What's cuzgeorge demo testing?

[sportbuf3]

I AM EXCITED.... I recently coded a new EA that I am evaluating which I hope to post in the future. I can't post it now since it does not yet meet all of Steve's criteria with length of time tested and testing with real $$ (I am not using real $ until the EA has several months of demo success).

Anyways... today's results from demo account morning trading looks extremely promising with $984 profit. 4 losers = $42 total. 46 winners = $1,026. This was on a $7,080 starting balance in the demo account. If the balance were higher there would have been greater profits but due to my margin min I did not have large enough balance to open other possible winning (or losing) trades. All trades were 0.10 lot size with 2 orders per pair that I traded. Close out for small profit on order #1 and let #2 ride to high profits. The EA seems very promising, but only additional time testing will determine long term profitability.

Attached is a screenshot of today's most recent orders closed. I will provide additional updates as I continue to test drive this new EA.

Thanks sportbuf,
look forward to testing your EA when you're ready to post it.
Your results look very impressive and i see your trades reflect trading all sessions. Most the EA's did rather well today, so correctly stated that you would need to run a few forward test runs and maybe a few Backtests if your code allows. Good luck and thanks for the post.
cuz

Thanks CuzGeorge. I have been working for many years to develope the perfect EA and this seems to be the most promising yet. It revolves around 10 custom indicators that I coded which shows change in both long/short price movement in D1/H4/H1/M15/M5 when all are pointing in the same direction. A day like we had today was perfect for this type of EA since price had long trends. It will be interesting to test in a more volatile market or when price movement is sporatic on short term basis. But I have another trendline indicator that will be referenced in the EA to help catch these short spike moves. You never know until you try whether it will work or not.

When it comes to inventing the perfect EA I swear that I have experienced more failures than Thomas Edison and Ben Franklin had combined!

[cuzgeorge]

It revolves around 10 custom indicators...

Hi Sportbuf,
please dont mention this to Steve, 10 indicators, he'll flip if he reads that,

When it comes to inventing the perfect EA I swear that I have experienced more failures than Thomas Edison and Ben Franklin had combined!

Now we know you are a real coder, as there is no perfect setup, or probably ever will be.

A quick question about the method, it sounds like you are looking for a trend and then sticking to it. Do you have any recovery positions worked-in, hedge, grid ? etc etc, as it sounds your EA will be appropriate for US traders too.
cuz

[sportbuf3]

Now we know you are a real coder, as there is no perfect setup, or probably ever will be.

A quick question about the method, it sounds like you are looking for a trend and then sticking to it. Do you have any recovery positions worked-in, hedge, grid ? etc etc, as it sounds your EA will be appropriate for US traders too.
cuz

Cuz,

Over several years trying to figure out how to develope a profitable EA of when to buy or sell in FOREX, I came to the conclusion that there are many strategies that work sometimes, and none that work all of the time. I conclude that the reason for this is that you are always trading a trend, but it could be a 4H or 1D or 1W or 15M or 5M trend that you are in. That sounds easy since a trend is a trend! But the problem is that if you trade based on 4H trend, it is also in a 15M trend which may be exhausting or reversing. Then your entry is too late into the 15M where it turns on you before realizing any profits. On the flip side, if you trade a 15M trend and it is also in a 1H trend, you get stuck in too many drawdowns or consolidations where the 1H trend changes from retracement into main trend direction leaving you with sudden change and losses in the 15M trend. Of course there is aways high TF consolidation that ping-pongs you into loses with lower TF. Ex: In a 1H consolidation and you trade a 15M trend. By the time the 15M trend shows a trend, you are at the top or bottom of the 1H consolidation where the both 15M and 1H reverses against you. The long and short is that I don't think that you can develope an EA on Empty4 that can analyze all time frames to compare and identify all time frames to ensure accurate entries since when you trade with a trend you are also always trading against another trend. (The tend is your friend only if you have picked the right trend! Identifying who is really your friend is the trick for succes.)

