Forex CobraSystem

[The_Snowman]

From a FF post, searched this forum but no mention of "Cobra" in any thread so far

Seems purely mechanical strategy, only uses a few Empty4 indicators, so would seem to be straight forward coding, please check it out.

Thanks, John

http://www.forexfactory.com/showthread. ... ost5462250

[astral77]

Hi John

It seems to be a good solid system and in couple of days got a good following. Lets keep an eye on the thread and see how it goes. I am too busy with other commitments to test it myself but I think a S&R indicator such as squalo`s would make the system even safer. I strongly urge anyone reading this thread to have a look at the system and post their comments here.

Kind regards

[garyfritz]

Looks interesting. I knocked out a quick system in Tradestation to backtest it. I'm pretty sure I have the rules correct, or very close.

It's not a killer system, but it does work. It only wins about 25-35% of its trades (without the reversal-candle filter), but in most pairs the winners are big enough to come out ahead. However the trades are small. Many pairs made about 800-1000 pips (before spread/slippage) in the last year, trading about 1-2x per week. Often it was a very bumpy ride, but if you traded it on all 21 non-NZD pairs it does OK. Only AUDCAD, EURCAD, GBPCAD, GBPUSD, USDCAD, and USDJPY actually lost significant money in the last year. No idea why the CAD pairs had so much trouble.

The rules are vague about the "no reversal candle" filter. I tested it a bit and arbitrarily chose to block any entries when abs(Close-Open) < 50% * (High - Low). That filters out about 1/4 of the trades, tends to raise the profits & win% a bit, and increases the average win size. It's a big improvement.

I tried trading on Friday & not, closing on Friday & not, and I got the best results trading every day and holding over the weekend.

Trading all 21 non-NZD pairs, 0.1 lot per position, the profits look like this:

Cobra1.gif

If you knock out all the CAD pairs, it looks a bit better:

Cobra2.gif

... but there's no guarantee CAD will continue to be bad and other pairs will continue to work well. And even without CAD, you've got a 6-month flat spot in there. All its profits came between June and September last year.

I think it would need some improvement before it would be worth trading...

[Forex37]

Here is the llink to the previous EA thread. The latest (and working) version is hidden somewhere with a link given within this thread.

http://www.forexfactory.com/showthread.php?t=168347

Regards


Sorry, misunderstanding!!! The above link leads to an EA for Cobra`s THV system. Much more complicated system. I`ll leave the link anyway - maybe some interest could be spawned for Cobra`s system: http://www.forexfactory.com/showthread.php?t=127271

[astral77]

Weekend back tests are not that good since we have wild fluctuations in the criminals spread. If you follow the thread in FF you can see that they have found this out. I think they are realizing that an additional filter is needed and IMHO it is the S&R. As Nanningbob says for any good system you must have horizontal as well as diagonal lines.

Kind regards

[garyfritz]

astral, if you were referring to my backtests -- the weekend spread is irrelevant. I was testing on Tradestation and it just uses the bid. Which means it's not entirely accurate for systems that scalp tiny amounts but you can get pretty close.

[astral77]

astral, if you were referring to my backtests -- the weekend spread is irrelevant...

I was talking in general and not just regarding your back tests. I think we all agree demo testing is the best way to test without risking our hard earned cash.

By the way we have an EA in FF and it is being tested. If we come to the conclusion that it has legs then maybe we can ask Steve to give us a hand, but at present there is no need to impose on him.

There is some speculation that since the website which was selling this system is down it must have been some sort of scam. We know from other threads and forums that sometimes the webmaster for some reason such as bad health can not carry on and the website is shut down. So we should not prejudge and just make an informed decision on the performance of he system.

Kind Regards

[garyfritz]

I think we all agree demo testing is the best way to test without risking our hard earned cash.

Nope, I think live demo testing is a terrible way to test a system. It's just the best way available in Empty4 because Empty4's backtester is such utter garbage.

Live demo testing gives you a good idea for how the system behaves in a very small time window. Unless you intend to demo test it -- with all possible input variations &etc -- for a year or two, you'll have NO idea how the bot performs in various market conditions.

A reliable backtester lets you see instantly how the bot performs over a long history. You can check that it's trading properly, see how well and how consistently it makes money, see how it responds to different market conditions, tweak your system logic or your parameters to improve performance, etc. You can't do that with live demo testing unless you want to devote months or years to testing every EA.

[The_Snowman]

EA has been posted over on FF anyone following this topic