Not logged in
Viewing 15 posts - 61 through 75 (of 82 total)

Forums>StrategyQuant>Extras & Strategies>EURUSD on the 1 hour chart

  • #241355|
    Customer
    484 Posts

    I think it would be very helpful and educative for most of the readers here, if you could provide a practical A-to-Z example of how you are using your auto-optimization code and how you are dealing with the issues I mentioned earlier in order to get some results, any results.. P.S.: I do not care much about your trade screenshots and life-style comments, in most cases this triggers a “false narrative” waring to me.

     

    Here some more “false narrative”, on par with the stuff on your website except this was built in 2004-5 and has maintained its performance:

    The thing about the portfolio is it is reoptimised every 4 weeks and traded for 1 week.  Sometimes the reoptimisation will only produce 1 trade but this performance has generally been consistent year in year out.  Don’t get me wrong I haven’t fully benefited from its performance; it quickly becomes tiresome having to manually re-optimise so often.  The software has about a 0.01% efficiency optimization wise when compared to SQ so it is a real drag.

     

    I am not bothered in the slightest if you believe it or not or believe me or not.  I couldn’t care less.  However, you’re missing a glaringly obvious point about sample to weight ratios however if you’re a 100% datamining type it shouldn’t make a difference anyway.

     

    Regarding the re-optimisation code; it is there, I have contributed it and if you think it is garbage don’t experiment with it.  Provision of the code is more than sufficient.   Why? Because it is a blueprint of how it can be done programmatically something we haven’t had so instead of your laconic tit-bits, try writing thank you.   The lack of gratitude astounds me.

     

    Regards,

    Notch

    P.S. The above post cost me 10 pips opportunity benefit!   ;)

    #241357
    Customer
    484 Posts

     

    … was built in 2004-5 and has maintained its performance:  

     

    The length of time I’ve been doing this is worth emphasizing.   I invested in my first algo’ builder in the year 2000 (I actually think it was 1998 but the memory fades because I#m not as young as you may think) called TradingSolutions and Neurosolutions.  I first started performing walk forward (rolling) tests in 2002 before it went mainstream.   I am formally educated in statistics, computer science, A.I. , finance, economics and to top it off Accounting – formally: UG, PG & PhD.

    Whether you decide to believe it is unimportant however I only offer opinions I KNOW work based on 20 years of research and practice.  So apologies if occasionally I brush off remarks such as the P-value comments because much of it is genuinely elementary stuff for me.

    I do not write this to be better, I just don’t want to be constantly flaming the innocent folk of the forums.  It is tiring for all.

     

    You don’t have to agree but if I take the time to offer an opinion, it is because it works.   There is zero guesswork at play here.  Zero.

    However, I still cannot build scalpers!!!!

    p.s. I cannot believe how well I profited from that GBPUSD 100 pip upward spike especially starting from a short position!!!  ;)

    I should start my own thread called “Notch’s rambling”.  lol

    • This reply was modified 5 months, 3 weeks ago by  notch.
    #241360
    Customer
    78 Posts

    Wow, 20 years in trading is indeed something, but I think in those 20 years you have missed two most important lessons:

    1. Respect towards other traders, especially newbies (people that need help), people that are as green as you were 20 years ago. You must understand that 99% of people here do not even understand your posts, because they cannot follow it. So indeed you are totally right, in that case you are just rambling :) LOL
    2. Some basic human communication skills without attacking others?

    I will happy to teach you 1 & 2 for free, no strings attached! Instead you will show us how to get the information out of 1 sample.
    Information from one data point which is totally buried in noise. So far I know only a handful ways to get info below the noise levels, like e.g.:
    – oversampling = increasing of the effective resolution
    – modulating signal and switching to frequency domain

    But, this cannot be successfully applied to trading (time series). So I’m (and all other people here) are really interested in your way of approaching this.

    #241364
    Customer
    397 Posts

    SQx have functions that allow You to test almost any type of way to find good strategies thus all features that have been implemented by requests from traders that I think combined have 500+ years of experience . You can even test statistical significance  which I have done and it has no impact either. Throw away the “fake” books and use SQx to find ways to improve stability from strategies going forward. Hankeys strategies with RDD around 20 is very good strategies. For evry dollar they lose they make 20. They are easy to find it just take different long time depending on what putors You have.

    #241367
    Customer
    484 Posts

    Wow, 20 years in trading is indeed something, but I think in those 20 years you have missed two most important lessons: 1. Respect towards other traders, especially newbies (people that need help), people that are as green as you were 20 years ago. You must understand that 99% of people here do not even understand your posts, because they cannot follow it. So indeed you are totally right, in that case you are just rambling :) LOL 2. Some basic human communication skills without attacking others? I will happy to teach you 1 & 2 for free, no strings attached! Instead you will show us how to get the information out of 1 sample. Information from one data point which is totally buried in noise. So far I know only a handful ways to get info below the noise levels, like e.g.: – oversampling = increasing of the effective resolution – modulating signal and switching to frequency domain But, this cannot be successfully applied to trading (time series). So I’m (and all other people here) are really interested in your way of approaching this.

