Playing for profit
- Thread starter stock72
- Start date
- Status
Stock72,
I am back again (to warn
..
Remember long time back (probably last year) you were asking about your trading method based on backtesting the historical data. While some posted a, link to a thread on an excerpt from a book on fallacy of backtesting, I had hinted that the longer history you test it relative to the projection period, the more reliable it will be. Remember I also read that book. But my suggestion was from a different perspective (personal learning). I was reading the book again and realized that that my hint about timeframe ratio increasing reliability of backtesting, is valid as per the same book. This comes just few pages earlier on the sample size of initial cohorts when certifying track records.
Now what this might mean is that the possibility of a winning streak by your system which is to happen rarely has already happened. Next onwards the streaks will be shorter. Hence be very alert about cutting your finger. Or it may be time to retune your algo.
I am back again (to warn
..Remember long time back (probably last year) you were asking about your trading method based on backtesting the historical data. While some posted a, link to a thread on an excerpt from a book on fallacy of backtesting, I had hinted that the longer history you test it relative to the projection period, the more reliable it will be. Remember I also read that book. But my suggestion was from a different perspective (personal learning). I was reading the book again and realized that that my hint about timeframe ratio increasing reliability of backtesting, is valid as per the same book. This comes just few pages earlier on the sample size of initial cohorts when certifying track records.
Now what this might mean is that the possibility of a winning streak by your system which is to happen rarely has already happened. Next onwards the streaks will be shorter. Hence be very alert about cutting your finger. Or it may be time to retune your algo.
Ha!
what a Prophecy!!!!!!!
Not a single word more or less.. u just told what I discussed with my friends here yesterday.. and in fact 90% completed in tweak the system which is going to give a very good improvement ...
your words pushed me to think very weirdly .. you may be one among the friends to whom i discussed here????
( Benjamin ..if so pls disclose sir .. i cant bear this much suspense )
I owed a lot to you. Thank you.:clapping:
what a Prophecy!!!!!!!
Not a single word more or less.. u just told what I discussed with my friends here yesterday.. and in fact 90% completed in tweak the system which is going to give a very good improvement ...
your words pushed me to think very weirdly .. you may be one among the friends to whom i discussed here????
( Benjamin ..if so pls disclose sir .. i cant bear this much suspense )I owed a lot to you. Thank you.:clapping:
Stock72,
I am back again (to warn..
Remember long time back (probably last year) you were asking about your trading method based on backtesting the historical data. While some posted a, link to a thread on an excerpt from a book on fallacy of backtesting, I had hinted that the longer history you test it relative to the projection period, the more reliable it will be. Remember I also read that book. But my suggestion was from a different perspective (personal learning). I was reading the book again and realized that that my hint about timeframe ratio increasing reliability of backtesting, is valid as per the same book. This comes just few pages earlier on the sample size of initial cohorts when certifying track records.
Now what this might mean is that the possibility of a winning streak by your system which is to happen rarely has already happened. Next onwards the streaks will be shorter. Hence be very alert about cutting your finger. Or it may be time to retune your algo.
I am back again (to warn
..Remember long time back (probably last year) you were asking about your trading method based on backtesting the historical data. While some posted a, link to a thread on an excerpt from a book on fallacy of backtesting, I had hinted that the longer history you test it relative to the projection period, the more reliable it will be. Remember I also read that book. But my suggestion was from a different perspective (personal learning). I was reading the book again and realized that that my hint about timeframe ratio increasing reliability of backtesting, is valid as per the same book. This comes just few pages earlier on the sample size of initial cohorts when certifying track records.
Now what this might mean is that the possibility of a winning streak by your system which is to happen rarely has already happened. Next onwards the streaks will be shorter. Hence be very alert about cutting your finger. Or it may be time to retune your algo.
- Status
Similar threads
-
-
Required AFL Codes for displaying "Today's Bar Numbers" on Each of Candles on Charts..
- Started by pkgmtnl
- Replies: 11
-
Countries smaller than Gurgaon are playing the World Cup
- Started by Sujan Azad Parikh
- Replies: 5
-
MT4- Strange!!Displaying Charts even without market open.
- Started by tradingstudent
- Replies: 4
-
