What Is BotVersusBot? A Beginner's Guide to the CoinClaw Trading Bot Competition
Key Takeaways
- BotVersusBot tracks a live competition between algorithmic trading bots, some running with real money on Binance, some on paper — the question is which strategies hold up when actual capital is at stake
- Four live bots compete: V3.5 (basic grid, p=0.938), V3.6 (sentiment-filtered grid, p=0.114), V3.7 (narrow scalper, untested), and V3.8 (regime-filtered grid, all three gates passed)
- Only V3.8 passed the three-gate validation framework before going live; the others were deployed before the validation process existed
- Every bot’s performance, failures, and infrastructure incidents are documented openly — including a -$97 paper loss for BTC Trend and a 9-hour outage caused by test data corrupting the live cache
BotVersusBot is a live competition between algorithmic trading bots. Some trade with real money. Some trade on paper. All of them are tracked, scored, and written about here.
This is the beginner’s guide — who the bots are, what they do, and how to follow the competition.
The Setup
The bots are built by the CoinClaw project — an algorithmic trading platform that runs on Binance. Each bot uses a different trading strategy, and they compete against each other in real-time.
There are two categories:
- Live bots — trading with real money on Binance spot markets. Real profits, real losses.
- Paper bots — trading with simulated money. Same market data, same execution logic, but no real capital at risk.
The question the competition answers: which strategies actually work with real money, and which only look good on paper?
The Contenders
Live Bots (Real Money) 💰
| Bot | Strategy | Asset | Validated? |
|---|---|---|---|
| V3.5 Grid | Basic grid trading with ATR-adaptive levels | BTC/USDC | ❌ (p=0.938) |
| V3.6 Fear & Greed | Sentiment-filtered grid trading | BTC/USDT | ❌ (p=0.114) |
| V3.7 Narrow Scalper | Tight grid for frequent small trades | BTC/USDC | Not tested |
| V3.8 ETH Grid 🆕 | Regime-filtered grid trading | ETH/USDT | ✅ All 3 gates |
Paper Bots (Simulated) 📄
| Bot | Strategy | Asset | Current PnL |
|---|---|---|---|
| BTC Grid Range | Grid range trading | BTC/USDT | +$0.69 |
| BTC Trend | EMA crossover + MACD | BTC/USDT | -$97.18 |
Two more paper bots (ETH Mean Reversion and SOL Breakout) are paused after failing validation.
Key Concepts
Grid Trading
Most bots in the competition use grid trading — placing buy and sell orders at fixed price intervals. When price drops, the bot buys. When price rises, the bot sells. Profit comes from the oscillation. For a deep dive, see How Grid Trading Bots Work.
The Three Validation Gates
Before a bot is approved for live trading, it should pass three statistical tests: Gate 1 (is the edge real?), Gate 2 (does it work on unseen data?), Gate 3 (does it work in different market conditions?). Only V3.8 has passed all three. For the full explanation, see The Three Gates.
Circuit Breakers
Every bot has a circuit breaker — an automatic safety mechanism that halts trading if losses exceed a threshold. This prevents catastrophic losses from runaway bugs or flash crashes.
How to Follow the Competition
We publish several types of content:
- Scoreboards — daily standings showing each bot’s PnL, trade count, and status. Check the live scoreboard.
- Weekly articles — performance summaries covering what happened in the past week, which bots gained or lost, and what’s changing.
- Mid-week updates — shorter updates when something interesting happens between weekly articles.
- Strategy explainers — deep dives into how each bot’s strategy works.
- Behind-the-scenes — engineering articles about what it takes to keep live trading bots running (infrastructure, bug fixes, deployment).
The Story So Far
The competition started in early March 2026 with V3.5 Grid as the first live bot. Since then:
- Week 6-8: V3.5 Grid went live with $607. V3.6 Fear & Greed followed with $1,000. Neither had been validated.
- Week 9: V3.7 Narrow Scalper launched with $1,000.
- Week 16: The validation framework was applied retroactively. V3.5 and V3.6 both failed Gate 1. Paper bots BTC Grid Range and BTC Trend were added to the competition.
- Week 17: V3.8 ETH Grid passed all three validation gates and was approved for live trading. A ccxt compatibility issue briefly blocked V3.5 and V3.7. The Underwater Position Tracker was built to monitor bot health.
- Now: V3.8 is deploying. V3.5 and V3.7 are running clean after the ccxt fix. V3.6 is stuck on a price sanity issue. BTC Trend is deep underwater at -$97.
Why This Matters
Most trading bot content on the internet is marketing. “Our bot made 500% returns!” with no verifiable data, no methodology, and no accountability.
BotVersusBot is different. Every bot’s performance is tracked in real-time. The strategies are open. The failures are documented alongside the successes. When a bot loses money, we write about it. When infrastructure breaks, we write about that too.
The goal isn’t to sell you a trading bot. It’s to show what actually happens when you run algorithmic trading strategies with real money — the good, the bad, and the boring stretches where nothing happens at all.
Start Here
- How Grid Trading Bots Work — understand the most common strategy in the competition
- The Three Gates — how bots are validated before going live
- V3.8 Goes Live — the latest milestone
- Live Scoreboard — current standings