Realistic Expectations for Automation

Discussion in 'Automated Trading' started by Norm, Dec 27, 2005.

  1. Norm



    I am leaning to use EasyLanguage for TradeStation. I was just wondering what to expect from strategy automation.

    If anyone can provide their real life experiences, good or bad, with automated trading, I would certainly appreciate it.

  2. Dustin


    Like they say..."Garbage in, Garbage out"

    Make sure you have a profitable strategy before diving into automation.
  3. If there were a fairly easy way to automate profits over time, everyone would have it. Even if there is a difficult and low probability way to automate profits, a few will have it.

    It appears to be eveyone's goal to discover the *Automated Holy Grail*... that way, they can just "pour in the money, go play golf, and count their money after the close".

    Odds are, it ain't gonna happen. Odds are worse that YOU will find it if it does exist.

    Yeah, I know.... some will pipe in, "I'M automatin' profits". That's for market conditions NOW... what about when they change? (One memorable quote I heard years ago... "All you have to do for the market to knock you off is keep doing what was successful." That is, the market changes and you'd better also.)

    Realistic Expectation? You spin your wheels and come up with bupkus.
  4. Once you as beginner start coding proprietary language
    and simple automation , these leads to other things like
    math, system languages, fix, gateways,....many terminologies
    ,system traders recomanding this, that,...they will brain
    storm you

    All of a sudden you become very confused now, You don't
    know where to start and where to end?

    The more you learn, the more complicate it gets, you can't
    apply simple methods anymore, because you don't know
    how many strategies or which strategy and where,when
    and how much/many times to apply them?

    I believe after having long experience, these confusion to
    go away and more simpler method will replace it
  5. The only order you can use in automated trading is the stop order. Market orders in realtime are delayed up to 5 sec. due to Tradestation processing so you can't count on the fills looking like historical results. Also limit orders are useless. There is no way to simulate the real world fills. Only sure way to have a fill is to trade through a price. The simulation will give it to you if price touches the limit. In real world you can sit in a queue and maybe get filled at the limit price. Since you want the sim and real trading to be equal forget about using limit orders.

    You've got to paper trade the strategy and look for logic errors such as placing a buy stop when the price has already gone above the target price. In the simulations it'll be filled as a market order. In the real world the trade will be rejected.

    You won't really be able to leave the system alone because your internet connection can be disrupted at any time (power brownouts, ISP hiccup, timeout disconnects at the broker side, the broker server going down, data feeds losing connectivity are some of what I've seen). So, the only real benefit to pure auto trading is it gives you some hands free time to do other work.

    Also, you'll see times when you use a trailing stop and it should generate a cancel order on the old stop. Sometimes you'll be left with multiple stops because the old order wasn't cancelled.

    Just some of what I've experienced.

  6. 100% accurate! Agree. Automation is the very last step in a system development. It is 99% about dealing with live orders and positions, broker messages, system errors etc and 1% = is about actual logic.
  7. Dustin


    To expand on my first post...

    I've had much success with automation in the past 3-4 years, but it wasn't until I found strategies that were already profitable. If you are still on the path to profitability I think messing around with automation can actually be a detriment because you should be spending the time learning to trade. Without these fundamentals, especially in such a low volatility environment, you will be lengthening your learning curve which ends in failure all too often.

    BUT, once you are in the green I believe automation is extremely important because it allows you to be trading so many instruments at once, and/or leaving you more time for research of new edges during market hours.
  8. Some of your statemnts are True, and some are not true:

    -5 second delay in Tradestation might happen due to
    sometimes laging data, Not because of Tradestation
    processing of your strategy orders, and this downside
    should affect you if you are extreem scalper only

    Don't forget somthing like X-trader also can cost you
    $800 a month for trading two simultinous markets,
    (No ability to trade stocks also) while Tradestation 8
    is almost free product today

    -You don't have to place simulation type (software side)
    orders, native orders for trailing stop (profit taking)
    only, then you are not bothered if your other stop loss
    orders are still active, unless they are reversing type

    Meantime with ES, NQ, YM tick size today, which used to
    be for many years ago, when volatility was much higher,
    I don't think most traders can't scalp today, unless CME,
    CBOT decreas Tick sizes, lets say ES=$4, NQ=$3, YM.....
  9. I've found that for me-simple = better = higher profits. I am no quant and so I have never had success when trying to develop a strategy that is overly complex. I have really only had success trading (relatively) simple strategies, while implementing strict money management rules.
  10. joesan


    #10     Feb 26, 2006