How To Build an Automated SCT Trading System (Part 2)

Discussion in 'Automated Trading' started by greaterreturn, Oct 6, 2008.

  1. Tums

    Tums

    Wayne...

    A tape can extend beyond 2 bars.

    The code I posted is NOT the comprehensive Auto Tape Drawing program...
    but rather, it is a first step in understanding how to use EasyLanaguage to build tapes.

    My posts focused on the basic 2 bar tape.
    Although its capabilities are limited as is, the tapes can still be used in understanding how price develop within the tapes... then breaks the RTL, and turns into an opposite tape... which in turn forms traverses, and then channels.

    I will go back to the posts to clarify the educational nature of the code.

    TUMS
     
    #21     Oct 7, 2008
  2. Sensational!

    My initial efforts at automating the SCT trading based on the automated logging were very fruitful!

    I started by using the the tape direction, position of price in relation to the tape, direction of price, momentum of price, direction of volume, and momentum of volume.

    After entering several combinations of these I was able to pick off the highs and lows for 340 pips (after costs) on one day of the USD/JPY.

    When I switched it over to another day, it still got 190 pips (after costs).

    It needs additional rules defined because it lost money on the third day I tested it. However, it still picked off excellent highs and lows, it just isn't complete and missed some other issues.

    But my confidence in totally automating SCT has been renewed.

    In fact, after sleeping on it, I see a number of potential simplifications, enhancements, etc.

    It almost appears that Jack's seemingly overly simplistic table of P,V relationship may really be that simple.

    Further experimentation will tell the tale.

    Sincerely,
    Wayne
     
    #22     Oct 8, 2008
  3. phubaba

    phubaba

    I didn't read all your posts after your discussion of what to do about volume buildups and logging (i don't understand what change x continuation means). but if you are convinced that most peaks happen near a "volume high" you might just want to send an order in, and put a stop loss close by, and then scale out as you make money. So if you are wrong you are wrong if not you make money:)

    another thing that might be valuable is looking a bit at ideas of point and figure, and trying to quantify a risk vs reward structure for your trades so you don't end up seeing a run up in volume, but taking a short with bad risk/reward characteristics relative to what you think is bullish or bearish
     
    #23     Oct 8, 2008
  4. Folks,

    While building this ATS for SCT, I have learned how to trade this discretionary.

    I learned the flaws and how to continue or change with the P,V relationship and practiced with my replay system.

    So I plan to trade it discretionary several hours per day while still working full time.

    Here's my question. If I'm comfortable trading this to get 50 to 100 points in a couple hours per day, how much money should I risk per pip?

    Is money management important with confident SCT trading?

    I have $800 in my Forex account so I have enough margin to risk $10 per pip or only $1 per pip or 10 cents per pip.

    My plan is to start trading live 2 hours per day for 1 week while risking only $1.00 per point, just for practice.

    So I expect to make from 50 to 100 pips during that 2 hours or $50 to $100 per day. After 5 days that could be from $250 to $500 profit.

    Then with over $1000 in my account, would it be unwise to risk $10 per pip from then on?

    That way each day will be from $500 to $1000 profit per day?

    I'm okay doing that for the first week or two till the account gets to a few thousand dollars.

    Then I can keep trading $10 per pip until the account get's large enough to risk $100 per pip, etc.

    Please, what are your thoughts?

    Many traders and money management guys recommend risking less than 1% of your account per trade.

    I do get small losses from time to time doing this of 10 pips or so during the day and then trades of 80 pips, etc.

    So I could right 80 cents per pip would would be $8 if a 10 pip loss so only 1% of account is lost.

    But I feel confident enough about trading this and it's so relaxing and non-stressful that I could risk 10% of my account loss since I know I doubt have 10 losses in a row and wipe out my account.

    But even if that happened as a mistake or two, and I'm still confident in my skills, I could easily replenish the account from my checking account.