Based on the above (and to get to answering your questions) I decided to pick a single long term TF to trade with and enter against all other short TFs (my choice was combo of 1H and 4H). This came about after noticing when I entered a trade on lower TF based on all indicators saying BUY, that the market would actually reverse the opposite direction. This led me to notice that when all 15M TF aligned as a SELL trend start, that actually it was the end of retracement for the 1H or 4H TF. My EA strategy is based on the 4H/1H being in the retracement stage of its trend, and 4H/1H signals aligning for trend change (or in consolidation) and either the 5M or 15M signals indicate exit their trend which is opposite the long TF main trend. I have found that if the 1H is in a main BUY trend, that within 2-3 ticks when the 15M indicators say SELL, it is actually time to BUY. Entering the 15M trend (against long 1H trend) is recipe for lossing, so that should also equate to the opposite being a recipe for winning...right? Knowing that trends generally move in 3 steps, I know that I have 6 profitable steps with the 15M TF going with the 1H TF (or 18 steps with 5M). There are 9 profit steps of the 15M TF in the 1H trend, but it takes the 3 steps of the 15M TF to make the 1st step of the 1H to allow indicators to signal 1H trend. So... once the 1H or 4H takes its first step, my EA enters with a combo of 5M and 15M moves in the opposite direction and continues this for 6 steps (15M) or 18 steps (5M) and stays in the trend until the long trend exhausts.

While the above is the foundation of the trading system, it includes many other factors (indicators) as confirmation or entry timing. I have found that RSI, AO, and even MACD are not reliable. The problem with MACD is same with trend scenerio above where the (12,26,9) entry on the 15M is actually reversal where another TF is ready to move back into its main trend. But I have found value making an indicator that combines all TF into a MACD indicator which can assist with entry confirmation based on the 1H main trend where the 15M is strong in the opposite direction. I also have identified oversold/overbought points, not with RSI, but average max PIP moves per TF step. I coded an indicator which shows these levels and once price touches the line it is high probability (90%+) that drawdown will occur. I use these points as exits and I will not allow entry in direction of low TF trend when price is above/below those lines. In the past price would be above these max/min points when my 15M trends signaled entry and then price would move against me 1 tick after I entered the trade. By identifying these max/min points in a 15M trend I know realize that is time to enter opposite the 15M in the direction of the 1H or 4H main trend (where 1H/4H has recently been moving against its main trend).

Anyways... sorry to ramble on and hopefully I haven't been confussing. But the above is a brief description of my EA strategy. Grids = Kind of with my max/min PIP indicator. Baskets = no, but I have noticed when a currency moves up in a pair, that other pairs with that currency follow the same trend. Hedging = No. Multiple orders closing out in profit stages = yes. SL points = Yes and No. I do not use pice points but instead I use different SL method with a special indicator to avoid the SL hunting broker criminals from "spkining" me out of my trades. Limited TF/sessions = My EA is currently set for all sessions, but my plan was to limit entry (and exits) to only certain busy sessions. I haven't added that criteria yet and I was pleasantly surprised that my EA seems to work well regardless of session. I contribute this to the fact that my primary trend is the 1H and 4H TF. MAs = big part of strategy. I previously came to the conclussion that MAs were a waste since they only tell you after the fact the direction of trend. But I learned that if you pick a long TF (1H or 4H) and go with its MA and trade against lower 15M/5M TFs MA, then you are entering as price is begins a new leg of the long TF trend.

Thats all for now, but I will definately keep you in the loop as my new EA testing continues. It may be a winner... or it may be a wiener!

[garyfritz]

sportbuf3, I don't think Steve requires you to test for months with live $$ before posting your EA. He posts his EAs as soon as he writes them, and gets debugging help, feedback, & ideas from the community from the early stages. No reason why you couldn't do the same.

[sportbuf3]

sportbuf3, I don't think Steve requires you to test for months with live $$ before posting your EA. He posts his EAs as soon as he writes them, and gets debugging help, feedback, & ideas from the community from the early stages. No reason why you couldn't do the same.

Thanks Gary. I was basing that on Steve's posting with Bold Heading stating to read rules before posing new EAs. Within his rules, using a real account x number of months was one of the requirements. While I am not new to FOREX, I am new to this forum and I didn't want to start off by breaking any rules. I will wait to see what others say before stepping out of bounds on Steve's rules.

