Hello all,
I'm new to creating strategies and am trying to wrap my head around some of the technicalities of testing. My initial goal is to teach myself the basics..... so in an attempt to accomplish this I have created a few strategies using different indicators, chart types and time periods.
The one in question is a strategy that I have applied to a Line Break chart which creates some basic buy and sell signals. I have set this up using 4 contracts for no other reason than just randomness. I should also say that my strategy uses an on bar close method so is set accordingly. When I go to the strategy and bring up the performance report for historical/real time I find that the strategy has produced a small profit. My first thought was O.K. that's great but how accurate is this method? I then did some reading on back testing and followed the help guide instructions to produce a back test. Surprisingly, the back test results showed a slight improvement over what the performance report had predicted. My next though was.... lets increase the time period so I downloaded (IQfeed) 360 days of historical and ran the test again finding the same results. Not exactly linear as the last year has had some good ups and downs but consistent. I then picked several days which produced multiple entries and exits both profitable and at a loss and downloaded the historical data to perform Market Replay tests. These produced a slight increase above and beyond the other two methods.
So..... how accurate will these tests hold true? This is just a time thing for me as the time required to connect my account to my existing IB account (simulated for these purposes, new to coding my own strategies but not trading) is just to time consuming to test every small iteration.
I'm questioning this because in researching this dilemma I have found some contradicting information as to the accuracy of back testing when using the Line Break chart type. This is persistent on other forums that came up while searching as well. One company even claims to have invented a new type of Line Break chart that eliminates any errors. So are these errors if they exist due to intra bar type strategies and not applicable to On Bar Close strategies?
I'm sorry for the long winded post but certainly appreciate any help/advise that anyone may have to offer. Now if anyone would like to talk about the finer points of trading options I can help! Don't......... futures just seem way simpler, I'll let you know how it goes.
GuppyDRV
I'm new to creating strategies and am trying to wrap my head around some of the technicalities of testing. My initial goal is to teach myself the basics..... so in an attempt to accomplish this I have created a few strategies using different indicators, chart types and time periods.
The one in question is a strategy that I have applied to a Line Break chart which creates some basic buy and sell signals. I have set this up using 4 contracts for no other reason than just randomness. I should also say that my strategy uses an on bar close method so is set accordingly. When I go to the strategy and bring up the performance report for historical/real time I find that the strategy has produced a small profit. My first thought was O.K. that's great but how accurate is this method? I then did some reading on back testing and followed the help guide instructions to produce a back test. Surprisingly, the back test results showed a slight improvement over what the performance report had predicted. My next though was.... lets increase the time period so I downloaded (IQfeed) 360 days of historical and ran the test again finding the same results. Not exactly linear as the last year has had some good ups and downs but consistent. I then picked several days which produced multiple entries and exits both profitable and at a loss and downloaded the historical data to perform Market Replay tests. These produced a slight increase above and beyond the other two methods.
So..... how accurate will these tests hold true? This is just a time thing for me as the time required to connect my account to my existing IB account (simulated for these purposes, new to coding my own strategies but not trading) is just to time consuming to test every small iteration.
I'm questioning this because in researching this dilemma I have found some contradicting information as to the accuracy of back testing when using the Line Break chart type. This is persistent on other forums that came up while searching as well. One company even claims to have invented a new type of Line Break chart that eliminates any errors. So are these errors if they exist due to intra bar type strategies and not applicable to On Bar Close strategies?
I'm sorry for the long winded post but certainly appreciate any help/advise that anyone may have to offer. Now if anyone would like to talk about the finer points of trading options I can help! Don't......... futures just seem way simpler, I'll let you know how it goes.
GuppyDRV
Comment