DAY 24: How to Backtest That Works for Discretionary Traders

On DAY 23 we discussed how the rule-making process in automated trading (system trading) reduces ambiguity in discretionary trading and essentially provides hints to enhance discretionary power.

Today we will explain in detail the indispensable “manual backtesting” for discretionary traders to verify their rules and methods, including concrete methods and key points to keep in mind.

1. Why is backtesting essential for discretionary traders?

1. You can check the validity of your rules in advance

   • Before moving to real trades, know how your trading rules would have performed in past markets in terms of win rate.
   • Reduce the risk of real money being melted due to rules that were decided somewhat superficially not working.
     
     

2. It trains market sense and mental discipline

   • As you advance through past price movements one by one, you naturally acquire an understanding of chart patterns and price action characteristics.
   • It also trains exit and take-profit timing, improving judgment.
     
     

3. Less deviation

   • Confidence that you will be thorough this time because “the past test showed this line worked well for stop-loss,”which leads to strength of conviction.
   • Even after a streak of losses, if you can think that “past drawdowns were within ○○ consecutive losses,” your mental state stays stable.
     
     

2. Basic steps of manual backtesting

(1) Setting the verification period

• Decide the period you want to verify on the past chart.Example: the last 1 year, 3 years, a period including the Lehman Brothers crisis, etc.
  • If too short, sample size is small; if too long, market conditions may differ too much. Strike a reasonable balance.
    
    

(2) Clarify the verification rules and conditions

1. Time frame: verify on 1-hour chart or 4-hour chart, etc.
2. Entry conditions:
   • Example: buy when the moving average (20SMA) is rising and the candle approaches around -2σ, etc.
   • Quantify where to place stop-loss and how to take profits.
3. How to perform the verification:
   • Advance the chart one bar at a time, enter when conditions match → record the results up to stop-loss/take-profit.
     
     

(3) Progress the chart “one bar at a time” to gain a simulated trading experience

• There are methods such as using the “Replay” feature on chart platforms like TradingView, or scrolling past charts leftward on MT4/MT5 and hiding candles to proceed.
• Be careful not to let past charts be seen in advance, which can cause a forward-looking bias..
  
  

(4) Recording trading results

• Record entry date/time and price, stop-loss and take-profit lines, and final net profit, etc., in a table or Excel.
• Also notemental state and questionsto help improve discretion later.
  
  

(5) Aggregation and analysis

• After completing a basic set of verifications, calculatewin rate, average profit, average loss, maximum drawdown, etc..
• If this rule yields “win rate 60%, risk-reward 1:1.5,” you can roughly judge whether profit is being made.
• Pick out concerning loss patterns or drawdown periods and consider improvements in “stop-loss width and entry conditions.”
  
  

3. Points to be careful about in verification

(1) Pay attention to the visible parts of past data

• When high and low have already been visible, there is a bias toward entering after the fact in a favorable way.
• To avoid thinking “I would have bought if this happened,” reproduce real-time progression with replay features so that bars move forward in real time.
  
  

(2) Don’t feel safe by looking only at periods of consecutive wins

• Verifying only a few months of trending markets may show high win rates, which can be misleading.
• Include ranging and volatile periods in the overall results; otherwise it may not translate to real trading.
  
  

(3) Are stop-loss rules ambiguous?

• If you lax stop-loss in hindsight by saying “if I held a bit longer it would have recovered,” you may suffer big losses in reality.
• During verification, adhere to strict rules and fully implement “if price hits the specified line, stop-loss/take-profit.”
  
  

4. Differences between backtesting and real trading

1. Real-time order slippage and spread fluctuations

   • Past charts assume constant spreads, but in reality liquidity shortages can widen spreads at moments.
   • Especially around major news releases, orders may not fill as per backtests.
     
     

2. Mental elements

   • Backtesting uses “virtual funds,” allowing calm decision-making, but real trading involves real money.
   • A string of losses may cause you to break rules; backtests struggle to reproduce such psychology.
     
     

3. Market environment changes

   • Backtests reflect past market conditions, but the future may not move in the same way.
   • Therefore, it is important to also conduct forward testing (demo account or small real trading) in addition.
     
     

5. Tips for applying backtest results to discretionary trading

1. PDCA cycle: adjust rules → re-test → implement

   • Example: stop-loss width was 20 pips but was being chopped off frequently → adjust to 30 pips and re-test, which improved win rate.
   • Be careful not to over-optimize. Check the broad rules, not overfit.
     
     

2. Separate good parts from bad parts

   • In which market environments (trending/ranging) or times of day was win rate higher?
   • What were the common points during losing streaks? (immediately after major indicators, low volatility periods, etc.)
   • In real trading, avoid similar situations or adjust position size accordingly as needed.
     
     

3. Compare your own intuition with the verification data

   • If your discretionary sense that “candlestick patterns tend to indicate reversals” is supported by actual data, it boosts confidence.
   • Conversely, discard beliefs not backed by data and choose strategies that are more likely to win.
     
     

6. Summary & next week's preview

Summary

1. Discretionary traders also need backtesting: confirming how well their rules performed in past markets is a major benefit for refining tactics.
   
   
2. Basic steps of manual backtesting:
   • (1) Set period → (2) Clarify rules → (3) Progress bar-by-bar and record results → (4) Aggregate and analyze.
     
     
3. Cautions:
   • To avoid bias from “seeing ahead,” use replay features.
   • Don’t feel secure based only on a string of wins.
   • Understand real-trade differences in stop-loss and spread fluctuations.
     
     
4. How to apply:
   • Clearly define failure and success patterns and reflect them in real discretionary decisions and position sizing.
   • Avoid excessive optimization and consistently operate with the assumption that “the future will be somewhat different,” iterating practice and fine-tuning.
     
     

Next (DAY 25) topic: Hybrid Trading Part 1 – Entry is automated, exit is discretionary

• By leveraging backtesting to solidify discretionary rules, it becomes easier to run an operation that is partly automated, i.e., a “hybrid trading” approach.
  
  
• Tomorrow we will detail the merits, drawbacks, and practical examples of a method where entry is automated while exit is discretionary.
  
  
• This should be useful for busy people or those who want to eliminate hesitation in entry timing, so please look forward to it.

If you are interested in automated trading, please also check the link below.
https://www.gogojungle.co.jp/users/147322/products

If this was helpful, we would appreciate it if you click “Read more.”
Thank you.