    But once the account gets to $10,000 or some larger number, I would prefer risking only 1% per trade just in case of any catastrophe to avoid wiping the account.

    I appreciate your advice.

    I'm able to trade all the B2B B2R R2R R2B etc confidently now--learned how to avoid all the flaws. So that means all the traverses are easy now.

    Later I want to master playing the traverses of tapes themselves to squeeze more efficiency out of the trading to make more money in part time work.

    Sincerely,
    Wayne
     
    #24     Oct 11, 2008
  5. Oh, the reason I want to do it part time is both to keep the full time job for now until I can routinely extract enough to replace that income.

    As soon as I achieve that goal, I want to drop that employment and finish automating this system.

    That's doubly because I love software development and secondly to increase efficiency and trade multiple forex pairs simultaneously.

    I could probably only handle at most 2 pairs live with simultaneous trading discretionarily.

    How AMAZING and a religious experience this is. Whew. I can't sleep now that I can SEE the markets as of today.

    I replay several days in accelerated time as drills and always profit handsomely.

    I print the charts and not any mistakes and review them to improve efficacy.

    Wow. Wow. Wow. It's all I can say.

    So amazing to trade in such a totally relaxing, non stressful way.

    Unlike the past experiences of day trading.

    It really almost does feel like reading tomorrow's newspaper today.

    After paying off some bills sorely behind, I can't wait to donate to charities and help society in other ways.

    How utterly and completely fun and exciting.

    Wayne
     
    #25     Oct 11, 2008
  6. Oh, by the way, now that I can see the market clearly with regard to continue/change, tapes, channels, the dom, etc.

    It will be a cinch to totally automate this.

    I'm just impatient to start generating money and would prefer to spend the time actually trading for profit than coding this up until replacing the day job income.

    Then I can spend full time automating this thing.

    I have it about 70% automated....that's when I realized, wow, I can do this myself now.

    I even wonder, who needs a computer when you can easily make $1000 an hour, part time? I might even find other things I prefer to do than coding on this.

    Maybe it's just beginner's "over confidence" but it almost feels like I could even trade blindfolded now!!!

    I'm all ears for any other advice anyone has about SCT now that I'm at this point finally.

    Wayne
     
    #26     Oct 11, 2008
  7. phubaba

    phubaba

    hey man, congratulations.

    Firstly, I'd say that you are right to try to trade using a very clear system that would be implemented into your automation.

    however, it seems to me that you must not be very confident in your ability to automate it because (if it is possible) the automation will absolutely destroy your ability to trade it.

    It'll trade 24 hrs a day (forex is 24 hrs remember!) and you can make sure that it doesn't risk as much in terms of a percentage until you trust it to just go on its own.

    This is wayy more valuable and will give you way more freedom than day trading.

    But if you feel like trading it part time will give you some more insight then by all means do it.

    But from what I can tell you are already making some pretty unbelivable claims. You are starting on an 800 dollar account and are already thinking of 1000/hr profits? are you insane? remember that there are many people who make millions paper trading and die live trading.

    either way, good luck! don't let your dreams of riches get to you. If people were making 1000/hr, you would definitely know it because this sytem has been so popularized.
     
    #27     Oct 11, 2008
  8. A couple things. One, is that even though Forex is 24 hours, there's only a certain range of hours during the day when there's enough volume to use the synthetic volume to do the SCT trading.

    Also, some P.M.s help me realize it's better to keep the risk as low as possible.

    Then after I have established a track record, (however small per trade), it will become easy to get more capital to trade and always risk a very small percentage per trade yet make a lot of money.

    Automating this rules of the SCT is easy, the hard part of automating is all the automating indirectly related to trading rules.

    For example, right now there is some technical issue that causes my quote server to intermittently get disconnected about once or twice a weak.

    It has logic to reconnect automatically and that works most of the time. But something happens on occasion and logs that it reconnected but fails.