     

    Let’s just agree to disagree and keep it moving.

    #241368
    Customer
    484 Posts

    I put 1000 people without a knowledge of C++ and send them away for a year to learn how to code, ask them to return in 1 year for a focus group which will be used to create material for advanced C++ programmers with 5 to 30 years in the game.  That means that would have 1000 years of experience to draw from which the advanced C++ will cherish and love.

    Absolutely asinine.

     

    I always thought the RDD looks at the final trade outcome relative to maximum adverse excursion then expressed this as a ratio whereas the profit factor was used to make a statement such as for every dollar they lose they make 20 because it actually compares actual loss to actual profit and not just how far a trade had moved against them.

    Hankeys average profit factor from his public models is 1.10;  The models are making money though which is great.

    Point is who cares if it produces RDD 20 or 200 straight out of SQX if it is only going to perform the same as a model that came out of SQX with 1.1 to 2 RDD.

     

    Anyway, I am done…at least before the discussion managed to cling onto basic logic even if we disagreed; now it’s just daft.

     

     

     

     

     

     

     

     

    #241370
    Customer
    78 Posts

    I see that cooperating, or even exchanging new or ‘other’ ideas will be very difficult here….on the other hand we can see people are achieving some good results using SQ (X). So maybe it is a good sign and everybody should keep doing things their own way.

    “There’s more than one way to skin a cat”.

    Moreover, sharing your best strategies in public is also not the best way to go…if you will share your best strategies here, people will sell them tomorrow on mql5 market using a different name ;)

    So what’s left? Exchanging the strategies with other SQ users? In a 1-for-1, strategy-for-strategy manner? In that way we can at least stay diversified, which is good for our portfolios. This, while accepting the risk you will exchange your good strategy for a shitty one ;)

    What do you thing about this idea?

     

    #241371
    Customer
    484 Posts

    Sorry skipped the detail.

    LOL 2. Some basic human communication skills without attacking others?

    1 word: Aspergers

     

    Information from one data point which is totally buried in noise. So far I know only a handful ways to get info below the noise levels, like e.g.: – oversampling

    Let me give this some thought.   After so long, much of the stuff in my head is now just a blob of knowledge so unpicking it and finding the beginning of the threads of thought may take a while.

    I’ll post stuff as it comes to me in my rambling thread to avoid the completely illogical and asinine stuff.  Not sure I can do 1 data point, but 10 for sure.

     

    modulating signal and switching to frequency domain But, this cannot be successfully applied to trading (time series). So I’m (and all other people here) are really interested in your way of approaching this.

    I assume you haven’t experimented with the Fourier Transform, Hurst Exponent or PCA, right?

    • This reply was modified 5 months, 3 weeks ago by  notch.
    #241373
    Customer
    78 Posts

    1 word: Aspergers

    Ok thanks for sharing this. This explains few things to me.

     

     

    #241374
    Customer
    484 Posts

    For clarification the RDD of 211 does not mean for every $1 I lose, I gain $211!!!

    More like for every $1 I lose, on average I gain 2.64.  Elementary stuff people!

     

     

    Works on the demo, recently transferred to live, so I really do not know at this point but with that PF and trade quantity…add a little leverage…$$$$$

    Report for 2003 to date.  The most significant take away is the need to use the bid, ask and real tick data with variable spread when building scalp models.  However, it only works on the demo and this is my most successful attempt at a scalp model so I won’t know until the live results are in.  Basically, the take away is low quality at this point and may leave a bad taste in your mouth.

     

    #241379
    Customer
    484 Posts

    What do you thing about this idea?

    I’ve never understood the wink  ;)  !

    ======================================================================================================

    1. Create discussions that are very tight and focused.

    2. Assign a discussion moderator to ensure the discussion stays focused.

    3. List clearly at the beginning of the thread the loose outputs one is hoping to generate in the thread.

    4. Have one main thread for idea generation where members can thrash things out.  The idea generation thread is the location where debates can be lively and raw; this also serves as the location where members decide if a particular idea should be elevated to its own thread for very focused and serious discussion.  Although raw and lively discussions can be like pulling teeth, it does produce valuable content.  This lively and raw thread has certainly been like pulling teeth but it has produced some interesting pieces from a number of members which for me are worthy of additional thought, for example, it been about 10 years since I performed some Fourier Transform work and models augmented by the Fisher Transform, perhaps its time for me to resurrect that work.  That probably would never have occurred to me if I hadn’t read the lines between the insults of a fellow member.

