Optimized Trend Trading | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Optimized Risk vs Reward Equation 335 replies

Loss Less Optimized Strategy 63 replies

Optimized Parameters in Metatrader Testing? 3 replies

"Optimized Parameter" feature in EA? 0 replies

Strategy with Optimized parameter 1 reply

Attachments: Optimized Trend Trading
Exit Attachments

Optimized Trend Trading

  • #21
  • Feb 1, 2008 2:28pm Feb 1, 2008 2:28pm
Quoting walb99
Hi giraia_br!

Are you familiar with muddbuddhas step MA? I have a screenshot and the indicator, but have difficulties with the settings.
sorry walb, but i haven't been with time to test his system by now. I hope Igorad or someone else make a EA so we can let the system running as it is 100% mechanical.

but StepMA is was made by Igorad and you will find details about it at ForexTSD forum.
  • #22
  • Feb 1, 2008 3:08pm Feb 1, 2008 3:08pm
"Seek" In who's opinion?

Quoting giraia_br
Mudd is one of the most respected guys here( like Seek, Smjones).
  • #23
  • Feb 1, 2008 3:13pm Feb 1, 2008 3:13pm
Hi giraia_br!

Thank you, just found these...
Attached Image (click to enlarge)
Click to Enlarge Name: step iondicators 2008-02-01 22.09.gif Size: 35 KB
/ ... clarifying decisions ... /
  • #24
  • Feb 1, 2008 3:29pm Feb 1, 2008 3:29pm
This is great news! So many good folk out there have indicators just as good (or even better) than mine, and MUCH more colourful!! How exciting!

I hope you are all making good money each month from them ;-)
  • #25
  • Edited 3:44pm Feb 1, 2008 3:34pm | Edited 3:44pm
My opinion is Mudd knows his stuff. He's made many posts worth reading.
I agree with his assessment that keeping an open mind is essential - but I doubt he would trade an MA system in trendless market.
Most of the trend based tools dont work on a trendless market - (My state the obvious line
Lots of things change as a trader matures in his craft, you never throw out a tool comnpletely unless it breaks
- for instance I am to a point where I forego charts all together - I find I am only using them any more cause I like them and am used to them.
[NOTE: this one the original poster posted is kind of nice - looks like it was generated from a tool though - not a platform I'm familiar with - I likethe 'No Candle Wicks - But cannot tell if this is H/L or O/C ]
- but for pure trend determination (Yay/Nay) I use statistical calculations. And guess what statistically - MA's fail in trendless markets.

[ Note 2 ] Did a search onTop Cat and only found Rotamatic 2.1 CodeBreaker is this what you used -I can see how it could accomplish the task?
Quoting tdion
"Seek" In who's opinion?
In trading, there is no bullshit. You either make money or you don't.
  • #26
  • Edited 4:23pm Feb 1, 2008 4:03pm | Edited 4:23pm
Good grief!! At last a contributor with an open mind. I have only been in FF for two days and had almost given up hope already. Thanks for the positive comments, MrGreen!!

As you will see from my original post, TopCat is called that because it is an acronym. The original chart just shows H/L as that is all the study shown uses. O/C are irrelevant so they are not painted.

The "tool" is not a "platform" like TradeStation etc. I saw no point in spending good money on OP's software. In fact, my entire set-up costs were £0.00 (I even got the laptop from work).

I just wrote TopCat from scratch (hoho) in C++. It is very flexible, I can add any indicator in a few lines of code and can interface to any data feed, most kettles, some toasters and get automated alerts via SMS text on my mobile.

I am trading with IGIndex so as to avoid paying Gordon Brown any sheqels of my winnings. I have heard mixed reports about them but the trading platform is quite nice and they seem OK for the time being.

But the best bit it that TopCat can "sniff" their live data feed so I am updating my studies from the SB data (with SB bias and all) and not from the "real" market. This means that I am analysing the actual prices I am trading, which given the bias SB companies use esp. during news updates etc, gives much more accurate signals than the "real" data could.
  • #27
  • Feb 1, 2008 4:42pm Feb 1, 2008 4:42pm
Quoting Codebreaker
Good grief!! At last a contributor with an open mind. I have only been in FF for two days and had almost given up hope already. Thanks for the positive comments, MrGreen!!

As you will see from my original post, TopCat is called that because it is an acronym. The original chart just shows H/L as that is all the study shown uses. O/C are irrelevant so they are not painted.

