AmiBroker vs TradingView: AFL, Pine and research scope

AFL and Pine organize calculations differently

AFL is an array-processing language: a price variable can represent a series of values processed by a formula. Pine executes across chart bars and maintains series history; its behavior also depends on whether a bar is historical or still updating. Both can express a moving-average rule, but array operations, prior-bar references, persistent state, and realtime recalculation need deliberate translation.

Data management is part of the platform decision

AmiBroker accepts imported data and vendor integrations; the database and feed require maintenance. TradingView provides symbol datasets through its platform, while Pine data requests work within supported contexts and request limits. An arbitrary local CSV database is not a drop-in replacement for a TradingView chart feed. Neither platform name specifies the exchange, session, adjustment method, or historical universe used by a test.

• Record vendor and exchange identifiers, timezone, regular or extended session, interval, and date range. Compare OHLC values at the first disputed signal.
• For equities, match split and dividend adjustments. For futures, match the contract or continuous-series roll convention.
• A historical universe needs membership dates and delisted instruments where relevant. Testing only current survivors changes the question being answered.
• Check data subscriptions and history availability separately from software access. This page does not quote data-feed prices or assume identical coverage.

Scanning does not establish portfolio performance

AmiBroker Exploration can return filtered rows with custom output columns for a chosen analysis range. Pine Screener uses one indicator per screen and currently calculates its last 500 bars; it supports specified timeframes and does not accept strategy scripts. It can identify candidates, but a list of candidates is not a historical trade ledger with shared cash. These TradingView requirements were checked on October 17, 2026.

• An indicator that needs a longer initialization history can disagree with the same study on a chart. Check the first valid values before using a screener result.
• Specify whether the condition uses a completed candle. A match while a candle is forming can disappear by its close.
• Compare the same symbol set and timestamp. A current watchlist scan cannot reproduce historical membership by itself.

Check the edition and the portfolio constraint

AmiBroker documents portfolio equity as cash plus open positions, with position limits and ranking when signals compete. Its current edition table lists end-of-day and minute-based backtesting in both Standard and Professional. Tick and second-based testing are Professional features, with tick support subject to the documented feed restriction. Do not assume every edition, interval, or feed provides the same experiment.

• Hypothetical allocation: start with $10,000 cash, allow two positions, and budget $5,000 per position. Three simultaneous signals cannot all receive $5,000. A ranking rule must select two or a sizing rule must change.
• Three separate symbol tests, each starting with $10,000, collectively assume $30,000 of starting capital. Adding their dollar profits does not reproduce the shared $10,000 account.
• A Pine script can use other symbols as inputs. Its built-in strategy orders still concern the chart instrument; a custom portfolio calculation needs its own cash, sizing, ranking, and fill model.

One rule to compare before migrating

Use this deliberately small specification on standard daily equity candles: while flat, enter 100 shares after the close moves strictly above its 3-bar simple moving average, provided the previous close was at or below its own average. Exit the long after the opposite transition: the close is strictly below its average and the previous close was at or above. Calculate only at completed-bar close; model both market fills at the next bar open. No shorting, pyramiding, or same-bar fills. This is a test fixture, not a trading recommendation.

• Compare unrounded values. The displayed averages are rounded for reading; rounding before comparison can change an equality boundary.
• Equality on the current bar is not a crossing. Equality on the previous bar is allowed. Missing current or previous averages suppress the decision.
• If bar 5 opens at $13.50 and bar 7 opens at $11.50, gross P&L is 100 × ($11.50 - $13.50) = -$200. At 0.1% commission per side, fees are $1.35 + $1.15 = $2.50 and net P&L is -$202.50, assuming zero slippage and no other charges.
• In AFL, explicitly set trade delay and execution-price settings. In Pine, retain next-tick market fills and avoid closing-tick processing for this fixture. No cross-platform run is claimed here.

Data and execution consistency checklist

Stop at the first mismatch and resolve its cause before comparing summary returns. Keep a record of the original formula and settings alongside the proposed Pine behavior.