DEV Community

Sergei Morozov
Sergei Morozov

Posted on

5 Ways TradingView Backtests Lie to You (And How to Fix Each One)

I ran a TradingView backtest that showed 247% annual return. Low drawdown. Clean equity curve. I was ready to go all-in.

Then I turned on realistic settings — commission, slippage, proper position sizing — and the same strategy showed -12%.

Here are the 5 most common ways TradingView backtests deceive you.

1. Zero Commission by Default

Open any strategy in TradingView. Go to Properties. The commission field says 0.

Bybit charges ~0.1% per side (taker). On a strategy that makes 200 trades per month: 0.2% per round trip x 200 = 40% of your capital gone to fees per month.

Fix: Set commission to 0.1% in Strategy Properties before you do anything else.

2. Zero Slippage

TradingView assumes perfect fills at exact prices. When your strategy sends a market order on a low-liquidity altcoin at 3 AM UTC, the order book is thin. Your "buy at $1.00" fills at $1.02.

Professional algo traders discount backtest profits by ~40-60% for live trading.

Fix: Set slippage to 10-20 ticks in Strategy Properties.

3. 100% of Capital Per Trade

The default position sizing is often 100% of equity. No professional trader puts everything into a single trade. One bad fill and you're done.

Fix: Set order size to 5-10% of equity. If your strategy doesn't work with proper position sizing, it doesn't have an edge — it has leverage.

4. Non-Standard Charts

Heikin Ashi, Renko, Range bars — they modify OHLC values. Your strategy executes at prices that never existed on the actual exchange.

Heikin Ashi open = (previous HA open + previous HA close) / 2. This is a smoothed value, not a real price.

Fix: Always backtest on STANDARD OHLC candles. Use HA/Renko for visual analysis only.

5. Repainting Indicators

Some indicators recalculate past bars using data that wasn't available at the time. The backtest shows "perfect" entries that never existed in real-time.

Fix: Test any indicator live on a 1-minute chart for a few hours. If signals shift after the bar closes, it's repainting. Don't use it.

My Workflow Now

  1. Backtest on STANDARD OHLC with 0.1% commission + 10-20 tick slippage
  2. Forward-test on demo for 2-4 weeks
  3. Compare demo vs backtest — if within 40-50%, the strategy might be real
  4. Go live with small size

We wrote a longer breakdown with more examples on our blog: Why You Shouldn't Trust TradingView Backtests Blindly


What backtesting pitfalls have burned you? Curious what I'm missing.

Top comments (1)

Collapse
 
kiploksrobustnessengine profile image
Kiploks Robustness Engine

Good checklist, the Heikin Ashi point is probably the most underrated one, burns a lot of people.

One thing missing from the workflow: even after fixing all five, you still don't know if the edge is real or curve-fitted to that specific time period. Step 2.5 between demo and live: walk-forward validation, split your history into IS/OOS windows and check if the edge actually transfers. A strategy can pass all your fixes and still be overfit to 2021-2022 BTC conditions.

I built Kiploks (kiploks.com) to automate this part , WFE, OOS retention, Monte Carlo in one report. If you're already doing the commission/slippage work, adding robustness validation is the natural next step.