The "tool" is not a "platform" like TradeStation etc. I saw no point in spending good money on OP's software. In fact, my entire set-up costs were £0.00 (I even got the laptop from work).

I just wrote TopCat from scratch (hoho) in C++. It is very flexible, I can add any indicator in a few lines of code and can interface to any data feed, most kettles, some toasters and get automated alerts via SMS text on my mobile.

I am trading with IGIndex so as to avoid paying Gordon Brown any sheqels of my winnings. I have heard mixed reports about them but the trading platform is quite nice and they seem OK for the time being.

But the best bit it that TopCat can "sniff" their live data feed so I am updating my studies from the SB data (with SB bias and all) and not from the "real" market. This means that I am analysing the actual prices I am trading, which given the bias SB companies use esp. during news updates etc, gives much more accurate signals than the "real" data could.
Interesting system you have there my friend.

Ahh and Topcat, classic cartoon!
  • #28
  • Edited 6:34pm Feb 1, 2008 4:57pm | Edited 6:34pm
Interesting on your 'sniffing' the data feed. I am working on a similar - using DDE feed from several individuel brokers to pull the data directly into my stat tool, but probably need to come up w/ an interface of some sort to act as a go between. I do reget not working with C more - but my career took me in a diff path.
My trading is longer term so the need for 'tick data' is not relevent though I may investigate its properties - at least store it somehow for analyses at a later time (I'm trying to convince my wife of the need for a few 800GB storage disks so I can create a "Home SAN" - but she is technically savvy so I got to justify it with something other than "I want it" LOL. -
I am automating the analyses piece so it will just spit out what I want after it checks the validity of the data. I am using MiniTab for analyses.
I agree that O/C is not relevent - it took me a while to grasp this concept. I make arbitrary O/C's based on the analyses of the data and the statistical significanse of one hour over another - but only for reference purposes, my thought process gets strange sometimes - but I like it like that

PS: Apologies for not thoroughly reading the first post - My questions were answered there.
In trading, there is no bullshit. You either make money or you don't.
  • #29
  • Feb 2, 2008 2:56am Feb 2, 2008 2:56am
Hi SeekingLight!

You are absolutly right in your statement that a crossover or whatever is not the point. But what is the point then???

Can only be "confirmation" also knows and "the ands".

By the way Codebreaker, in my life I made the experience that I learned most from highly arrogant people. Strange isn't it?

Best regards
/ ... clarifying decisions ... /
  • #30
  • Feb 2, 2008 5:19am Feb 2, 2008 5:19am
Don't give up on the FF members just yet, Cb. There are some extremely gifted traders and programmers 'within these walls'. Unfortunately, I cannot count myself amongst them and have to work very hard at what I do, which is to survive in this business and hopefully show a profit at the end of the month. I'm getting better at it...!

Gripping stuff so far and v interesting. I too use a SB co (Finspreads) for exactly the same reasons... so G Brown can't give my money to all and sundry, except where he should (but that's another story for another place). So your 'sniffing' paragraph perked up my Britney Spears, I found this to be an interesting concept using the SB data, which often gets me stopped out of trades (where the 'true' price wouldn't have) and often varies considerably from the 'true' data.

Looking forward to reading more
Great thread n don't despair
Dave
  • #31
  • Feb 2, 2008 5:25am Feb 2, 2008 5:25am
Perhaps great minds do think alike, MrGreen! I like your ideas of doing stat analysis on data from various sources. In another life I dealt with "intelligence" and there we set much store on confirmation of raw data from independent sources.

BTW, on you disc problem, Maplin were having a sale just before Xmas and I bought a very nice Maxtor 400Gb stand-alone disc that just plugs in the USB port. Cost £59.95. Jolly good value - and large enough to back-up my entire brain if I thought I might be going to a particularly wild party where too many brain-cells might get killed.....

Like you, I don't use tick-data. TopCat's companion software, Mongoose, just sniffs live feeds by talking directly to the ethernet packet drivers and pulls in every tick. It then combines these into bars for any time frame you want. Could be 1s, 30s, 1m, 30m, 1h, 1d etc etc etc. You could have 4.435 second bars if you were so minded! It works like the GlobalServer in TradeStation but an order of magnitude more flexible.

