Simple ESMini Strategy?

[cd173]

Following on from my simple DAX strategy scenario, here is a system for the ESMini.

I actually pinched this one from Pristine's ESP software which signals a buy/sell after the break of the high/low of the of the first 60mins (2.30pm to 3.30pm). Target was set at 5pts with an initial stop of 3 pts, moved to breakeven after 3pt gain.

The attached results show the system turning a nice profit over the last 3-4 months of around 89 points (no comm/slippage included though). Only bad month being Sept with 8pt loss.

Not sure how this works out longer term, but any thoughts or improvements greatly appreciated!

Cheers
Chris

 

[MGBRoadster]

That's an interesting one, Chris,

I get the impression looking at recent data that using the 5 pt target misses quite a few big moves.

I've started paper trading this using a 3 pt trailing stop after the 5 pt target is hit, then closing at 21:00 if not stopped out. (wouldn't begin to know how to code this for backtesting!!!)

Thanks for sharing

 

[X-Ray Specs]

cd173 Nice simple setup.

Looks good but don't have reliable e-mini data to verify results.
Something I did notice is trading only Tues and Weds changes current W/L from 37W, 24L to 20W, 5L.

 

[JonnyT]

Taking into accout costs etc it is not viable IMHO.

JonnyT

 

[donaldduke]

Assuming this system works in the long run as it has over the last three months then it is perfectly
viable.

If the system nets 300 pts a year, you might lose max 100 pts in costs that still leaves a nice 200 pts profit.

The only question is does this system work this well in the long run...

 

[cd173]

Hello Chaps

The 5 point target and 3 point stop were added purely for assessment on a systematic level.
This system could no doubt be improved upon to generate more profit.
Baring in mind that I've only got 3 months worth of data to go on, long term suitability is the issue!

JonnyT - Interested to know why this may not be viable?

Cheers
C

 

[JonnyT]

It's worked out at about 1 point per trade.

Allow for Spread 0.25 pts, slippage, dealing costs and I reckon its lucky to have made 1 tick per trade net.

Much too thin for my liking. i.e. wheres the edge?

JonnyT

 

[donaldduke]

I think it shows that JonnyT trades FX and not ES.

ES is the most liquid futures market there is.

You wont get much slippage on ES with your stops (unless you are trading 100 lots), even when trading
breakouts.

If this system works in the long run then its completely viable.

 

[cd173]

JonnyT

Understand your point. So, when factored in a 1point slippage/cost implication per trade, you end up with a 4pt win and 3pt stop or 4pt win and 4pt stop?
I feel a better exit strategy is needed..............

C

 

[donaldduke]

JonnyT

Understand your point. So, when factored in a 1point slippage/cost implication per trade, you end up with a 4pt win and 3pt stop or 4pt win and 4pt stop?
I feel a better exit strategy is needed..............

C

No all you need to factor in is 5.25 profit exit.

Some times your 5 pt exit wont get filled, if the market touches your price and reverses
because there are orders ahead of yours, however (some times you will get filled too,
so 5.25 is a worst case exit).

By using 5.25 you garantee your profit order taking getting filled.

As far as your stop loss is concerned you might want to assume that 1 in every 2 exit stop
slips by 0.25 (ie 3.25 instead of ) 3.

Your system should work fine with ES.

 

[cd173]

Hi donaldduke

(Sorry just missed your post)!!!!

With reference to my previous post to you feel that a win loss of 5 to 3pt in reality would work, or that a 1point slippage cost factor is reasonable or excessive??

Cheers
C

 

[donaldduke]

Assume 0.25 slippage.

So you win 4.75 on the winners and lose 3.25 on the losers.

In reality with ES it shouldnt even be that much.

 

[cd173]

donaldduke

Thanks for the feedback!

C

 

[JonnyT]

ES is very liquid but don't be fooled about not getting slippage.

It very much depends on order type and how you apply them.

I'll test this strategy on ES using some out of sample data from this year and report back...

JonnyT

 

[donaldduke]

I did a quick test on this stragegy the results were not good on ES.
The system was actually made a loss for the period tested (July 2003- July 2004).

This may be due to the fact that most ES breakouts are false.

Heres a summary of my results:

1 means a win (either full 5pts or partial win at close)
0 means a loss (either 3 pt or partial loss at close)
- means break even trade
Days when no signal was generated are not shown:

10000-000-1010--10111--000000001001-001-0-01000--000001100100-1000000010100--110101-000001010-00100-00001100010-00100010000000-0111110010-0110000-0-0-01-0-0-011-000000-1010-11101000100--01100---0000--10-000--1-10-00101-000010100000-0-01011-010100

Only about 30% of the trades were profitable...

Ofcourse there could be a bug in my testing code or my ES 5 min data that i have could
be unreliable.

 

[cd173]

donaldduke

Very interesting. You didn't happen to test Aug - Oct 2004 just to clarify the results I posted?
Would be interested to see if we match up. :!:
Looking at the results, this strategy seems to work a couple of weeks at a time and then becomes unreliable. Too many 0000000 for my liking!!
Thanks for posting the backtest!

Profit factor needs to increase greatly to make it anywhere near worthwhile...

Cheers
C

 

[JonnyT]

My first results for the March 04 contract show a 132.25 points from 55 trades. An encouraging start.

I will now test on the more recent contracts.

JonnyT

 

[JonnyT]

The June contract shows 133 points from 63 trades.

I'm going to manually verify the results as they are rather exciting...

JonnyT

 

[JonnyT]

The results were too good to be true.

I have found a bug. Both periods were slightly negative...

However my tests suggest that a 3 point stop is too tight, and one should not profit take but let the market run.

Moving the stop to 5 with a BE at 3 and closing at 21:00 makes both contracts profitable.

It might be possible to enhance the entry criteria to wait for a pull back etc...

JonnyT

 

[cd173]

JonnyT

Thanks for the test results. The stops and targets were created off the top of my head and could no doubt be improved upon. By removing the target and letting the profits run until 21:00 would probably produce better results.

BTW there is also a Pristine strategy for a 3pm break instead of 3.30pm, but I believed at first glance it was less profitable......could be wrong though......

Cheers
C