DAY 30: Retrospective – Strengths and Limitations

On DAY 29, we learned the importance of keeping a trading notebook and tips for continuing. This week's theme (Week 5) is “thorough verification—and more verification.”

Today, DAY 30, we will delve into the merits and limits of backtesting (past verification).

For both discretionary traders and EA users, reviewing past market conditions—past verification—is a highly valuable exercise.

However, there are dangers if overtrusted, so let’s cover the points to use it in a balanced way.Balanced utilization points.

1. Merits of past verification

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

   • You can verify how well a rule you devised (e.g., the combination of moving averages and RSI) performed on past markets.
   • Before real trading, you can somewhat determine whether there are major defects.
     
     

2. Win patterns and loss patterns become clear

   • You can understand in which market conditions (trending, ranging, sharp drops, etc.) you tend to win or lose.
   • Based on this analysis, it’s easy to build a strategy that focuses on strengths and avoids weaknesses...
     
     

3. Verification speed is fast (especially for EAs)

   • With automated trading (EA), you can verify several years of data in a short time using MT4/MT5 Strategy Tester.
   • Even discretionary traders can, using chart replay features, learn a large number of market patterns in a short period.
     
     

4. Psychological reassurance

   • If you know, for example, “this method can withstand a drawdown of up to ○○ yen over the past 10 years,” you’re more mentally stable after a losing streak.
   • Helps you stick to trading rules.
     
     

2. Limits of past verification

1. Does not guarantee future market conditions

   • Past movements are not guaranteed to recur in the future.
   • The best method in the past ten years may not work in the next ten years. Environments change with monetary policy and world events.
     
     

2. Slippage, order rejection, etc., cannot be reproduced at indicator announcements

   • Past charts often assume constant spreads, and rapid changes or liquidity shortages in real time are hard to account for.
   • In particularinstant price jumps right after economic indicatorsare risks not visible in backtests.
     
     

3. Risk of over-optimization (overfitting)

   • If you optimize too much for past results, you may achieve an apparently high win rate with parameters tailored to a specific period.
   • In reality, the drawdown may become large in different environments.
     
     

4. Discretionary psychology and real-time speed cannot be reproduced

   • Backtesting shows results clearly, so there is no emotional fluctuation like in the real trading.Single misclick on manual stop loss or increasing lot size after a string of losses—these real-world mistakes cannot be reproduced.
   • 
     
     

3. What market conditions and how much should you verify?

1. At least one year of data

   • Even short-term traders should ideally look at a year of data to see seasonal trends and the influence of important indicators.
   • If possible, choose a period that includes both trending and ranging markets.
     
     

2. Cover multiple environments

   • Example: periods with a long uptrend, periods with sharp declines or crashes, periods with consecutive narrow ranges, etc.
   • Check how the rules perform under diverse environments.
     
     

3. Include an out-of-sample periodespecially for EA

   • Split data into in-sample for testing and out-of-sample for validation.
   • Verifying whether an over-optimized logic works in a different period (out-of-sample) helps reduce overfitting.
     
     

4. Steps to apply past verification to real trading

1. Decide that it is “generally effective” from backtests

   • Example: win rate 50%, risk-reward 1:2, expected value seems positive.
   • Understand guidelines such as acceptable maximum annual drawdown, etc.
     
     

2. Forward test (demo or small lot)

   • Track real-time movements with small funds or a demo account,experience slippage and order rejection.
   • Try for 1–3 months to see if you keep winning, and verify no logic collapse during drawdowns.
     
     

3. Full operation & regular checks

   • Invest a reasonable amount of capital.
   • Evaluate performance weekly or monthly,and adjust parameters or lot sizes according to market conditions.
   • Discretion traders should compare daily trading notes with validation data.
     
     

4. Beware of overfitting

   • If a string of losses occurs, reassess whether the market has changed or the logic has reached its limit, and consider pausing or adjusting rules.
   • Prepare for “string of losses” in advance (anticipate maximum drawdown and number of losses to keep your psychology in check).
     
     

5. Mindset to effectively use past verification

1. Don’t have excessive expectations

   • Thinking, “If backtests show 10% monthly return, real trading will be the same” is dangerous.
   • There is no guarantee that the past market and future market are the same.Treat it as a reference material.
     
     

2. Also focus on loss patterns

   • Don’t just review winning trades; thoroughly check cases that led to retracings or large losses.
   • Having mitigation strategies like “reduce lot size or stop if this pattern appears” lowers the risk of ruin.
     
     

3. Be aware of psychological differences

   • In real trading, fear of capital loss and pressure from consecutive losses create emotions that backtests do not have.Drawdown often feels emotionally tougher than theoretical numbers.
   • 
     
     

4. Continuous verification and updates

   • Markets are constantly changing. A logic that worked a year ago may not work now.
   • Weekly or monthly real performance should be validated and compared; adjust parameters or pause if needed. Flexibility is key to success.
     
     

6. Summary & next forecast

Summary

1. Merits of past verification: Check rule effectiveness in advance, clarify win/loss patterns, and establish a mental foundation.
   
   
2. Limitations: No guarantee of future markets, slippage cannot be reproduced, risk of over-fitting, and difficulty reproducing discretionary psychology.
   
   
3. To make it more reliable, forward testing is essential: Use demos or small lots to confirm suitability to real markets, then move to full operation.
   
   
4. How to utilize: Do not have excessive expectations; treat as “reference material,” regularly check and update performance while remaining wary of overfitting.
   
   

Next time (DAY 31) theme: Forward test – demo trading and small-lot operation

• After past verification, forward testing is essential.Forward testing.
  
  
• Tomorrow, we will look at specific methods such as “how to test in real-time environments” and“how long should you run demos or small lots?”
  
  
• This step helps experience risks like spread widening and order delays that were not visible in past verification, and finalizes the logic. Please look forward to it!

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

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