|---|---|---|
| ldraney/botfire | Forgejo | Application monorepo | Active |