Is it possible to trade this system?

Discussion in 'Automated Trading' started by billyjoerob, Jul 9, 2009.

  1. Would it be possible to a) cut and paste a whole bunch of stock symbols into a system then b) set a trade to trigger at the new high (meaning that I don't have to put in my buy stop for every symbol) and then c) the trade gets executed if daily volume (to that point in the trading day) is some multiple of the usual level (say 1.5) and d) trade lot size is some function of the ATR and e) the system automatically puts in a trailing stop as a function of ATR and f) stocks would be sold after x number of days if they didn't reach some multiple of the ATR.

    Even better, would it possible to create a scan which would then spit the symbols into a watchlist, and then the system would select the stocks to buy based on b-f above?

    I don't have a particularly sophisticated system (KISS), but I got tired of setting stops and manually entering symbols into AMTD.

    So can some of the geniuses here point me in the right direction and say if a) this plan is possible, b) which broker should I use and c) how much time is this going to take to set up?

    Or am I full of it? Let me know!
  2. foible


    A) it's possible
    B) any broker, but those with a good API are easiest (IB, TS, etc.)
    C) Skilled programmer could do it in a day or two. Since you have to ask, a month or two.
  3. Sumeet


    What do you mean by ATR and ATMD?

    Would appreciate your reply.

  4. edbar


    ATR = Average True Range

    True Range = Highest Price of the day minus the Lowest Price of the day.

    ATR(14) is the Average ATR over 14 days.

    A better indicator is ATR%, because knowing an ATR is .50 isn't helpful as a .50 move for a $100 stock is radically different than a .50 of a $2 stock.
    If you know a stock moves 5% a day, you can set your profit goals and stop losses accordingly. i.e. don't expect a 10% return in a day if the stock only moves 3% a day.


  5. Two rules in software development

    1. It always takes longer than you think it will

    2. See rule number one.
  6. this statement is not intirely true as the TR is the difference between

    highestvalue of (close last bar - high this bar)
    lowest value of (close last bar - low this bar)

    in easy langue the 14 bar atr woul be calculated like this

    ATR = average(( MinItems( Close[1], Low ) - MinItems( Close[1], Low ) ), 14 );

    you are not full of it this is a small project of a couple days
    it need only litle programming skills, but without any
    programming experience it will not be easy

    good luck
  7. Thanks for the replies. After poking around the AMTD strategy desk, I found it does most of what I want. I can dump my symbols in and it will automate the orders and trailing stop. I'll see how that works. I can't use ATR to modify position size, but now that I think about it, that isn't as important as I thought.

    My plan is to automate some form of breakout trading, buying stocks at new highs, high alpha stocks (from the bar chart list) with nice bases. Standard Dan Zanger/William O'Neill methodology. In addition, that methodology should keep me out of the market during bear markets, like the one we've just seen. Breakout trading also benefits from the 20/80 rule: most of your profits are going to come from a handful of stocks. If you're grinding out 1% per trade, you need to be right much more often.

    My question is whether any system based on mean reversion indicators like Stoc and RSI isn't fundamentally flawed. As you know, those tend to work until they don't. Stocks can stay oversold much longer than you can stay solvent. Why try to grind out wins in the trenches when you can benefit from trending stocks at new highs with no resistance above? (I avoid shorting because the arithmetic of going from 100 to 50 is 1/2 as good as going from 50 to 100).
  8. auspiv


    Ninjatrader can do all this and more. I actually just built a strategy very similar to this, but it currently only works to the downside. In my opinion, stocks and commodities fall faster and harder more often than they rise, so that's why I focused on the downside first, but adding the upside calculations to this strategy would increase profitability drastically.

    ..and I'd share the codes, but you've gotta struggle a bit before anyone hands you anything.
  9. I've thought about that re shorts, that they fall faster. If you read about any big trader, none of them made much of their money shorting. The other problem with shorting is that you get less long as the postion goes with you, and get longer as it goes against you. You could always pyramid the short I suppose, but I've never heard anybody do that. And then shorts tend to be more volatilie, ie the biggest rallies occur in bear markets. So you'd need a larger stop unless you wanted to always get stopped out.
  10. edbar


    Billy, every day trader I know, Shorts. Look at this last crash (last 8 months), it's the short players who cleaned up. And, they didn't have to wait long to do it either. When the market it goes down, it goes down good and hard. Whereas trying to make money on the Long side is slow with a lot of fake dips causing you to get stopped out, as the stock turns and heads back up.

    I've heard a lot of complaining about losing money (as the mutual funds dropped 50% last year), but all from Long traders. The Short traders are too busy partying!

    To me, getting stopped out, simply means that your strategy didn't do what you expected it to do, before it went the other way (or before it did the opposite of what your strategy thought it would do). So, the Stops aren't the problem. The problem is in setting your Profit Goals and then pulling the trigger on the profit goals to take the profits. This is the advantage of going 100% automated. The software has no problem grabbing the profits without emotion. Very seldom do I ever get stopped out. Not that it doesn't happen. Sometime the market does something really crazy and I eat a trade, but it's rare and hardly affects the bottom line.

    To answer your original question of this thread, Yes, the CoolTrade system can do most of what you are asking and it doesn't require any programming (just point and click with a tab-driven strategy builder). Very easy to use. However, it does not have the historical volume like what was the volume this time yesterday kind of stuff. However it has current volume and you can use a host of other indicators readily available for added trade entry indicators, including sector movement indicators that tell you what % of stocks are actually going up vs down in the market, minute by minute. So if 90% of the stocks in the market are heading down, you can wait a bit before going long, etc. Very powerful.
    So, yes. Automated systems can do what you want AND A LOT MORE!

    CoolTrade CEO
    #10     Jul 11, 2009