Can anyone test the following strategy ??

And this is where statistical analysis will help in determining stop loss, trade times and the expected move.

To add to this I may look to trade just once in a day if stop hit I sit out. So will not get whipsawed whole day.

I want to know statistics of how many days can we get a move > 4-6 pts in a given direction with minimum reversal in between.

I am ok if it goes UP 3 pts comes down 2 or 3 pts and goes back UP 5 or 6 pts.

Once entered I will not exit until my stop hit and that will be end of my trading that day.
 
I think if you want someone to backtest this, you will need to be much more specific. You have specified the entry, in that you enter long if it initially goes up X points or short if it initially goes down by X points. But, where are the initial stops? Are there any? Are they fixed or trailing? What is the initial target? You can't just say the target is somewhere between 2 and 4 times the distance between the open and the X point entry. You must pick one and stick to it (until you get to the optimisation phase of the tests anyway). Etc. Etc. Etc.
 
Quote from abhay:

I want to know statistics of how many days can we get a move > 4-6 pts in a given direction with minimum reversal in between.
More...

Also, the term "minimum reversal" is rather vague as well. What are you defining as a minumum reversal. I assume this is implying that you want to use a trailing stop? Does the system reverse on stop, or just quite for the day?
 
don't have time to give you details, i'll just say don't be dissuaded by the automatic naysayers, get yourself some historical data, and test this out. alternately, use a minimal amount of capital to forward test it or even paper trade it.

i guarantee you, you will learn something. :-)

Quote from abhay:

So say if S&P closed today @996.

Now tomorrow if S&P goes UP then it will go atleast to 1001 - 1002 and if Market goes down then it will go atleast to 990 given that market moves atleast 6 points in a direction.

So if we blindly place buy order @998 to be sold @1000 or 1001
or on other side blindly place an order to sell @994 (if market moves down) tobe covered @992 or 991 or 990
More...
 
Yes get all the days you want and us a "if" exel formula and get the percentiles laid out. For instance, when through you may have 10% over x points, 11% over x+1 points, up to 100% over x+100 points. Then get the increments.

John
 
Actually what you are trying to do is a well known strategy. The pitfall you have to avoid is that what happens to so many traders is they test 3 years of data, get a system, have a few losers in a row and quit that system. The system you describe has to have all trades taken for it to test out.

John
 
Stroma,

I have given the target points as 2pt or 3 pts or 4 pts.

I believe once you have written a strategy or formula then it is a minimal task to change the values and compare the results.

I will be happy even if I can get 2 pts/contract every day, but if probability of 3pts also looks good on historical then I will take that.

As far as Stop losses are concerned, frommy experience strategy reports tell you what would be drawdowns and how many times based on that I can fix my stoploss and test again.


It is all a question of writing a strategy or formula ot test this, after that there are os many scenarios I can play with and see what will work best.

My conviction is something good can be made out of this...
 
Quote from abhay:

Hi,

Just throwing an idea to test.

I have observed that when the Market moves in any direction it goes to atleast certain distance, building on that if anyone can test the following :

Say Market closed @ X previous day

then if Market moves to X + ( 2 or 3) then it is very likely to go to X + (4 or 5 or 6)

Similarly if market moves to X -2 then it is very likely to go to X - (4 or 5 or 6)

I would like to see %winners, this number 4,5 or 6 can be optimized later.
More...
Suppose that someone has already tested and optimized this strategy, and is currently trading it successfully. Do you think he would reply to your post? This is a rhetorical question only and should be disassociated with what follows.

As others have mentioned, there are still parameters which you have not defined. As it happens, I have looked into your idea or, at the very least, a variation thereof. All I'm prepared to say is that it is worth investigating. John (jficquette) makes a valid point when he talks about the discipline of sticking to the system after a series of losses. My self-written testing software would probably handle your strategy (once fully defined) with only small modifications. I may be inclined to test something specific for you (on ES tick data) if it doesn't require too much time and fits into my schedule - this of course precludes any optimizing.

For obvious reasons, I have been intentionally vague. However, the offer of assistance, with its qualifications, is genuine. I would prefer to keep all future communication, if any, here publicly on ET (no PM's or emails).
 
I thuoght there are some good folks who can write good strategies and if they have not tried this idea may be interested in doing some research on this, like I have seen on some other threads.

To begin with my specifications will be something like :(1st scenario)

1. Say Prev closing price is S&P 995

2. Next morning if price goes to 997.5 then BUY and sell @1000 or @999.5 with STOP @992.5 (this figure can be played around with)

or if price goes to 992.5 then SELL which is to be covered @990.5 or 990 with a STOP @996.5

Whichever target hits first trade is completed on that and quit for the day.


Summary: if price rises +2 or +2.5 pts from prev close then BUY and Sell at Prev close + 4 or 4.5 pts with STOP = Prev close - 2 pts

If price drops by -2 or -2.5 pts then SELL at Prev close - 2 or 2.5 and cover at Prev close - 4 or -4.5 with STOP = Prev close + 2

Just one trade and quit after 1st trade completed.
 
You must log in or register to reply here.
Back
Top