Anyways... yesterday I quoted "I am Excited" and today my quote is "I am depressed". My EA is taking a beating this morning, but that is to be expected since my trend is based on higher TF and yesterday's market moved big with the main trend. Yesterday resulted in $2,246 profit (32%) and so far today I am down about $1,000 (-12%) where I am giving back a large chunck of my yesterday profit. That means that I completed my 6 profitable steps in the 15M TF and I am currently in the 2nd step of the 15M TF or in the 4H drawback stage. I identify this fact and I could manually close my trades but I want to let them run to see the final results of when my EA identifies 3rd leg of 15M and whether it closes the trades as 1H reversals or leaves trades open recognizing the 3 steps of 15M trend as 1H tracement being complete. Once I get these results then over the weekend I can tweek the EA to exit 1/2 way into the 2nd step of the 15M TF. I will need to confirm with some chart analytics but I suspect if the 15M TF goes into a 2nd step against the 1H/4H then I could have exited the trade and saved 50% of the retracement losses. Exiting 1/2 into the 2nd step I estimate would have limited and cost me about -$600 today against $2,246 yesterday profits from main trend move. That would have been $1,800 total profit on a $7k demo account, which still is not too shabby. Thoughts?

[rosst]

Profit is still profit.... I don't think there would be a problem releasing your ea for others to test. There are a lot of smart folks here and it could speed up your process...

[garyfritz]

Where did you see that? I'm wondering if possibly you misunderstood the post. You wouldn't be the first to post an EA without extensive live testing.

$1800 profit on a $7k account would not be too shabby for a YEAR. If you can do that in a few days, AND do it consistently, you've got a monster!

[sportbuf3]

Where did you see that? I'm wondering if possibly you misunderstood the post. You wouldn't be the first to post an EA without extensive live testing.

$1800 profit on a $7k account would not be too shabby for a YEAR. If you can do that in a few days, AND do it consistently, you've got a monster!

Gary - Here is the link that I mentioned with rules for posting EAs. I may have misunderstood something in the post, but being a newbie I would rather error on the side that would not give me a bad name. http://www.stevehopwoodforex.com/phpBB3 ... f=36&t=519

As for the $1,800 being good. Well.... yea it would have been good, but the bleeding didn't stop at $1,800 profit. When all was said and done (and as of 10M ago I checked my account and it still is not "done" yet) I ended up with +$200 combining yesterdays monster profits with today's blood bath. But I am still confident that I can make changes to the EA where it would have saved close to $1k just by closing 1/2 way into the 15M reversal which was opposite my open trade direction. Closing 1/2 in 2nd step would be 50% of total 15M 3 step and even more if the drawback were an actual reversal. I didn't execute manually since I wanted to let it play out to see how the EA would react so that I could make appropriate changes to eliminate this from happening in the future. My way of thinking is that anytime the 15M TF is in it's 2nd step and price closes lower than 1st step low, that is strong signal that it is going to be a large drawback in the 1H/4H TF and possibly even a reversal. Currently my EA is not coded to close in the 2nd leg of the 15M step down, but rather is coded to close when my MA combo indicator confirms a 1H/4H trend reversal. (That will get changed this weekend!!!) Eitherway, even a basic drawdown in the 4H TF is major losses when entry was based of 3 steps of 5M and 15M TF moves.

As I stated earlier I do not use SL due to crooked stop hunting brokers, but instead my SL consists of price + MA averages confirmation. A day like today shows how I must change my SL exit strategy to get out more quickly. I will change the exit (SL) this weekend to get out when 15M enters a 2nd step in direction opposite my trade. I am also going to code to limit entry for specific sessions and tighten my indicators for quicker entry/exit during those busy session. I will still leave more room for drawback in slow sessions since these move are slow and usually small (painless), but in busy sessions drawbacks are huge and hurt like the devil.

I look forward to hearing your feedback on both the rules link that I pasted above, as well as any feedback on my planned EA changes. Also, thanks to both you and Cuz for the warm welcome!

[sportbuf3]

Profit is still profit.... I don't think there would be a problem releasing your ea for others to test. There are a lot of smart folks here and it could speed up your process...

Thanks Rosst for the comment. I am sure that there are many people here that are smarter than me (at least I hope so, if not we are all in trouble)

Believe me when I say that my coding is below par compared to some gurus that I have run across. In college I started out majoring in computer programming but I switched my major to Business. About 5 or 6 years ago I got hooked on FX and wanted to code EAs to test my stategies. I have learned enough coding to be dangerous but by no means an expert. Funny story (not funny to me but probaly will be to you ) when I first started I thought that I had figured out the perfect trade strategy that nobody else had ever discovered before. I coded my EA and it ran flawless (with good profits) for 1 week. I was so excited and opened a live account with $10k of my savings. For 2 days profits soared and I was ready to resign from my job. On the 3rd day of live trading I hear machine gun sounds coming from the alerts in my account. There was a glitch that had never happened in demo trading where on a double cross it was opening and closing orders (buy..close...buy...close) before I could figure out what was happening and disable live trading, my account was eat with commissions from machine gun trades and I had a balance of $4.36. To this day I have never traded another penny of real money and it will take a lot of positive historical performance to get me back into live trading (though that is my dream).

