Here are the 2nd Test Settings for the CowabungaEA.
The EA is being developed from the forum version(s), other information and custom indicators, as well as, testing settings from the trade machine version.
If you are new to my posts and to what I’m trying to achieve with the Cowabunga EA’s open development please first view my post “Cowabunga EA addons” in the Expert Advisors and Automated Trading thread.
[B][U]Analysis of the Mechanical (Manual) Cowabunga System Trades[/U][/B].
With the second test we will try to establish a simple method for making the EA continuously profitable. Readers of my few posts and other forum posts will know that the Cowabunga manual system is quite profitable, but most attempts to maintain this within an EA have not been successful.
Thanks to Pip Surfer’s blog posts including his trading histories we have hopefully established some realistic and more sensible settings for MinTP, MinStop and MaxStop.
Here are some interesting stats that Pip Surfer himself might not be fully aware of As of 01/21/2010
Total Profitable Trades: 302 (+11096 pips) Avg. TP: 36.74 pips
Total Loosing Trades: 194 (-5204 pips) Avg. SL: -26.82 pips
So folks that’s a total of +5892 pips since Jan. 2007 (i.e. the system works!)
Using these statistics we have chosen the following new settings to try and make the currently available Cowabunga EA perform as close to the manual system as possible.
[B]MinTP 20[/B] (10 default) - Pip Surfer’s avg. TP is almost 40pips ! However, it’s unrealistic to use this as the MinTP of course. The value MinTP 20 may [B]not[/B] be optimal. However, it’s much better to take the MinTP at 196 total trades above 20 pips than it is to take 283 total trades at above 10 pips! i.e.
196 x 20 pips = +3920 pips
283 x 10 pips = +2830 pips
Take Profit is still reachable, but below the avg. threshold and this should maintain a much more reasonable profit to risk ratio. N.B Only 15 manual trades were below the 10 pips TP which is of course negligible overall and also not allowed with some brokers.
[B]MinStop[/B] 15 (15 default) - This is probably stlll a reasonable value to test.
[B]MaxStop[/B] 30 (50 default) - Pip Surfer’s manual trades have an avg. loss of around -30 pips. Again, this should be a much more acceptable loss to risk ratio. Only 69 out of 194 loosing and ‘stopped out’ trades are more than -30 pips. If this 35%+ of the more than -30 pip loosing trades can be ‘capped’ at this level without affecting the swing high/low then this will greatly increase the EA’s total profit in pips.
[B]TPMustBeGreaterOrEqualToStop[/B] false (default false) - This was previously set to true to test a wider spread (50 / 50 ratio in the 1st test), where some lots and initial trades are to close to the 00 or 50 levels (as per Pip Surfer’s recommendations and within the EA settings of course). The relatively large stop loss and therefore TP was both beneficial and problematic here! Given the above statistics it’s much easier to understand where the big high/low - profit/loss swings stem from.
This 2nd test will be much closer to PipSurfer’s recommendations and trading times and will hopefully more closely ‘mirror’ his future trades and profits. We will be adding the ‘AvoidNews’ features in this 2nd test from Pip Surfer’s postings.
We expect to see less overall profitable trades than Pip Surfer. However, these will hopefully be offset with our stop losses. N.B. this maybe close to the optimal performance for the EA, before the implementation of it’s new features and settings with the Neural Network additions. Valid entry points also need much more work, because their is no ‘automatic’ exit once a trade has begun to swing in the opposite direction. These are ‘human’ factors which are difficult to determine. This will be the main focus of a Neural Network style addition to the EA.
[B][U]CowabungaEA 2nd Test Settings for GBP/USD only[/U][/B].
[B]StartTimeGMT[/B] 08:00 (default 04:00)
[B]EndTimeGMT[/B] 22:00 (default 20:00)
[B]AutoCloseOrdersAtEndTimeGMT[/B] true (default true)
[B]LotSize[/B] 0.1 (default) 0.01 for mini-account
[B]MagicNumber[/B] (user defined)
[B]MinTP[/B] 20 (10 default)
[B]MinStop[/B] 15 (15 default)
[B]MaxStop[/B] 30 (50 default)
[B]TradeIfStopIsGreaterThanMaxStop[/B] true (default true)
[B]TPMustBeGreaterOrEqualToStop[/B] false (default false)
[B]EMA4HrFastPeriod[/B] 5 (default)
[B]EMA4HrSlowPeriod[/B] 10 (default)
[B]EMA15MFastPeriod[/B] 5 (default)
[B]EMA15MSlowPeriod[/B] 10 (default)
Edit: Live MT4stats are no longer available.
N.B. A test for GBP/JPY will also follow. However, the ‘optimal’ setting will be somewhat different. Moreover, we think that the GBP/JPY could be potentially more profitable using the Cowabunga EA than it might be with the GBP/USD !
Active Currency Pairs - Avg. Pip Range
London Session (3am to 12pm ET) 8am to 5pm GMT (8 hrs)
- GBP/JPY - average range 140 pips
- GBP/USD - average range 110 pips
New York Session (9am to 5pm ET) 2pm to 10pm GMT (9 hrs)
- GBP/JPY - average range 120 pips
- GBP/USD - average range 90 pips
Tokyo Session (7pm to 4am ET) 12am to 9am (9 hrs) crossover - depends on BST etc.
- GBP/JPY - average range 110 pips / 4.5 (approx)
- GBP/USD - average range 60 pips / 4.5 (approx)
Here’s to avg. pip ranges and volatility
Happy Trading and GL !
Thanks.
XeroFX.