I only EVER use 1h bars. The reasons for this are deep and based on my long analysis of the Shannon Entropy (and hence Shannon Mutual Information) of the market at different timeframes. If you plot a curve, you find that tick, 1s, 1m etc have an awful signal-to-noise ratio (SNR). This is because price movement is purely due to market makers buggering around with the prices. Likewise, very long timeframes such as 1d+ are not "technical" becuase they are driven largely by world events or sometimes, some irresponsible idiot from the ECB popping up on Bloomberg and saying something off-the-wall.

The 1-hour bars are at the minimum of the curve and allow the best signals to be created. It is also a nice, leisurely timeframe! I don't want to be frantically trading from second to second!! I get on with other things and just look at the screen when it goes "ping". I think the secret is not to trade frequently (people who do this are trading for entertainment) but to trade successfully!

When I opened my IG account, after about a week my "account manager" emailed to say that he noticed I hadn't traded and was everything OK. I replied as above that I was here to me ME money, not him. Haven't heard from the joker since....
  • #32
  • Feb 2, 2008 8:45am Feb 2, 2008 8:45am
I am reviewing a means of using price objectives as a time delimeter. I got the idea from a point and figure chart. Such charts remove the time element and I would like to crossreference this data to what I term statistically significant trade hours. This sounds similar to what you have already done w/ the analyses of the Shannon Entropy. I pulled up some research data on this and plan to take a closer look.

Great Idea.... A packet sniffer may solve my interface issue to Minitab. I have a Solaris 10 x86 server in the basement that I use as a sandbox. My Oanda Java based acct is configured on it already so it would be just a matter of configuring a few more java based platforms and then sniffing their individuel packets.
I will check into the USB drive idea as I should be able to attach one to the SOl10 box. This would also take some of the load off of my PC.
In trading, there is no bullshit. You either make money or you don't.
  • #33
  • Feb 3, 2008 2:19pm Feb 3, 2008 2:19pm
Quoting Codebreaker
Good grief!! At last a contributor with an open mind. I have only been in FF for two days and had almost given up hope already. Thanks for the positive comments, MrGreen!!

As you will see from my original post, TopCat is called that because it is an acronym. The original chart just shows H/L as that is all the study shown uses. O/C are irrelevant so they are not painted.

The "tool" is not a "platform" like TradeStation etc. I saw no point in spending good money on OP's software. In fact, my entire set-up costs were £0.00 (I even got the laptop from work).

I just wrote TopCat from scratch (hoho) in C++. It is very flexible, I can add any indicator in a few lines of code and can interface to any data feed, most kettles, some toasters and get automated alerts via SMS text on my mobile.

I am trading with IGIndex so as to avoid paying Gordon Brown any sheqels of my winnings. I have heard mixed reports about them but the trading platform is quite nice and they seem OK for the time being.

But the best bit it that TopCat can "sniff" their live data feed so I am updating my studies from the SB data (with SB bias and all) and not from the "real" market. This means that I am analysing the actual prices I am trading, which given the bias SB companies use esp. during news updates etc, gives much more accurate signals than the "real" data could.
Hi Codebreaker

I am also trading at IG and the difference between the price movement in my account and on the charts is terrible. I have no programing knowledge to date to do the kind of thing you have done. Have you come across any charting packages that 'match-up'. If so please let me know, it would be a great help!

Good trading

Leon

Ps. Great sunbathing weather were having isnt it?! lol
Living the adventure in my head.
  • #34
  • Edited Feb 4, 2008 6:08am Feb 3, 2008 6:25pm | Edited Feb 4, 2008 6:08am
Hi MrGreen, Acko, Leon

While not the thrust of my original thread, this business about data sniffing is clearly of great interest to a number of folk :-)

When I started in this business, I opened a "player" account with TradIndex. I chose them simply because you can keep the account for as long as you like (I had mine for 3 years) whereas IG only allow you 2 weeks!

I had every intention of trading with Tradindex (not knowing any better) and tried to refine my strategy. I was getting increasingly concerned about the differences between the TI charts and the "real" market. I didn't understand what was going on. It was only after doing some reading on several forums that I leared about spread bet bias and the various other tricks these people use to con suckers out of their cash. The last straw was one day when I opened a position and the (real) market moved 62 points IN my favour and yet I LOST 12 points with TI. Some are worse crooks than others but all have some devious ways.

