I have been comparing live trades to sim trades to what the back testing shows. I have concluded that Sim account trades are very very close to my live account trades. I run my strategies on a sim account and the live account at the same time so I can see if there is any slippage. I have found this to be minimal with my setup. How ever when any and every trade is calculated in a back test 50% of the trades have the minimum slippage it seems. This is why.
Take a basic cross over strategy that fills a order on the live account and on the sim account at lets say $100.00. This moment is time stamped. If you go to back test at this very moment all 3 prices should be all the same. BUT lets say the market goes up in the next 10 minutes, and now your moving average has changed. the "crossover" point where the initial trade triggered has now moved. If you go and back test at this time it will show a different time stamp and price for what should be the exact same trade.
Why does the back testing software apply calculations based on what I would call the "whole finalized picture" verses calculating the trades based on bar after bar.
Think of the data being tested as a wave of information. Why does the analyzer calculate EVERYTHING based off of a static finalized wave. NOT on a bar to bar basis.
Yes you can I guess add in the slippage value yourself IF you know the value. But this to me is a huge drawback and I don't see why it can't be integrated into the software a proper way. Yes it may take a considerable amount of time to actually compute the data but it will be realistic results. Just like the default optimizer and the genetic optimizer options.
In other words trades in the back testing should be exactly the same as if they were done with a sim account.
Take a basic cross over strategy that fills a order on the live account and on the sim account at lets say $100.00. This moment is time stamped. If you go to back test at this very moment all 3 prices should be all the same. BUT lets say the market goes up in the next 10 minutes, and now your moving average has changed. the "crossover" point where the initial trade triggered has now moved. If you go and back test at this time it will show a different time stamp and price for what should be the exact same trade.
Why does the back testing software apply calculations based on what I would call the "whole finalized picture" verses calculating the trades based on bar after bar.
Think of the data being tested as a wave of information. Why does the analyzer calculate EVERYTHING based off of a static finalized wave. NOT on a bar to bar basis.
Yes you can I guess add in the slippage value yourself IF you know the value. But this to me is a huge drawback and I don't see why it can't be integrated into the software a proper way. Yes it may take a considerable amount of time to actually compute the data but it will be realistic results. Just like the default optimizer and the genetic optimizer options.
In other words trades in the back testing should be exactly the same as if they were done with a sim account.
Comment