The 20-Day Channel Rule: How Our Bot Decides a Breakout Is Real

A green candle is not a signal. It's a rumor. Our paper-trading bot's majors and mid-cap tiers run on one of the oldest trend ideas in trading — the Donchian channel, a band drawn from the highest high and lowest low of the last 20 periods — but a bare breakout isn't enough. Every hour, each entry must pass three gates before the bot commits a single paper dollar.

Gate 1 — Break the 20-period high

Price must print above the highest high of the previous 20 periods. That's the Donchian-20 breakout: it means the market is doing something it hasn't done in nearly three weeks. Fresh ground, not noise inside a range.

Gate 2 — Volume must confirm (1.5x median)

Gate 3 — The trend must agree

On the 4-hour timeframe, price must sit above its 200-period average, and that average must be rising. This is the regime filter: breakouts inside a larger downtrend are statistically the worst kind. The bot only buys breakouts that go with the bigger trend, never against it.

The guardrails around every entry

• Volatility guard: the stop distance (2× average true range) can't exceed 10% of the price. Too violent? Skipped.
• Position sizing: the SAFE tier (BTC, ETH, SOL) risks 1.0% per trade, the MEDIUM tier (majors like ADA, DOT, LINK) risks 0.5% — each with a maximum of 2 positions.
• 5-day time stop: if a trade hasn't worked within five days, the bot closes it. Capital sitting still is capital dying.
• Fees modeled: 0.80% per side on Kraken spot data — because a strategy that only works before fees is a hobby, not a strategy.

Three gates, one guardrail stack — and a published paper-trading record showing exactly where it still loses. That's the point of running it in the open: you get the rules and the receipt.