So... what to do. Clearly the data from the real market didn't bear sufficient correspondence to the SB data to allow much of a profit. So I decided to gather the SB data and run my algorithms on that, to see what could be done. This involved writing some code to talk to the packet drivers to directly hack the SB servers to get the data I wanted. Admittedly, if you are not a programmer this is a daunting task but if you know what you are doing it isn't too hard a job. I had everything up and running in a couple of days.

The SB data turned out to be very interesting. The SB's dare not move prices by tens of pips at once, so they "slide" them up and down to suit themselves. Modelling this, we find a new and totally "synthetic" market with its own trends etc. OK, it bears only passing resemblance to the "real" market, but we aren't trading the "real" market - we are actually trading the SB data so that is all that matters.

So why bother with the real market? Just think of IG-EURUSD as a different cross to REAL-EURUSD and get on with it. In fact, the SB data tends to trend MORE than the real market due to the "sliding" of the prices. I am slowly learning to make MORE pips from these crooks than from brokers who actually give you the real prices!

I strongly recommend this idea of sniffing the SB data. If you want, you can feed it into TradeStation, MetaTrader or whatever platform you use. As I said, I wasn't prepared to spend a single penny (other than my own time) on this business until I was sure I could make a profit. So I just write all my own software from scratch.

I think we are on to something quite interesting here and if we put our heads together, we might have the seeds of a very powerful system.

Good luck...
  • #35
  • Mar 12, 2008 7:20am Mar 12, 2008 7:20am
Very interesting topic started here. Surprised it faded like it did. I agree w/ Codebreaker in that you guys are onto something quite interesting indeed. Unfortunate that it's so many naysayers, though. In my two years of discovering and researching trading and the subcategories within, I've found that there an awful lot of haters. I think it was Albert Einstein that said, and I paraphrase, "Anyone who has never failed, has never tried something new". Don't let the non-believers quell your progress. I digress......

I am a user of all the diff types of noise filters. More specifically Jurik indicators (which Codebreaker mentioned) and Digital Filters. I've yet to find anything better commercially available then the two aforementioned methods (though it looks as if CB may have created a contender). Ehlers and Hurst are pushing some good stuff as well, but I'm still trying to wrap my mind around their theories.

I do not want to advertise one forum on another (I believe someone already beat me to it in this thread) but there is quite a bit of information for such topics on Forex TSD (though they've also experienced a similar interest degradation). It's refreshing to see that there are still a few de-noising loyalist left. Hoping you guys are still around. Will wait patiently.

Little reading material for those interested:

Supressing Market Noise:

http://www.investopedia.com/articles...arketnoise.asp

Digital Filters:

http://www.fin-ware.com/article1.html

Also, attached is a little piece regarding Juriks

Peace,

F_F_L
Attached File(s)
File Type: pdf Jurik[1].pdf   540 KB | 7,324 downloads
  • #36
  • Mar 13, 2008 3:38am Mar 13, 2008 3:38am
Hello everyone,

I am still here, but snowed under marking exam papers etc (tribulations of the day job!).

The data feed sniffer is working well and I now have 2 months of IG-Index data to play with. I optimsed a few simple parameters (nothing sophisticated) and am trading live with IG. I now have 3 weeks of track record and have averaged around 300 pips/week. Trading the IG data is VERY difficult due to the bias and general crap on the feed. Using Dukascopy data, I average 731 pips per week. We are lucky to get 300 from IG.

The robot gets off to a rocky start when he starts up, as he is learning the market and optimising internal parameters, so I don't trade for the first day or so. Then, he starts to "bite" and the pips start coming in.

Anyway, been in the FX game for around 2 months now and reasonably optimistic that a few shekels can be made. I have a load of ideas for improving the robot (it currently contains only 2 naively simple rules!!) so will press on.

All the best

CB.
  • #37
  • Mar 13, 2008 4:45am Mar 13, 2008 4:45am
Hello

I was going to include in my earlier post a link to the live web page where the robot places his output (so I can access it from anywhere). But on reflection, I thought better of it, not least because it might be construed as "trading advice" which is strictly illegal under UK law. So I am attaching a 24-hour snippet of the robot output up to about 10 minutes ago as I sit and write this post. The robot was started 1630 on Tuesday with a profit of 0.0 and so you can see the rate of win. These trades are on the IG EURUSD with a 2-pip spread.

