Phoenix - Development+Suggestions - MQ4 in Post#1 - page 30
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
In profit = worthwhile.
Updated settings = fewer losses. (4 days with these new settings, no losses yet)
New Settings - No losses??
What settings be these, Daraknor? Can you upload a modified EA or Settings file? Perhaps we should try them out on other brokers.
PContour's modification produces a constant string of messages about losses and Breakit 90 50 10. Can't tell whether the trailing stop is working though.
I'm not sure what I'm going to do with settings files. Our settings file thread seems dead, and you were the person who released the most aside from me. I released files to assist optimization, and nothing seemed to really happen because of it.
I'd like to release settings files, but I was hoping other people would release too. So far it's been a very small number of people. We had more guides released than settings files released. :/ I'll continue working on code for free, but I thought the community would be more involved regarding settings since it is something anyone can do.
Settings files
You mentioned that you have been loss-free for some time with new settings. These were the ones I was referring to.
The sum total result of hundreds of hours of backtesting and optimizing on my part has been a big goose egg. I wish I could contribute something, but I have drawn a blank. Optimize on a six month period, and the next month, two months, etc. is a total loss. There is no consistency. I presume others have the same experience. On top of that, Bert and Azbo keep telling me to go back to 5.7.0 and run default settings in mode three. So I guess that's as good as it gets.
This is why I started looking at the signals, and considering a trending system. I would like to at least understand how the signals function in order to be able to tweak them more effectively. If I can visualize a signal with an indicator in real time, perhaps I can make constructive adjustments to the inputs. On the other hand, intuitively, it makes sense to me to switch over to trending, since I think losses with countertrending are bound to be too large and trades open too long. EASY provides some interesting alternatives here, and it is all very visual (indicators already exist). I have tried to follow your guidelines for optimization, but often find my head spinning. Visual strikes me as a simpler approach.
I appreciate your time and effort, and that of all the others on the thread. Maybe we need to organize into committees so that each task gets the amount of attention it needs. Haphazard testing will not give you the feedback you need to push the program, even though some of the suggestions have been very positive. For example, Pcontour's tweak of the TS seems very good, but I need more information in order to put it into practice. I'm not sure what I did is actually working.
That said, my demo accounts are up close to 10% in just a couple of days running default settings on 5.7.0 and 5.7.2W (modified) in mode 3. So there's a silver lining somewhere... I'd like to find the gold lining though.
CCI Indicator
A friend of mine just brought the Channel Index Indicator (CCI) to my attention. I wonder if this is not the same bird as our Signal 3.
See http://www.forexheaven.com/woodies-cci.html
Not at all. Signal 3 is a momentum derivative. A number that reflects the rate of change in the rate of change. CCI is a longer term subjective comparison between moving average and price. the difference is like pressing the gas or the break, compared to how close you are getting to the crosswalk. Your current position would be price, and the crosswalk would be the moving average. Some people think that the moving averages indicate a future price, but I don't think there is strong evidence to support this stastistically for future predictions. It is a perfect measurement of the past, but a poor indication of the future.
Many people think moving averages are like annual rainfall, that you have an average rainfall that is steady year to year. You approach the average over time. A minute to minute rainfall indicator that has rapid fluctuations in rainfall indicates almost nothing about rainfall in the next minute. On the other head, if you notice either A) there is a natural limit to rainfall or B) rainfall increases and decreases according to a general formula (derivative and second derivative have a pattern) then this is a *moderately decent* indicator for what rainfall should be.
When approaching an intersection CCI says the crosswalk matters, Signal 3 says gas and brake matters.
Hi everyone! (and Bertbin!) :-)
About backtesting. Phoenix will be (mostly) used on FXDD data, but backtesting is mostly done on Alpari data. Wouldn't be nice if we would have FXDD history data?
Ta-da!
Check out this undocumented and unreleased URL:
http://www.fxdd.com/mt1m_data.html
My history data from Alpari must be bad since I didn't manage to get any good settings file - whatever I did, backtesting chart went down.
Maybe this data will fix the problem.
Mario
P.s. (I don't post much, but when I do, I try to be usefull) lol lol
I was afraid I needed to host that myself! thanks for the updated link
It has been a long day, and Windows is *almost* installed on my Windows machine where I do development. Soon i won't be flying blind anymore. I did a little progress on Phoenix 6, but not as much as I would like since I did FAT table analysis and data copying about 30 hours this week. Using a text editor is annoying for development :/ I can read the help files on linux, but not the API documentation unless i go to the website. Enough excuses...
While windows was installing I updated the website and moved some stuff around. http://bestforextools.com/pf/index.html PhoenixFund has money on the wire, we are getting IB tracking login and information Monday, IB system is live, EA Hosting is ready for new customers, and I'm debating the whole settings file issue.
I noticed two new settings files posted, and this warms my heart very much. I will make sure Phoenix always has at least one currency with current settings available for new users. It would be great if we are able to keep 5 more currencies current all of the time as a community. I'd like to follow Hendrick's original advice, if we have more than one bad trade in a row (Mode 1 Maxtrade 1) then it is probably time for new settings. I'll post my updated settings file this week, it performed well in the first week of live trading.
If we can't get strong community development of 5 settings files for other currencies, I'm seriously considering spending a week and optimizing all week long. Since I was able to create one settings file that has (fingers crossed and knock on wood) done well so far I should probably spend the time and do it for other currencies as well. At this point, we have a few leaders, a few followers, and a lot of sleeping sheep. If I do spend a week optimizing 5 currency pairs, I would ask for $50 from people who haven't donated or signed up with PhoenixFund as the IB. I see Phoenix as a free (as in 3-5 people doing a ton of work) cash machine. Having one ATM card for free and easy is available for everyone. Having 5 more ATM cards probably means more money - and if one works, why not spend time to figure out more? If you can't figure them out, and the community isn't doing it, then I'd like to see some other way of doing this. Any suggestions? Reactions to the idea of 1 Free ATM Card and 5 more for $50? I'm tired, but this is something that has been on my mind for a few days. I'll gladly support any free competition to the settings files I create by testing them, putting them on a website for download, etc. If the community does well, then nobody will *need* to buy the other settings files at all. I just didn't think I'd be writing code, doing banking errands on Money I can't touch, and making settings files through optimization.
With that said, I'm still trying to get the Windows machine back up and running so I can backtest and optimize.. Night all!
What settings be these, Daraknor? Can you upload a modified EA or Settings file? Perhaps we should try them out on other brokers. PContour's modification produces a constant string of messages about losses and Breakit 90 50 10. Can't tell whether the trailing stop is working though.
I removed the Print statements a couple of days ago. By printing out these things, I was able to verify that Decrease factor and the TS were working correctly. There's not much to say about it now since they are gone. We take a lot of things with faith in Phoenix. I felt it would help people to see what was happening and more readily believe in the code.
Losses, was showing the number of losses in the last 3 trades and the number of lots being used as a result. Breakit was making sure the code was breaking it (it being the decrease factor entered by someone), into 3 parts correctly.
These two things, the TS and the Decrease factor, were kind of dangerous before I modified them.