Do people get same results with real trading as backtesting?

Discussion in 'Automated Trading' started by cunparis, Jun 14, 2008.

  1. whenwood

    whenwood

    If you trade 5% volume of a stock for the day you are definetly going to start affecting it signifigantly. Even one percent.

    Small size would be a few ES futures contracts, or a few hundred shares of AAPL. For apple that would be something like .00001 percent of daily volume.

    But as you said the market on open orders might help.
     
    #31     Jun 18, 2008
  2. I was thinking I could get buy with 1 or 2%. This is going to be a challenge.

    For now I'm trading 100 shares, which obviously aren't earning me anything. 5% of 100 shares = lunch! So I'm thinking of doubling this slowly and observing the effect if any.

    I'm also considering limit orders to help keep me from affecting price.
     
    #32     Jun 18, 2008
  3. edbar

    edbar

    If you float your limit orders in public view, you will broadcasting your intentions to every professional trader and institutional computer. That will affect the market/price.

    Whereas, if you trade with a system like CoolTrade, it tracks your orders internally (on your PC only), essentially hiding your orders from the public. Then, when your target is reached, and in many cases exceeded, it fires the order to just take you out of the position.

    Good luck!

    Ed
     
    #33     Jun 22, 2008
  4. whenwood

    whenwood

    Ed I understand you are a paying sponsor so it gives you the right, but it is SOOOOO annoying to see you interject your product into every discussion. We get it, you have a nice auto-trading platform. Lots of advantages over manual trading. If someone wants to find out more about your product he/she will say something like 'tell me more about CoolTrade' or look at your add and visit your website.

    Besides just because you hide an order does not mean it will not affect market price. It will just affect it immediately upon sending, assuming it is big enough. Whereas the limit price's effect is already built into the market. Your choice, affect market now or right before you want to trade. Not sure which one is worse.
     
    #34     Jun 23, 2008
  5. But if I understand correctly, this is what IB's scaling feature is designed to avoid. It puts out 1 limit order at a time. you pick the # of shares.

    for example

    limit 100 10.00
    limit 100 10.01
    limit 100 10.02

    but AFAIK the public only sees one at a time.

    I appreciate you recommending a potentially useful software but I'm still learning what I have and I'm trying to minimize costs and keep it simple. So far IB and Quotetracker have been terrific.
     
    #35     Jun 24, 2008
  6. QuantBot

    QuantBot

    Greetings:

    I ‘m guessing those who are keeping up with this thread may want to hear from someone who’s actually executing a fair number of trades daily using the Assent Robot which is powered by the TI engine.

    The Odds Maker is simply telling you what has happened when a certain event is acted upon (Long or Short) with the application of a user defined set of rules for managing the trade. Using The Odds Maker I can easily access the exact trade data(exact time of event, price of stock when event was triggered, etc..) to go back and review charts for accuracy. Any trader in his right mind would do this before risking $1.

    Here is a screen shot of one of the strategies I run completely automated, all entries come in the 1st 5 mis of the mkt and the hold time is 10 minutes, this is a long strategy.

    (see attached screen shot)

    So I’m seeing that this strategy kicked out 441 trades in the past 15 trading days and applying a certain rule set with hold time and stop loss, generated 94 points. Is it realistic to think I will actually make 94 points? If I were able to run mkt orders and the market existed in a perfect vacuum yes. Unfortunately we operate in anything but a “fair and orderly” market. With a strategy of this nature it comes down to throwing out bids below to get long. How many times have you clipped an offer to get long(even with a small spread) and you’re immediately down .15-.20 cents? ;> You get some, you don’t get others, but the ones you get still have high odds of success and you’re adding liquidity, and every penny counts in this market. There are a fair number of strategies I run on market, it all depends on the type of event you are trading. Is it a violent event(something that’s spiked $2 in a 5 min candle) or is it tame (a consolidation or nr7 type play)? The more violent, the bigger the slippage on a mkt order.When you’re trading the spreadier ones, bidding and offering is the only logical route.

    Anyone that spends a moderate amount of time running scans of this nature will quickly realize that this is simply a tool that greatly accelerates the rate at which one can learn, and identify short term trends. Going back to the above strategy-try manually back testing 441 trades. When you’re done (assuming you would finish) and wonder hmmmmm…what If I changed the hold time from 10 minutes to 20 minutes, would it hurt or help? What if I risked .50 per trade as opposed to $1? If you wanted to re-crunch all the data you could find out; Now you can find out in a matter of seconds.

    Happy Trading

    QB
     
    #36     Jun 24, 2008
  7. QuantBot

    QuantBot

    I hate to shatter the veil good feelings here, but most of the market isn't on display, try 85% of it.

    It doesn't matter if you're visible or not. If you're long at $20 with a mkt sell stop in at 19.49; the stock proceeds to trade through 19.49, you've just been "invisibly" stopped out. The algo's that make the markets know where to take it and shake it. Welcome to the real world Neo.

    QB
     
    #37     Jun 24, 2008
  8. edbar

    edbar

    This is my point EXACTLY. If 85% of the market (institutions and professional traders) is invisible then it stands to reason that showing your orders is BAD.

    I have floated limit orders and stop loss orders in public view in the past only to see the price either repell against my order or immediately go to my stop loss order and take me out early, and then rocket to new prices. With public limit orders, even a .05 goal becomes hard to achieve. Not so, with hidden orders. The price move through my goals like hot butter.

    Whether you believe you can affect the market or not, you are not being paid to show your hand ahead of time. So why do it.

    If you were playing poker, you would not show your cards ahead of time, lest you want to be a guaranteed loser.

    I'm also sorry that mentioning my product offends you. I'm just proud of the product that I have created that automatically takes these kinids of behavious into consideration. If this was a basket-weaving forum, I could not contribute at all. But I will be consciencious in the future. Thanks for pointing that out!

    Ed
     
    #38     Jun 24, 2008
  9. So, to hide orders does that mean we should avoid limit orders?

    Does IB have an order type for hiding orders?

    I'm not yet convinced that my 100 share lots are going to impact the market but I'm hoping to build up capital and confidence to trade larger lots. :)
     
    #39     Jun 24, 2008
  10. edbar

    edbar

    I don't know what brokers have for hiding the limit and stop loss orders. I know my software has it built-in called "stealth mode".

    To convince yourself whether your 100 shares of MSFT limit order matters, do this simple test:

    TEST1
    Buy 100 Shares of MSFT
    Place a Limit Order to sell .05 higher
    Place a Stop Loss to sell .05 lower

    See which one gets taken out first.

    TEST2
    Next Buy 100 Shares of MSFT
    Write down on paper to sell .05 high
    Write down on paper to sell .05 lower

    Now see which happens first.

    I contend that with TEST1 you will get stopped out 90% of the time. While with TEST2, it will be 50-50.

    I've already done it, and those are my results.

    Ed
     
    #40     Jun 24, 2008