Enjoy.
Attached File(s)
File Type: txt LIVE_TRADES.TXT   119 KB | 2,134 downloads
  • #38
  • Edited 9:02pm Mar 13, 2008 8:22pm | Edited 9:02pm
Congrats to CodeBreaker on his recent triumph. Though only in it's infancy stages, I can see nothing but good things ahead as it matures. I'm sure most (non-haters) will agree that its always refreshing to see a fellow trader successfully profiting from his spilled blood, dripped sweat, and bitter tears.

I've not been numerically centered in my past which currently seems to be to my detriment. Now I'm playing catch up. Why, oh why didn't I pay attention in Algerbra II . I've discovered that such setbacks are a direct result of poor discipline. As a result, I've been working on spending more of my time on programming my mind since I am currently unable program an E.A.: I said that to say to all those who aren't "Math Mavens", focus on self for now and the external factors become much easier. I'm still @ my infancy stages, though I'm chronologically 22 yrs. old, but I am so grateful for learning this early and I want to pass it on to all who will listen. Codebreaker not only optimized his robot, he first optimized his mind.

It's a high probability that CodeBreaker's inbox is filling w/ request and propositions regarding his approach as I type. I was one of those when I first started. Now, I want nothing more than to find like-minded individuals w/ whom to bounce ideas back and forth. I've been working on reducing my exposure to forums because I've found that the majority are either newbies acting as digital beggars or oracles who forecast the doom of others. Very few positive and sane peeps. I saw CodeBreaker's post and the subsequent cadgers and cynics. Though it looked like he had it understand, just wanted to give him an assist because I know how discouraging it can be when first coming to the forums.

I digress..........

Regarding systems, the more I look and test, the simpler I wanna keep it. Juriks and Digital Fiters are great and all, but sometimes I feel like I'd be better off dealing solely w/ price action and support/resistance. Ideally, I'd be able to incorporate V/OI (Volume / Open Interest), however this isn't available on Spot FX. Only on centralized exchanges like currency futures on CME. Maybe do the ultra-smoothed and digital filtering on the side.

Speaking of smoothed indicators, I'm not spending anything for the Jurik indicators. There are several free packs on TSD forum. I spoke w/ Mark Jurik via email a couple of months ago and he confirmed that he is now programming mathematically correct indicators for MT4. He said that there is a flaw the MQL language of which is is trying to work around. He has a few interesting setups already created for other platforms that look promising. Check out the site. Worth a look.

Codebreaker,

I'm not sure if you're aware but there is a way to automatically optimize e.a.'s on MT4. Should make things a little easier for you. Might be able to adapt it to do the same w/ your Dukascopy data. Link below:

http://articles.mql4.com/506

Sorry for hi-jacking your thread.

Peace,

F.F.L.
  • #39
  • Mar 15, 2008 6:34am Mar 15, 2008 6:34am
Hi Guys

Attached the TRADES file from where the last one left off, up to 2046 last night when I shut the robot down for the weekend. (I was going to a party with my MSc students and suspected I might not be in a good state to do things on the computer when I returned......).

The robot started 1630 Tuesday, so ran for about 80 hours and made 420 real pips profit. This is about 5 pips/hour which doesn't sound a lot but I guess it mounts up over time.

I am considering a much longer post when I get some time as I am much encouraged by FFL and folk like him. I have some quite deep thoughts on this that I think might be of benefit to share.

All the best,

CB
Attached File(s)
File Type: txt TRADES_2.txt   177 KB | 1,215 downloads
  • #40
  • Mar 15, 2008 8:05am Mar 15, 2008 8:05am
Quoting Codebreaker
The robot started 1630 Tuesday, so ran for about 80 hours and made 420 real pips profit. This is about 5 pips/hour which doesn't sound a lot but I guess it mounts up over time.

Hey CB,

just curious what this actually means in real world ROI / %.
All over the forum people perpetually use pips while they mean nothing(positive pips over a year can still mean negative income and vice versa).

How much return did the robot generate in that time period?
Is there an average size initial stop?
How much risk in % is placed on that?
Is there a set TP or a trail or simply a counter-signal?

This is the interesting stuff imho

It's also interesting to see a purely mathematical / automated approach being tested.

No worries, I still loathe MA crossovers though


Have fun at the party,

SeekingLight
Trust price. Know yourself.
0 traders viewing now