Anyways... thanks for the comments and look forward to exchanging ideas in the future.

[cuzgeorge]

Hi all,
Quick catch-up on whats been happening my side.
first, the bad news.
Ive lost one of the PC's and the really crap part is the Trendy Gary that was doing so well was on that PC. i'm going to rebuild it but because of my constant head-aches with the Ava demos on it, i have decided to slowly phase out all my Ava demos.
Starting from now, all the demos that are still showing, will be using the publisher method to show up and update on the fxbook thingy. I really cant figure out why this keeps happening to my demos, but going to start spreading out to other crims now. The bad thing is we wont be able to see the 'comments' tab with the publisher method. So really sorry about that, but all the newer tests will be on normal demos where everything will show again. I just dont want to lose some of the demo history that i think is still very helpfull to coders, followers and developers.

-So , TG that never lost had a very big loser in the GBPUSD, somewhere round the 10k mark. this is really a big loser, so if one keeps an eye on the GBPUSD for the near future, then we can take a guess asto where it would have ended up. The others werent so bad and there was also a winner on the USDJPY that was up at about 2,5k.

- AL3XX's 10.2 has finally come to an end with the demise of the PC, so lets see if Baluda can retro fit MultiBob with 10.4 and we can start a new test on that.

- a very profitable test also came to an end similarly with 2Sides from Harry45. I have some other demos running different versions of 2sides, but this one was seriously kicking with a huge profit recently. Going to keep on that EA until someone notices something we cant see. Russian translations may be a problem though. Not sure if the EA likes to be translated.

- Gimme Your Hair has also run its course and i have stopped on that one. For a scalper, it was very slow, and would have more winners than losers, only problem was the losers were bigger in value than the winners. Played with the settings, but i couldnt get to a decent set file to post.

- Dropped some of the older xmeters and working on only the newer ones, except i'm still running one of the V4 Speed EA's. I'm starting to stop the use of hedge and Follow and taking out the xmeter magic number to '0' so i can also make some manual trades that xmeter can take over. On one demo, i am buying and selling at the same time to see how xmeter will trade the two positions.
Evo 1 is doing nicely with the US set file i posted . JL's US set file has taken more winners, but was higher on DD. I will wait a while to gather some more info before posting more info on them in their thread.

- I have not been able to participate in JL's new threads and EA's. I'm absolutely , over-the-top this side with work, so following everyone there with their testing. Really looking good for the first 2 weeks on the EA's. Good luck there guys.

- BigMac will be updated this week to 0.15, still running with my twists including 0.8.2 with MPTM for TM and one from Renexxx 0.80. Both are doing reasonably well, although bit slow to trade. My 0.14 settings are similar to what acostafulano is testing and all are showing a very slight profit. it seems the testing has gone into over-drive, or at least the guys have taken the testing to a much higher level than i could possibly run, and i applaude AGT, Acostafulano, Wiggy, dudest and a few others (the names escape me right now) as they have covered every base i could think of. Well done people, brilliant work coming out from Bunker, Gary and the team that side.

- The Crossfader tests are coming on nicely, but watching one that seems to be getting out of its 'safe place'. It might be going towards a death-trade, but would have manually intervened if it were live by now. Will post the results in their thread. Another brilliant idea to combine the huge demo run on the EA. Well done VM and AGT. Please give me till Tuesday to get all the set files and info ready for a post there. I believe most must have been OOTB, but i think i tinkered with the pip gap settings, but

- Simple Beast seems to be trading much smoother on Global Prime demo. Dont know why, settings are exactly as they were on other demos crims. I'm starting to like this one more and more. Have not needed to interfere yet, but have already retired all the other, older Beast versions. This version was modded with the help from JL and snoopier and available in the thread by ribello under Simple Beast. Pippo is keeping a watch on the positions, so good to know as it appears Pippo is well experienced with the Beast's trades.

-FXPoolers EA is also doing okay and will be following that one up in the course of the week. Gathered some history so will take a closer look during the coming week.

-Havent got to The 9's yet, although thread looks interesting. RJo has quietly been moving along with the development there. Will post link and info in the thread during the week. I think i got enough history to work with here also.

Will see if i missed something and post again. Very tired this side, so will catch up with all in the threads.
good luck for the week and may the pips be with you.
cuz