Some troubles about the difference of backtest results between SQ and MC

       Probably most sq users use MT4/% for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)
3.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits in their respective annual cycle reports. The            main reason is that MC will calculate the Openpositionprofit in each year period for trade that span two different years. It is not according to the entry time or exit time of the trade.
     b.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report sq4_9660.

     c. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     d.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.

     e. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928

To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.

       Probably most sq users use MT4/5 for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)
3.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits in their respective annual cycle reports. The            main reason is that MC will calculate the Openpositionprofit in each year period for trade that span two different years. It is not according to the entry time or exit time of the trade.
     b.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report sq4_9660.
     c. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     d.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.
     e. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928



To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.

       Probably most sq users use MT4/5 for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)
3.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits and monthly profit in period reports. Many strategies that have a long average holding time will show a profit of 0 for many months.The  main reason is that MC calculate the Openpositionprofit in each year,month,day for trade. SQ calculate simply according to the entry time or exit time of the trade.   Some statistical metric/column related to cyclical profits may gernareate bigger bias in SQX, such as Sharp ratio, Sortino ratio. So you will find that even though the default value of the risk-free return in SQ is 5%, but Sharp ratio in SQ is often much higher than MC's value even risk-free return be set 2%.

     b.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report sq4_9660.
     c. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     d.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.
     e. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928

To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.

       Probably most sq users use MT4/5 for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)

3. metric calculation bug, for example Calmar ratio which can be referred to the task  https://roadmap.strategyquant.com/tasks/sq4_010020

4.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits and monthly profit in period reports. Many strategies that have a long average holding time will show a profit of 0 for many months.The  main reason is that MC calculate the Openpositionprofit in each year,month,day for trade. SQ calculate simply according to the entry time or exit time of the trade.   Some statistical metric/column related to cyclical profits may gernareate bigger bias in SQX, such as Sharp ratio, Sortino ratio. So you will find that even though the default value of the risk-free return in SQ is 5%, but Sharp ratio in SQ is often much higher than MC's value even risk-free return be set 2%.
     b.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report sq4_9660.
     c. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     d.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.
     e. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928

To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.

       Probably most sq users use MT4/5 for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)
3. metric calculation bug, for example Calmar ratio which can be referred to the task  https://roadmap.strategyquant.com/tasks/sq4_010020
4.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits and monthly profit in period reports. Many strategies that have a long average holding time will show a profit of 0 for many months.The  main reason is that MC calculate the Openpositionprofit in each year,month,day for trade. SQ calculate simply according to the entry time or exit time of the trade.   Some statistical metric/column related to cyclical profits may gernareate bigger bias in SQX, such as Sharp ratio, Sortino ratio. So you will find that even though the default value of the risk-free return in SQ is 5%, but Sharp ratio in SQ is often much higher than MC's value even risk-free return be set 2%.
     b.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report sq4_9660.
     c. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     d.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.
     e. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928

     f. differences in the treatment of max refernce bars. As you can see in task: Differences in the treatment of max reference bars between MC and SQ - Strategy Quant 

To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.

       Probably most sq users use MT4/5 for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)
3. metric calculation bug, for example Calmar ratio which can be referred to the task  https://roadmap.strategyquant.com/tasks/sq4_010020
4.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits and monthly profit in period reports. Many strategies that have a long average holding time will show a profit of 0 for many months.The  main reason is that MC calculate the Openpositionprofit in each year,month,day for trade. SQ calculate simply according to the entry time or exit time of the trade.   Some statistical metric/column related to cyclical profits may gernareate bigger bias in SQX, such as Sharp ratio, Sortino ratio. So you will find that even though the default value of the risk-free return in SQ is 5%, but Sharp ratio in SQ is often much higher than MC's value even risk-free return be set 2%.
     b.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report sq4_9660.
     c. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     d.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.
     e. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928
     f. differences in the treatment of max refernce bars. As you can see in task:

          Differences in the treatment of max reference bars between MC and SQ - Strategy Quant 

          Second bug about max reference bar - Strategy Quant

To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.

