| stevehopwoodforex.com https://www.stevehopwoodforex.com/phpBB3/ Print view |
|
| What's cuzgeorge demo testing? https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=848 |
Page 14 of 15 |
| Author: | AGT [ Sat Sep 14, 2013 7:08 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Hi George ! Many thanks for your analysis of the hugh testing work you did the last months .... years ... ! Could you perhaps give any (at least rough) recommendation of the bots you have tested: I mean which of them you would consider as possible live trading candidates (and which others better NOT) ? Are there any better/stronger candidates among the good running bots ? BUT: of course at the end anyone has to decide on its own what is good for himself.  |
|
| Author: | cuzgeorge [ Sat Sep 14, 2013 9:26 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Hi AGT,all. I will leave a comment when i complete the reviews about where to throw some attention but the bottom line is one has to come up with ones own conclusions as you state. My demos can help to see what was tested and with what parameters. I include the EA's , not to make it easy, but because i sometimes mess around in the code. I forget with all the changes, but a simple web comparisson site, or examdiff, if you have it, will spot all the differences, if any. I'm not a coder, so changes would be minimal. I was able to add JL's code for the cover on some previous EA's, but JL has added to his latest so not necessary for me to temp ol luc to a party. Just explaining my skills are not good in that dept. Xmeters are good for me because i understand enough to trade them live, and again, years of trading them demo helps alot. Ive recently had the good fortune of putting alot of info gathered here through TIG, TIB, TIG and of-course SH, to good use with xmeter. Running on my live accounts i have xmeter do all the thinking for me, but picking up manually some positions and letting xmeter do the rest. Mackoazy and i were following some trades together and were lucky enough to pick up the AUD rise early enough to get some cash back into the accounts. Personally i take big risks so i dont advise anyone to try and follow me unless they willing to reload a blown account. I still over-do the lot sizes when manually trading, still working on that.Currently took on the metals and expecting a big short on gold. Big lots as i got over zealous and now i'm sitting with big losses. Xmeter does not trade the metals for me, there is a way to make it but will only trade L1 positions. I prefer to trade the metals manually anyway. NB's advice on waiting for a pull back has been invaluable, also the captain has brilliant points in his naked trading thread. So, using pending orders at pivot lines to get back in. Again, as always mentioned by Steve, one must understand a little of coding, or as much of as possible, and then patience. NB's posts are very informative in finding pivots. I'm a silent lurker in NB's threads but picking up pieces along the way. With a lot of words, i'm trying to say that manual trading is playing a big part in my real trading, but watching all the demos to see which EA's are picking what pairs , long or short etc, then i decide and enter with xmeter as my EA and magic at 0. If i'm wrong, xmeter will go into recovery and hopefully, not take too long to get me out of a bad choice of positions. Most of the time i'm getting it right lately and xmeter TP's me out in L1. You can set a TP in the pending order and xmeter doesnt take it over if you prefer. Another EA that almost always gets the direction right is Phill and Niepces Marylin. Its what happens afterwards thats the problem there, but that EA seems to almost always get the direction right initially for most of its trades, not all , but most. For eg, it picked up the gold rise almost immediately, i closed most of them manually, but will see if any were closed by Marylin when i get there. Those systems are system 4 and 5 that have the majority of demos so its going to be a while before i get there. Will make detailed feedbacks on those if i can recall the changes or if i have noted changes correctly. VM's Crossfaders are also extremely interesting and JL's NSNA is now the one that has me mesmerised. Getting those demos still set up, and watching to see what initial results look like. Looking good though. The Crossfaders are still battling , with the older versions doing better for me. WIll also respond with those when i get there. Too many EA's and i could go on all night here, so lets be patient and we'll get a good idea where to concentrate and what works better. SR for instance is definately great in theory for me, but terrible in practice. The various EA's will help us get a better idea where to follow. Watch JL where he's following as he is looking for that trigger that will make all the difference. Unbelievable coding from him, extremely versatile and great to see his versions coming out blended into xmeter. Hope i didnt babble too much,  kind regards and thanks for reading here. cuz |
|
| Author: | AGT [ Sat Sep 14, 2013 9:57 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Hi George - its every time great and encouragingly to read your posts !  I'm curious about your forthcoming posts - infering from the number of bots you tested they will be overarching ! Especially I have a question about "Merylin": its still not clear for me which mq4 version the testers use - I mean: does there really exist any version of the bot equal to the windev version Phil is using all the time so successful ? As I understand no completly bugfree Empty4 mod is available - thats why I'm in doubt what the existing/posted Merylin Empty4-EA's are doing indeed.  |
|
| Author: | cuzgeorge [ Sat Sep 28, 2013 1:36 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Hi all, updating post 1 with new links for 3 fxbook accounts. One will be for live accounts, a second for current EA's, and a third will house the Blown or Stopped EA's for all to view and compare. All info, including settings and what attempts-changes were made, should always be available for years to come. Please be patient with me. Priority right now is to get all the GP accounts that are stopping tomorrow, updated in their correct accounts, with info uploaded and saved. kind regards and have a great week end. cuz |
|
| Author: | Wiggy1965 [ Sat Sep 28, 2013 1:51 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Your the man.... |
|
| Author: | garyfritz [ Sat Sep 28, 2013 7:06 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
You're doing great work, George, and thanks for it! I looked at your "stopped tests" from post 1, and I wondered why you stopped your Gday 2 test? It looked like it was doing great after your April changes, averaging 10 pips per trade ... |
|
