How Candle Theory actually works

We run a multi-timeframe moving-average system. Three timeframes, each with a specific role:

Per-timeframe moving averages (closed candles only — we never act on an in-progress candle):

Per-timeframe bias rule:

• Bull if fast EMA > slow EMA and price ≥ long SMA.
• Bear if fast EMA < slow EMA and price < long SMA.
• Neutral otherwise, or whenever the MA gap is below 0.1%.

A bias flip is only published after it persists for at least 2 closed candles. This suppresses single-bar noise but means signals are deliberately late.

The three timeframe biases then compose into one label:

• STRONG BULL
  30m bullish + 4h bullish + 1d not bearish.
• STRONG BEAR
  30m bearish + 4h bearish + 1d not bullish.
• WEAK BULL
  30m bullish but 4h not aligned — fewer of these work out.
• WEAK BEAR
  30m bearish but 4h not aligned.
• COUNTERTREND
  30m flips against the 4h/1d trend — explicitly flagged as a skip.
• NEUTRAL
  No clear setup.

The honest read: STRONG signals are the ones to pay attention to. The rest are context.

Candle Theory scans the top 100 USDT pairs on Binance plus a curated set of commodity perps from Hyperliquid (gold, silver, crude oil, natural gas, copper, uranium and more). Commodities are tagged with a COM badge in the Markets list and can be isolated with the Commodities only filter.

They flow through the same multi-timeframe engine as crypto: 30m entries, 4h trend, 1d regime, same MA rules and guardrails. We only surface commodity markets that are actually executable on a venue we already cover, so every signal you see is tradable.

Note: commodities can move very differently from crypto — lower intraday volatility but sharper macro reactions (CPI, OPEC, weather). Size positions accordingly; the same R-based risk model applies but you may want a smaller risk-per-trade for energy markets.

Four virtual wallets (one per timeframe) start with $10,000 each. They auto-execute long-only spot trades on confirmed signal flips across the top assets. No leverage, no shorts — bearish bias means "exit and hold cash".

The wallets dashboard shows real-time portfolio value plus closed-trade metrics: win rate, average R (avg win ÷ avg loss), max drawdown, and longest losing streak.

We emphasise the 4h and 1d wallets. Lower timeframes have more signals but more noise — they're useful for engine validation, not for setting expectations.

• Late by design. Moving-average systems lag. By the time a bias is confirmed, a chunk of the move is already gone.
• Chop kills them. In sideways markets the engine produces a stream of small losses. The guardrails reduce flips but don't eliminate them.
• Past ≠ future. The public wallets are a live walk-forward, not a backtest you can extrapolate. Sample size is small.
• Execution slippage. Real trades pay spread, slippage, and gas/fees that our virtual engine simplifies. Your fill won't be ours.
• Outages and gaps. Exchange downtime, API errors, or stale candles can delay or skip signals. We retry but cannot guarantee real-time delivery.
• Black swans. No trend system protects against gap-down events, exchange failures, or sudden regulatory shocks.

On the Alerts page you can configure two numbers that get embedded in your Telegram messages:

• Risk per trade % — how much of your wallet you'd lose if the trade hits its stop. Conservative range: 0.5–1%.
• Max daily loss % — a personal circuit breaker. Stop trading for the day once cumulative losses hit this.

These are reminders, not enforcement. Candle Theory never places trades — sizing and execution are entirely on you.