       Probably most sq users use MT4/5 for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)
3. metric calculation bug, for example Calmar ratio which can be referred to the task  https://roadmap.strategyquant.com/tasks/sq4_010020
4.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits and monthly profit in period reports. Many strategies that have a long average holding time will show a profit of 0 for many months.The  main reason is that MC calculate the Openpositionprofit in each year,month,day for trade. SQ calculate simply according to the entry time or exit time of the trade.   Some statistical metric/column related to cyclical profits may gernareate bigger bias in SQX, such as Sharp ratio, Sortino ratio. So you will find that even though the default value of the risk-free return in SQ is 5%, but Sharp ratio in SQ is often much higher than MC's value even risk-free return be set 2%. This way of calculation will lead to a larger error when calculating Portfolio Sharpe/Sortino ratio.

    b. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928
     c.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report sq4_9660.
     d. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     e.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.
    
     f. differences in the treatment of max refernce bars. As you can see in task:
          Differences in the treatment of max reference bars between MC and SQ - Strategy Quant 
          Second bug about max reference bar - Strategy Quant

To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.

       Probably most sq users use MT4/5 for trading. But i'm Multicharts user, because it is a legal system trading software in both the US market and emerging markets for trading futures and stocks. As we all know, the strategy developed in sq must be finally tested and verified under the trading platform. Unfortunately, The backtest reports of the two platforms are usually more or less different, sometimes very big.Include net profit, total number of trades, and profit distribution per year. It has to take me quite a bit of time to do this, especially when i am building a big portfolio. The differences due to different settings are not within the scope of this discussion. I think there are several reasons:
1.Some strategies trigger a bug in SQ(For example sq4_9829, recently I discovered that SQ generate false source code once using MA indicator).
2. Opening times have been adjusted historically in some emerging markets .For example, the opening time is adjusted from 9:15 to 9:30.This makes a carriage return difference to the M30's strategy.The default cutting bar method in MC is 9:15-9:45,9:45-10:15,and so on.(Of course, this can also be understood as session setting is different.)
3. metric calculation bug, for example Calmar ratio which can be referred to the task  https://roadmap.strategyquant.com/tasks/sq4_010020
4.Three computational differences the SQ engine and the MC engine.   
     a.Comparing the backtest reports of sq and MC, even though Net profit and the total number of trades are exactly the same, we will find different annual profits and monthly profit in period reports. Many strategies that have a long average holding time will show a profit of 0 for many months.The  main reason is that MC calculate the Openpositionprofit in each year,month,day for trade. SQ calculate simply according to the entry time or exit time of the trade.   Some statistical metric/column related to cyclical profits may gernareate bigger bias in SQX, such as Sharp ratio, Sortino ratio. So you will find that even though the default value of the risk-free return in SQ is 5%, but Sharp ratio in SQ is often much higher than MC's value even risk-free return be set 2%. This way of calculation will lead to a larger error when calculating Portfolio Sharpe/Sortino ratio.
    b. Huge portfolio correlation difference bettween SQ and MC, which can be reffered to the task https://roadmap.strategyquant.com/tasks/sq4_9928
     c.The logic of slippage calculation is inconsistent between the two platform, which can be referred to the report Error of slippage calculation between Multicharts and SQX - Strategy Quant.
     d. Slippage calculation bug when using the money management method, which can be referred to the report https://roadmap.strategyquant.com/tasks/sq4_010008
     e.Differences in the calculation of commision. Although error in this reason is the smallest of several factors. For many futures and stocks in emerging markets,  commission is usually calculated persentage          based by entry amount and exit amont. So SQ should be add percentage based per side option.
    
     f. differences in the treatment of max refernce bars. As you can see in task:
          Differences in the treatment of max reference bars between MC and SQ - Strategy Quant 
          Second bug about max reference bar - Strategy Quant

To sum up, I think it is very necessary to optimize these three known calculation differences in order to greatly save the time of the last step, the comparison and verification of the backtest reports of the two platforms. Once optimized, finding other bugs or problems can be very efficient.