Avoid positive slippage

Discussion in 'Order Execution' started by Aetey, Aug 9, 2010.

  1. Aetey


    Before entering a trade I'd like to get confirmation by the market that it's headed in my direction. That's why I enter my trades with a stop-limit order that's placed in the direction of my trade.

    Example: let's say I identified a short trade; yesterday's close was 100. I'd then place a stop-limit sell order at 98 (with the stop somewhere around 98.5). Now when my stop gets hit it's not unusual that I get filled at 99.

    This slippage is in my favor (I wanted to sell short at 98 but I sold short at 99) so I cant really complain -- if it isn't for the fact that it's completely negating the reason why I put the stop-limit where I put it in the first place: a move from 100 to 99 is not as convincing as a move from 100 to 98.

    Is there any way to ensure that I do not get this "positive" slippage? Are there any types of orders I'm missing that I should check out? What about conditional order executions? I'm with IB...

    Any help would be greatly appreciated!
  2. Pita


    if you would prefer to get negative slippage instead of positive then just omit the limit for your entry stop:p

    joke aside your entry is not good that's all, nobody ever will complain about positive slippage.
    You are apparently a (too) safe player or perfectionist while both approaches don't work out in trading.
  3. zdreg


    please explain the order more clearly. why should the order ever be filled at 99? the stock is at 100 the sell stop is at? the limit is at?
  4. Orders are instantaneous actions. They ensure nothing about what will happen in the next instant. otherwise, we would all be rich by now, I mean rich, filthy rich, you know:)
  5. The best way I have found to achieve a desirable positive slip is to do it synthetically.

    When your target entry is reached, look at the trade data (ticks) - price action rarely moves linearly in one direction for very long.

    Wait - when the desired spread is reached between you target entry price and your desired slip - submit your order - yes you can even submit a market order (if your on the CME it is recast as a limit).

    For this to work effectively you need low latency - ping the server your placing order through and/or the server your getting your data from. Look at the number of hops and the computed latency.

    Second - check the time stamp on the trades you receive as well as the time stamp on order confirmations and compare them to your clock.

    You will probably have to do some extra programming to accomplish this - but like I said at the beginning, this technique works well.
  6. Oh Yeah,

    It is always good to check the actual latencies - if they are too high - consider other providers. The best providers can be quite expensive, but the quality of the data can be superior (under 10 milliseconds)

    For the absolute fastest fills - you will need to co-locate your trading system - this can be very expensive - but if your system is profitable, it may be worth the cost.

    Look at Reuters - great data
    Look at TT - great order fulfillment

    Townsend Analytics data is also good

    Can't comment on NT's Zynfire - if that is what your using, verify it - don't take anyone's word for it.
  7. ...Not endorsing any providers here, the three listed are all excellent.
  8. what if the desired spread is not reached and the target price never returns? if the answer is don't take the trade, then you'll be faced with adverse selection. there is usually much more involved in decreasing slippage and not at all trivial.
  9. I am tired, I will answer tomorrow...If I remember...ping me just in case.

    My response may be lame,...But accurate (at least for the brief period that you have to digest and respond (thank goodness for open dialog)),...

    ...I don't have a magic bullet yet,...Maybe never,...But I am solving a lot - and I will gladly share my solutions - and your rebuttal is expected.

    (I) Just - re-sunk my teeth into this (figuring out the trade and all of its nuisances) after 10 months.

    (I am) Running Models - so far so good - live money this week (I hope) - not expecting a profit - just want to be certain it all works (the trading app and all I developed over the past three years).

    I will be testing - slip control (and other stuff) - aggressively.

    Stay in this loop, PM me if you desire - I will let you know what I am doing and how well it works.

    Maybe it is something that you can use.

    (Then again) Maybe I don't know what the hell I am doing!

  10. P.S. (My) Models are showing results of 7 to 12 percent return a day - and my models are (usually) within ~3.3% of real money trades.


    #10     Aug 10, 2010