Is trend your friend?

[oxusmorouz]

Ajay,
Notional & Actual ,Investor & Trader are all very clearly known by you.Noise constitutes even Arbitraggers ,the Safest Return path ,yielding over 12% per annum minimum ,is even better than 7.8% ,as shown by you.This you know best than me.
Trend or Counter Trend is the Best,it all depends how we identify.My personal experience says it is not in Entry/Exit blah blah but sensing the Yellow light early before the Red is most Imp when on Trade.Everyday if one looks at the screen relaxedly ,putting actual money days , months & years together develops an uncanny sense of the forth coming tempest.
Ask any experienced 'Dealer' ,where though he may not have put his real money,but client's pally to him,you will get an input,which are very difficult to codify.
Asish

Noise trading, sir, is basing entry-exits on irrelevant information. Arbitrage when used in an absolute context is much different than when viewed as "relative asset price mismatching", which by no means is noise trading. In technical terms, identifying opportunities where mispricing exists between two assets and creating a zero beta portfolio which need not necessarily give an assured return, but is a function of probability. I've been working on a similar concept and I hope to use this in my trading, after a period of beta testing in real market conditions.

The 12.6% return on DJIA futures-Bond was an outcome of a compeltely passive investment strategy which requires very little time and devotion. The question is, how many of these "technical trend" traders who started trading in 1970 (forget 1970, let's talk about 1995) are still existing? The one's who have been fortunate enough to earn a positive return, though not necessarily on a risk adjusted basis, are the ones who write books on trend following, give seminars etc.

 

[beginner_av]

I cannot prove or disprove anything. My testing on the data from 1970 to 1.9.2005 gives healthy returns with an annual return of 23.02%. Just a 5,20 EMA system for entry. 2% initial SL and Trailing SL of 1% (risk).
The SL and TL are pretty bookish and standard that may not even work in today's markets. Needless to mention that nothing is optimized (canned as you say.)

 

[oxusmorouz]

Posting the results BAV. Brokerage is .5% both sides, initial stop is 2% and trailing stop is 1%. Orders delayed by 1 bar and executed on the maximum of close or open of the next bar for buy and short cover and min of open and close for shorts and buy exits.

 

[amitt29]

I assume that the test has been done on all the the systems(uy hold,5-20 whateva) that assumes trades to be done continuosly from 1800's to 2000 whatever.

 

[beginner_av]

http://img232.imageshack.us/img232/8079/djiarepsp1.png

Your test result is absolutely correct. There are fundamental differences between your testing and mine.

best wishes,
BAV

 

[oxusmorouz]

http://img232.imageshack.us/img232/8079/djiarepsp1.png

Can you post a point's test BAV? It's easier to compare.

 

[beginner_av]

Sorry Sir, never heard about such a thing. As I said your results are fine in its context. Anyway my method of testing will be completely unacceptable to you for comparison (from your point of view), so wont be of much use in comparison.

 

[beginner_av]

Here is a test from 1928 to 2007.
By using blind data like you have done where S/W doesnt know which came first H or L, I have tested the same with buy at Open next day, Sell at Close next day etc (buy, short at Open, delay 1),(Sell, Cover at Close, delay 1). All stops with a delay of 1, i.e. intraday you cannot get out. All reentry is only next day. (Impractical for an active trader, but harshest and most reasonable with blind data).
It gives a CAR of 14.53%. It improves to 16.19% if you buy/short at next days open but sell/cover at today's close.

Thats enough playing for me. Now back to work. Buy the way an active trader would be out of the market during non-trending periods which 1970s is. Why would anyone do that repeatedly throughout that period? So the crux is that if we imply wrong strategy at the wrong market, we will be chopped off. NO two ways about it. Building something like CV's template and then putting the whole thing to test would be more like a real one.

 

[uasish]

Abhijeet,
No the best is 1 day delay,becoz here Ajay is using min. overhead expense hence let us assume one is using only EoD data.

Asish

 

[beginner_av]

Well you may be right, but then....
Using the optimistic and pessimistic levels the results can be a CAR of anywhere between 61.48% to 14.53%, pretty useless test result, cos too many hypothetical situations (not Ajay's problem, problem is with data availability). BTW using all of Ajay's parameters, the results are encouraging and the system definitely qualifies for further evaluation and optimization. (This is again a completely seprate matter, not comparable with Ajay's initial intention of B&H vs TF).

BTW Oxy mine is not a point only system (if thats what you meant by Point's test), it is 100% A/c no leverage, position size = 1.