ICC trading strategy: indication, correction, continuation

Where the ICC terminology comes from

ICC is associated with Trades By Sci's educational material. His official YouTube channel publishes Trading Course Day 4: Entries; its description references the ICC community. That identifies an educational source, but it does not establish a standardized algorithm or independently verify trading results. The linked TradingView publications show why the version matters. TradesByBrev describes a confirmed swing, a correction zone based on a candle before the push, and a later close beyond the swing. The ICC - SCI publication by lucianasouzaramos describes a pullback to the indication level followed by a reclaim. These are authors' implementations published on TradingView, not a TradingView-endorsed standard. Neither description justifies treating all ICC indicators as interchangeable.

Turn the three phases into observable conditions

Before using ICC strategy trading rules, decide what each phase must leave on the chart. A label such as strong indication or healthy correction is not enough for two independent tests to agree. Also record when the information became available; a swing identified later cannot authorize an earlier entry.

One illustrative long-only ICC specification

This is a deliberately narrow research variant on standard 15-minute candles. Its 20-bar window, time limits, and 2R target are hypothetical choices. They are not attributed to Trades By Sci or either community indicator. Fix a symbol, data feed, session, timezone, and test dates before evaluating it. Evaluate states at completed bar closes, allow one position at a time, and ignore new indications while a setup or position is active. A new cycle may start only on a later bar after a cancellation or exit.

• Indication: with no active setup, require the close to exceed the highest high of the previous 20 completed bars. Freeze that high as B and the lowest low of the same window as L. Reject the indication if its own low is at or below L.
• Correction: within the next 10 bars, take the first bar whose low touches or crosses B while its close remains above B. It must be a later bar than the indication. Freeze this correction bar's high as C and its low as Q.
• Continuation: within the next 5 bars after correction, require a completed close above C. The correction bar cannot trigger its own continuation. Submit one market entry after that close.
• Cancellation takes priority over entry: from the bar after indication, cancel if a bar closes at or below B or its low reaches L. After correction, also cancel if a bar's low reaches Q minus one minimum tick. Cancel when either waiting period expires without its required event.
• Exit: use Q minus one minimum tick as the initial stop. Set a target two times the distance from actual fill to that stop above the fill. If a gap produces a fill at or below the stop, close the invalid position rather than calculate a negative risk distance. Add a time exit after 20 completed bars in the position, counting the fill bar as the first.
• Sizing: choose a fixed quantity for the first comparison and state its currency exposure. A later risk-based version must account for point value, quantity increments, and costs; it is a separate test.

A hypothetical setup and its failure branches

Suppose B is 100 and L is 96, and an indication bar closes at 102. A later correction bar has a low of 99.8, a high of 101.4, and a close of 100.6. It touches B but closes above it, so C becomes 101.4 and Q becomes 99.8. A subsequent bar closes at 101.6 without hitting a cancellation level. That completes the example signal. With an assumed minimum tick of 0.1 and a next-bar fill of 101.7, the stop is 99.7. The initial distance is 2.0 price units, so the 2R target is 105.7. These are invented prices for arithmetic, not historical prices or a performance result. Fees and slippage change the net outcome.

Check timing before comparing backtest results

TradingView documents two relevant timing limits. A pivot needing right-side bars is known only after those bars finish; moving its label back to the pivot does not make it available earlier. An unfinished higher-timeframe candle can also change. If you add hourly context to a 15-minute rule, specify which completed hourly values are permitted. Under the usual bar-close strategy calculation, a market order created at the close fills at the next available tick, normally the next bar's open in a historical test. Keep that fill separate from the signal price. Model commission and slippage. If stop and target fall inside one bar, inspect the broker emulator's assumptions and whether Bar Magnifier changes the result. Use standard candles for the main test.

• Audit a few setups bar by bar: indication timestamp, frozen levels, correction timestamp, cancellation checks, signal, and fill.
• Compare one rule change at a time, such as the correction timeout. Keep the symbol, session, costs, and sizing fixed.
• Report trades, drawdown, and net results alongside the tested version. A high win rate alone does not establish positive expectancy.
• Reserve later data for a holdout test. Choosing the best historical setting is not evidence that it will remain best.