    By structuring discussions in a fashion similar to the above, members are able to draw a line beneath one type of discussion then take its most valuable outputs for a higher quality, disciplined and focused level of discussion which would be moderated.

    I think throwing the baby out with the bath water approach would definitely be a mistake for many here.

    #241389
    Customer
    78 Posts

    For clarification the RDD of 211 does not mean for every $1 I lose, I gain $211!!!

    RDD in SQX is just netprofit/drawdown, so in your case = $4396.2/$20.75 =211.865 as visible in your stats below the chart.

    Scalping FX markets is not easy, I’ve spent many many months (years) on that and have only one scalping system that works and only on one broker. Modeling on broker A tickdata and running it on Broker B is pointless. Spread is on thing, but there are also slippages, latency, commissions and swaps. Scalpers will also perform differently on different VPSses and different broker server connections. Scalping is like being Alice in wonderland….

    I do not think you will get much out of the discussions here, if you really want to learn something new e.g.: going into frequency domain, doing FFT stuff etc… maybe you need talk to people like John Ehlers: http://www.mesasoftware.com/mesa_phasor.htm

    Below fresh testing results of my scalping system on real account.

     

    coensio scalping system

     

     

     

     

    • This reply was modified 5 months, 3 weeks ago by  coensio.
    #241391
    Customer
    484 Posts

    For clarification the RDD of 211 does not mean for every $1 I lose, I gain $211!!!

    RDD in SQX is just netprofit/drawdown, so in your case = $4396.2/$20.75 =211.865 as visible in your stats below the chart. Scalping FX markets is not easy, I’ve spent many many months (years) on that and have only one scalping system that works and only on one broker. Modeling on broker A tickdata and running it on Broker B is pointless. Spread is on thing, but there are also slippages, latency, commissions and swaps. Scalpers will also perform differently on different VPSses and different broker server connections. Scalping is like being Alice in wonderland…. I do not think you will get much out of the discussions here, if you really want to learn something new e.g.: going into frequency domain, doing FFT stuff etc… maybe you need talk to people like John Ehlers: http://www.mesasoftware.com/mesa_phasor.htm Below fresh testing results of my scalping system on real account. coensio scalping system

     

    RetDD yes and DD is in SQ is average close to close MAE but semantics… the point is PF is used to gauge $1 loss vs $gain

    Thanks for the link.  Been following him and his site since the Swiss Army Knife indy days.  The SWAK used to be an outstanding NN input.

     

    Nice scalper, am I reading you report right?  Has your scalper only made 72 trades since October 2018?  At this point, building a decent scalper is more about the challenge.  It is something I have been biting at for years but I simply fall flat every time.  I use TDS with slippage and latency etc. In short, I will not feel complete until I have built a good scalper.  Let’s hope my current scalper works on a live account then I can move on.

    • This reply was modified 5 months, 3 weeks ago by  notch.
    #241395
    Customer
    78 Posts

    Yes I’ve also tried short term (minute) scalping also with several other ‘experienced’ traders, we tried “wick hunting”, peak detecting, peak reversals, candles with wick without wicks, candle patterns, all standard indicators etc…you name it….I’ve coded several different proprietary indicators for that investigation. The issue was that all systems behaved very differently on live accounts than on demo and in the backtesting. Modeling was very difficult, nothing worked. So instead of real scalpers, we concentrated on ‘fast’ in-and-out intraday high probability trades. Even then it took several months to find something with profits.

    Scalping is not easy…

     

    #241401
    Customer
    484 Posts

    Yes I’ve also tried short term (minute) scalping also with several other ‘experienced’ traders, we tried “wick hunting”, peak detecting, peak reversals, candles with wick without wicks, candle patterns, all standard indicators etc…you name it….I’ve coded several different proprietary indicators for that investigation. The issue was that all systems behaved very differently on live accounts than on demo and in the backtesting. Modeling was very difficult, nothing worked. So instead of real scalpers, we concentrated on ‘fast’ in-and-out intraday high probability trades. Even then it took several months to find something with profits. Scalping is not easy…

     

    You ought to build an anti swap charge EA which closes all negative swap positions 1 minute before Wednesday’s rollover then reopens positions (or places limit orders etc) after rollover using the same magic number as before.  I bet that would sell.  I’m sure it would cause a few issues but it may be worth considering.  (it could also give the option not to close positions but to widen stops)

     

    Scalping is not easy…

     

    Agreed.

    • This reply was modified 5 months, 3 weeks ago by  notch.
Viewing 15 posts - 61 through 75 (of 82 total)

You must be logged in to reply to this topic.