this is my first post here and it is driven by a very serious issue I have with the cowabunga system.I hope that someone can point us in a new direction.
Me and a friend are trying to create an EA based on the cowabunga system. We were not using the ones already available in this forum because they tend to give completely different results than the original system from the blog and we understand best what we develop ourselves.
The issue we are currently working on is the win rate. We did a lot of backtesting, examining the performance of the system since 01/01/2000. We always got bigger average win than average lose (about 1.2 to 1.5 times), as the guy in the blog did. Unfertunetly, our win rate never went above or even approached 50%. We usually got an overall win rate of 45-46% (Blog: 57%). This leads to massive drawdowns (more than 50% during the first half of 2011) and makes the system unprofitable for the entire first half of the first decade (2000-2006). In fact, we were never even close to break even during this period. Without changing anything, the system provided more or less consistent profits after 2011. These results were obtained trading EURUSD on 15 min charts. When we are testing GBPUSD, the results are even worse.
I would like to ask you guys and maybe even Pipsurfer for some suggestions on how to improve the win rate above 50% since this seems to be the major factor for the massive loses we are currently experiencing. We sticked to the rules as close as possible, as listed below, and now we are running out of ideas :(.
Our entry and exit rules (for a long trade) are listed below.
I do not think that its a money management issue since we are not risiking more than 3% per trade with average loses of less than 2%.
EURUSD (10 am - 10 pm GMT):
on 240 min: EMA 5 > 10
on 15 min:
EMA 5 > 10,
MACD negative and increasing or switched from negative to positive within the last
2 candles.
RSI (9) > 50
Stochastics (10,3,3 on High/Low): Signal line increasing value and still below 80.
Exit: First target is Bid +/- stop loss, if we have a clean break, we adjust the target in steps of 5 pips. We DO NOT use .00 and .50 levels because we havent figured out an efficient way to implent that in the code yet.
Your system is quite a bit different from the vanilla, so I’m not surprised you’re getting different results.
A small side note, why do you disregard stochastic and RSI values on the 4 hour chart?
I also tested it with an EA using the cowabunga indicator The result were so bat that I didn’t continue with it. I found the indicator was also not correct and made some changes to it.
Making an EA with MA crossovers are very difficult to make profitable. The human eye sees the difference between a false signal and a good signal The cowabunga is not trading all signal he get from his system and that can make a big difference I think. The same is with the Ribbon system but I want to give that an other try. I think the problem is in flat market that gives to many false signals Toekan found a solution for it in an old thread. 301 Moved Permanently
What is the problem implementing 0.00 and 0.50 levels?
How did you test this system? How reliable is your data from 2011? How do you work with different time frames in your EA? Are they causing problems? I recently have a Multi time-frame indicator that gives different values to the EA then appear on the chart and the EA is making trades that it shouldn’t