Backtesting strategies in TradingView: Pitfalls

Backtesting Strategies in TradingView: Pitfalls

The Trap of Illusory Profitability

TradingView is the de facto standard for technical analysis and trading system development today. Its built-in Strategy Tester module allows any user to evaluate

Starting out is always easier when you’re allowed to make mistakes. Hone your trading skills on MEXC’s virtual balance without risking real losses. Give it a try: https://promote.mexc.com/r/aep0hTSdh1 #ad

a trading algorithm’s efficiency on historical data in seconds. However, this accessibility hides the main danger. Beginner traders often fall into the trap of perfect equity curves, where the capital line heads toward infinity at a 45-degree angle. In reality, most of these results are the product of methodological errors and technical platform limitations, rather than the genius of the strategy. For a professional analyst, it is crucial to understand that backtesting is not a search for validation, but an attempt to disprove the viability of a system under conditions as close to market chaos as possible.

The Repainting Phenomenon in Pine Script

One of the most insidious problems is repainting. This occurs when an indicator or strategy changes its historical signals based on data that became available later. In Pine Script, this frequently happens when using functions that call higher timeframes via the request.security parameter if look-ahead is not correctly handled. On the chart, this looks perfect: the algorithm always knows where the local high or low was. However, in real-time trading, a signal may appear and then vanish as the price moves in the other direction. To combat this, one must use the barmerge.lookahead_off qualifier and carefully verify entry logic to ensure the signal is only confirmed after the candle closes.

Hidden Look-ahead Bias

Look-ahead bias is a methodological error where an algorithm uses information from the future to make decisions in the past. In TradingView, this manifests not only in repainting but also in order execution logic. For example, if a strategy is programmed to enter a trade at the candle close price but uses conditions that were only known at that exact moment, the execution should technically occur at the opening of the next bar. Ignoring this one-candle nuance on lower timeframes can completely distort results, turning a losing system into a profitable one. A professional test always assumes an execution delay of at least one tick or uses limit orders with verification of their reachability within the bar.

Neglecting Trading Costs

Many traders forget that the figures in the Strategy Tester are sterile data. By default, TradingView may not account for broker commissions and slippage. In high-frequency strategies or scalping, commissions can eat up to 50-80% of potential profit. Slippage is a critical factor for illiquid assets or during moments of high volatility. If your strategy shows an average profit per trade of 0.1% and total costs amount to 0.05%, you are operating on the edge of your mathematical expectation. Without configuring the Slippage and Commission settings in the tester, any backtest results can be considered invalid.

The Intra-bar Precision Problem

The standard TradingView tester operates based on OHLC prices. It does not know what happened inside the candle: whether the high was reached before the low, or vice versa. This is critical for strategies where Take Profit and Stop Loss levels are reached within the same bar. The system might incorrectly credit a profit, even if in reality the price hit the stop-loss first.