| Author: | cuzgeorge [ Sat Sep 28, 2013 9:57 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Hi Gary,  ,really good to hear from you.Referring to the demo you mentioned; http://www.myfxbook.com/members/cuzgeor ... for/395163 and this pic i'm adding; The pic is on page two and if you click on the blimp you'll see the comment for GDAY 2 , thats what would usually happen with GDAY, small wins, and a big drop like that. All profit would be eroded when this happened. After starting GDAY 2 when Steve re-introduced the newer version, i battled to understand GDAY, and subsequently lost all hope for anything to do with stochastics. I initially enjoyed stoch indis, but it was my turning point to prefer RSI. This was my lesson learnt from GDAY and GDAY 2. Then i decided to compare with xmeter that i understood much better, and joined xMeter v2.3.3 (the latest and greatest at the time), just to see what would happen when they both traded on the same pair. Gday 2 would take same pair trades, but end up in losses, whereas the xmeter series would close in profit. ALL the gains made on this demo were from xmeter and initially the buy only method speared by Fx8000. If you recall we were getting those tremendous gains (longs only) on anything XXXJPY at the time and been so obviously long, the attempt was to catch the pair again and again, when it would retrace and then rocket up north. This worked until the bull run ended with a huge retrace. Xmeter took a hiding with that bull run and taught me to open up the grid values. Gday wouldnt take some of the buys it was meant to take, so i gave up. Took my lessons and moved on, preferring RSI and xmeter with controlled M. The growth chart is for all closed positions, and open positions are still running actually with xmeter 13k, much more tame animal compared to xM v2.3.3 and still messing around with xM13k here. Will relabel and restart in the new relevant portfolio. I never found the perfect xmeter settings and i know i promised to send you them when i did. I still intend to find them, and i promise to post here when i get there. Its just that the magnificent profits are slowly getting less and less, as we look for safety from M. Thanks to your guidance/warnings here. JL is working on totally avoiding M and i'm looking at using it with medium-to-more-risky method. SOmething i wanted to ask you about referring to the demo linked below; http://www.myfxbook.com/members/cuzgeor ... red/508413 This demo was for Bears N Bulls and when i realised it wasnt working out, i started to manually play around. Then, noticed a script you wrote for the 'perfect hedge.' If you look at the comments tab, you will notice a number, like '1' or '2' or '3' etc etc. Each time i ran your script, i would change the magic to see if it would close its own series in profits. Then i thought adding MPTM to the mix, but never quite got there because some of the trades were supposed to lose, and therefore closing the winners, would negate the whole idea of the perfect hedge. Subsequently, that thread got a 'go to jail' card, so i couldnt look up the whole idea behind it, and what the thoughts were. I really liked the idea though and wondered what you thought of the script and method? did you ever continue looking into it? Basket closing is/was never my favourite, but here it made sense when the group, with the same magic, would hit a small profit, then closed it all out. I couldnt work on this alone, and hoped you might throw some light on this. regards and really good to read you in the threads again, hope to have everything updated soon for all. cuz |
|
| Author: | garyfritz [ Sun Sep 29, 2013 3:22 am ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
?? I don't think that was me. I'm almost as big a fan of hedges as I am of Martingales. It annoys me that people write hedged EAs, and then us poor USians can't trade them. And hedging doesn't buy you ANYthing, it just costs you more slip/spread/commission. So I don't remember ever writing a hedging script... |
|
| Author: | cuzgeorge [ Sun Sep 29, 2013 7:13 pm ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Hi Gary , Heres the script i was playing around with. I seem to think it was posted by you, but i also know you have a keen interest in helping many on this forum, so you must have a gazzillion EA's and code snippets on your PC. It seems fairly simple, and i was in fact using the jumping stop method with MPTM, but not as a basket. Each time the group/series would recover to a near-profit situation after closing the winners by JS method, i would open another series with a new magic number. Idea is great if i remember Brians thread, but my execution with MPTM was over-zealous and i dont think it would have worked like that in the long run. especially if there were long trend runs against the losing 'floaters'. There was no closing method except for the basket method per each magic number so maybe it would be interesting if we could expand on the script with a different approach. I dont think you continued otherwise you would remember the script, but would very much like you thoughts at the end of the day. Crim needs to have low spreads and allow very quick execution if a basket should go in profit. That was the theory, i tried something different with MPTM. If anyone would like to try it, remember its a script and should be used from the 'script' folder, moving the script onto a chart each time one wants to execute a series of trades, therefore do not install in the Experts folder. cuz EDIT- (just waiting to see if there is no copyright issues with the script, then will post here) |
|
| Author: | AGT [ Thu Dec 26, 2013 11:34 am ] |
| Post subject: | Re: What's cuzgeorge demo testing? |
Hi George ! Looking here your test with the XMeters anyone can see very impressive results ! You could mostly close the gaps between balance and equity !!!  So congraz to the professional !  I haven't such good achievements. But this only in consequence of my stupid running the bots only OOTB. Never changed any settings - and that's wrong. George is the master of tweaking them longest studying these EAs coded by Jose generously. So my question is: is the success with most of your XMeters an implication of the manual interventions you described already otherwhere ? Or perhaps a result of your special setfiles ... or most probaly of both of them ? Anyway: bcz I and no one else can copy your trading strategy, and I dont have your experience: its better for me to stop so far my humble XMeter attempts. Or is there any possibility/hope for an XM-EA/setting doing the job fully automatically on day ? Plz excuse my daft depressive remarks.  Surely I'm a bit disappointed - my last hope/bot is now the "NewCF7SR" doing (with changed settings - no OOTB !) pretty good (so far).  |
|
| All times are UTC | Page 14 of 15 |
|
Powered by phpBB® Forum Software © phpBB Limited |
|