    I need to add more logging messages, etc. And eventually get that debugging. Those kinds of intermittent issues can be a bear to debug.

    Another, (last example) issue was my discovery that while it auto-traded individual days quite well with a couple exceptions (before the new logic I learned) when I tested all the days together they mostly all failed except the first one.

    Turns out that since my strategies use multi-timeframe bars, they weren't synchronized. So joining them together through out the timing of the bars related to each other from one day to the next.

    That took over 12 hours of design coding and testing to fix it so all the timeframes of data select the same sync point in time and remain synchronized.

    Etc. Etc.

    I would say that after months of work including many hundreds of hours, I'm about 80% complete with all the logic and code needed to fully automate the trading itself.

    But you know that 80/20 rule, right? Well it could take an equally long time and hard work to get it perfect.

    Plus, before I can trust it with major amount of money, I need to build a monitoring/recovery system that will check for any failures and restart.

    Not to mention redundancy of synchronizing too servers so that if one fails or dies due to hardware, the other can log in and take over from there.

    So once I can make a living trading part time, I can spend the time to perfect the automation.

    In fact, it would be cool to make enough to pay a software developer to help do the rest. We'll see.

    Wayne
     
    #28     Oct 11, 2008
  9. Specterx

    Specterx

    I would trade simulated for at least 1 month (many people will say 6 months), and thereafter only enter the market with real money when you're completely sure of the system. If you have a particularly bad day, a couple of normal down days close together, or you just don't like how it "feels" (maybe you've had some large drawdowns) then extend the trial period for another week, and so on.

    From what I've understand you've been trading off accelerated replays. If you're going to do this discretionary then I would strongly suggest a long period of live market work as above. Your sense of the market's normal 'pace' will be thrown off by just using the replays. And of course during the day you'll have issues like boredom, distractions etc. that will affect your net results.

    Other than that, good luck to you, I hope you keep us updated :)
     
    #29     Oct 19, 2008
  10. Thanks for the tips. Some responses...

    Actually, I have been delayed from getting started because my app couldn't run at work due to a lot of outbound ports blocked. But a few are open.

    So I built software to tunnel all the ticks from the dedicated server to the office PC using an open port.

    Then I added features to my home-grown trading app so that it has a "compact" mode whereby it shrinks to a little space on the right side of the screen (like the toolbar).

    That works because once the "context" for the day is clearly in mind, all you need to see to trade is the current few bars including volume and the channels. So it works great as a very small app on the right side of the screen.

    In compact mode, it saves space by not showing the title or scale on the chart and abbreviates other stuff.

    That resolves boredom because it's easy to trade while working on other stuff.

    I will consider buying a LCD monitor to take to work for additional screen real estate.

    Also I programmed a couple hot keys to buy/sell/reverse/flat that work no matter what primary application I use. That saves the delay of having to click the mouse or alt-tab to get focus to the trading app to do an order.

    That way the market info shows on the right side of the screen while I work on other stuff.

    Tomorrow will be the first day to really give this a try during work hours. I spent all week (in evenings) working out the kinks in the network communication, making it SSL, secure, auto recover, etc.

    I'll post back with progress.

    FYI, I can't stand simulated trading for 2 reasons. First, the broker uses different trade date that is also "simulated", the simulated servers are less reliable (they go down at times).

    Plus it's too easy to lose focus or lose interest since no money is really at stake. I never learn as much that way.

    What I love about forex is that I can trade "micro" size so that, while money IS involved, it's very LITTLE money.

    In micro trading, every tick is only 10 cents. CENTS. Do you realize how little that is?

    Even if a really bad trade lost TWICE the daily range of about 100 pips, that would still be only $20.00 loss.

    So micro trading works excellently for forward testing after back testing is finished.

    Simulated trading, in my opinion, is for the birds.

    Does anyone agree/disagree?

    Sincerely,
    Wayne
     
    #30     Oct 19, 2008